mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-10-11 23:08:38 +02:00
e8c8ba3c19c7e0fbd5ae138cb28684e283c04e36
1432
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 DottaandPaperclip
|
e8c8ba3c19 |
feat(apps): add experimental MCP aggregator connectors (#13755)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Its tool gateway applies company access rules and approval controls to connected apps. > - MCP aggregators expose many apps through one provider endpoint. > - Each aggregator needs its own credential, catalog, grants, and lifecycle in Paperclip. > - This pull request adds independent Zapier, Arcade, Composio Connect, and Executor setup with a common Access → Connect layout. > - A default-off MCP aggregators flag lets operators opt in while we complete provider acceptance tests. > - Agents use the normal Paperclip permissions, Test screen, and gateway after setup. ## Linked Issues or Issue Description **Subsystem affected** Apps, connection setup, shared contracts, and the remote MCP gateway. **Problem or motivation** Aggregator endpoints need clear provider setup and correct MCP sessions. Generic setup does not explain each provider's authentication or broad execution tools. Provider approval must preserve the original execution instead of replaying a write. **Proposed solution** Add four separate connectors behind Settings → Experimental → MCP aggregators. Start with human and agent access, then connect the endpoint and read its tools. Enable tools by default. Use the existing Permissions and Test screens after setup. Keep legacy Composio API-key and child connections intact. **Alternatives considered** A shared connection for all providers would mix credentials and access rules. Separate provider-specific permission and test screens would duplicate existing controls. Vercel Connect is outside this change. **Roadmap alignment** Extends the existing MCP Tool Gateway & Apps capability and the Connected Apps roadmap area. This work was requested and reviewed by the maintainer. Related work: #11894, #12630, #12632, #12634, and #12906 concern the legacy Composio broker. #13102 also covers remote MCP pagination. This change preserves the broker path and adds initialized sessions, response matching, and provider resume handling alongside pagination. ## What Changed - Add branded setup and interactive Storybooks for Zapier, Arcade, Composio Connect, and Executor. Use the existing access controls and normal action tests. Do not request a connection name or action choices during setup. - Add the default-off `enableMcpAggregators` flag to settings, managed feature metadata, the catalog, and setup guards. Hidden connections keep running. Legacy Composio connections remain unchanged. - Reuse the vault, grants, policy, and catalog models. Support OAuth discovery, bearer tokens, custom headers, and credential-bearing URLs. Add no database tables or migrations. - Initialize and retain Streamable HTTP sessions by connection and effective credentials. Read paginated catalogs and match streaming responses to request IDs. - Classify unfamiliar aggregator tools as writes despite upstream read-only hints; only exact reviewed read capabilities enter the read-only allowlist. Legacy Composio child behavior is preserved. - Preserve provider authorization links and execution IDs. Support Executor approve/resume, decline, and cancel without automatic replay of uncertain writes. - Preserve Off and Ask first choices during refresh and reconnect. Allow new tools and retire removed tools. Keep agent access updates atomic and preserve an empty agent selection. - Document connector UX rules, provider branding sources, and live acceptance results. - Stabilize the existing Sentry release fixture after its repeated CI failure by reusing one module mock; production Sentry behavior is unchanged. ## Verification - Final head `d11781970`: [CI run](https://github.com/paperclipai/paperclip/actions/runs/35633534900) passed, including broad typecheck, test shards, build, and E2E. All 54 checks pass; 2 optional checks are skipped. Greptile is 5/5, Security Scan passes, and all review threads are resolved. - Passed 27 focused connector Vitest checks and 18 connector-only Storybook browser checks before the flag change. All 85 stories rendered at desktop and narrow widths. - Passed 5 connector lifecycle/server checks and 7 selected flag checks after adding the flag. The latter cover settings, managed defaults, cached catalog visibility, and all four setup routes. - Review fixes passed 13 risk/handoff/lifecycle checks, dedicated session-expiration and transport regressions, 13 selected connector/gateway CI cases, and 10 selected setup/reconnect UI cases. A real Composio connection-list call also succeeded through the refreshed UI on `9ab115f71`. - UI and server TypeScript checks passed. UI build, Storybook build, token gates, and diff whitespace checks passed during implementation. - Real browser and real Paperclip agent tests passed for Arcade, Composio, and Executor. Tested action permissions, denied agent access, reconnect, disconnect, and isolation. Tested Arcade catalog additions/removal and Executor provider approve/resume, decline, and cancel. - Zapier live acceptance is incomplete. Its dedicated provider server is configured, but its credential-copy dialog returned an empty clipboard through browser automation. No live Zapier action is claimed. - The three isolated Sentry release cases pass after the CI fixture fix. - Local verification is deliberately narrow at the maintainer's request. The full local suite, recursive typecheck, and repository-wide build were not run. CI provides the broader checks. ## Risks - Shared MCP transport changes affect other remote MCP servers. Protocol fixtures cover initialized sessions, streaming response matching, pagination, and isolation. - Broad execution tools remain broad permissions. The provider governs actions inside those tools. - Provider handoff links are retained briefly in memory. After a server restart, a one-time link may require reopening the provider dashboard. Paperclip does not replay the original call. - Zapier remains unproven live. Custom-header imports and self-hosted endpoints have fixture coverage rather than a separate live account for every variant. - Turning the experimental flag off hides setup; it does not revoke existing credentials or stop existing connections. ## Model Used OpenAI GPT-6 through Codex, with reasoning, repository tools, shell execution, and browser automation. The exact runtime model ID and context-window size are not exposed in this session. A separate Anthropic-backed Paperclip agent performed live gateway acceptance tasks. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
3790ca2f13 |
fix(runner): repair approval and Stop races and eval infrastructure (#13750)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Runner tasks must continue after approval and stop when the user presses Stop. > - Live evals found races at approval delivery and provider startup. > - Browser readiness and CI setup errors also hid the actual task results. > - This pull request fixes those races and the related test infrastructure. > - Regression tests and saved live reports show which cases now pass. ## Linked Issues or Issue Description Companion eval definitions PR: https://github.com/paperclipai/paperclip-evals/pull/25 (AgentCore paused and provider/environment infrastructure). Related: #13741 now supplies the late-startup Stop fence and warm-attachment recovery; this PR retains that fence and extends startup tracking and regression coverage to both native backend paths. #13539 introduced queued approvals during active runs. #13738 fixes child assignment, task replies, and warm process continuity and is already in the base. #13291 concerns automatic continuation of interrupted legacy sandbox runs; this PR fixes native startup cancellation and does not change that recovery policy. **What happened?** An accepted service approval could wait after its source run stopped. Stop could return success before the provider handle existed. Work could then start after Stop, or a cancelled run could be recorded as failed. Some E2E tests also failed on unloaded browser content or irrelevant reply wording. Runner CI could fail before model work because of dependency or sandbox setup. **Expected behavior** Deliver each settled approval once after its source run stops. Do not start work after an acknowledged Stop. Preserve the audited cancellation. Test the intended product behavior with a ready browser and verified runtime dependencies. **Steps to reproduce** 1. Approve a service request while its source run is active. Let the run finish. Check that its result starts one continuation. 2. Delay provider startup. Press Stop before its handle is available. Check cancellation, then submit `/new`. 3. Run the browser, warm-workspace, and Stop-and-redirect cases from the linked report. **Paperclip version or commit** The branch includes master at `9d19f98b5`. The report records the original source for each focused attempt. **Deployment mode** Isolated local development instances and disposable Daytona sandboxes. ## What Changed - Deliver settled tool-action results for the exact company and source run during final cleanup. Keep the existing idempotent receipt and periodic recovery sweep. - Wait for startup to hand off its provider handle before acknowledging Stop. Reject first-turn admission after cancellation. Preserve a matching audited pending or acknowledged cancellation. - Wait for mounted task history and connector controls in browser tests. Record failure evidence. Grade workspace contents and process continuity separately from exact reply wording. Require each warm-turn marker once and in order, allowing surrounding prose. - Stop-and-redirect now checks that the source file exists and work is active before Stop. - Resolve target dependency locks in an uncredentialed CI job. Verify the lock artifact hash. Keep orchestration and publication on the trusted workflow revision. - Materialize the pinned OpenCode executable and configure the exact Codex executable's user-namespace profile before provider credentials are available. - Compress Daytona directory uploads with gzip. Preserve files, executable modes, symlinks, empty directories, and confinement checks. - Classify file-transfer RPC deadlines as infrastructure. Keep unrelated runner RPC failures visible. ## Verification - [Focused live report with screenshots and original attempts](https://pages.paperclip.ing/runner-reliability-20260921/): 14 of 15 selected Product E2E cases pass across the recorded revisions. Claude and Codex Stop → `/new`, Claude service approval, delegation, both hiring/reuse cases, and native Daytona warm continuity pass. - Two credentialed Runner smoke cases pass. These are not full protocol coverage. - E2E harness after the master merge: 429 tests pass. E2E and server TypeScript checks pass. - Daytona plugin: 239 tests pass, 6 skipped. Plugin TypeScript build passes. The compression test fails against the old code and passes with the change. - Runner backend/runtime regression group: 161 tests pass. Cancellation/startup selection: 26 tests pass. Approval delivery: 34 real-database tests pass. - Workflow security: 7 tests pass. Both edited workflows pass actionlint. Runner TypeScript and Rust builds pass. - After merging master, all 389 native executor tests pass, including both native backend paths and late startup after the Stop deadline. - Post-merge `pnpm -r typecheck` and `pnpm build` pass. The monolithic local `pnpm test:run` was interrupted to integrate master and is inconclusive. The [hosted CI test partitions](https://github.com/paperclipai/paperclip/actions/runs/35620461738) pass on `50a3e43822bcba1e0d07b1b45b0be91cbf9312da`. An unchanged sandbox callback schema test initially received HTTP 503. It passed five isolated local runs, its full local test file, and one failed-job CI retry. No assertion was weakened. ## Risks - Stop can wait for the bounded startup handoff. If it cannot settle, the existing pending-recovery state remains instead of a false acknowledgement. - Immediate approval delivery must remain idempotent across cleanup and recovery sweeps. Tests cover duplicate delivery and company/run boundaries. - The workflow changes still need hosted Linux verification. They retain the trusted workflow and credential boundaries. - Gzip reduces the observed provider upload from about 1.8 GB to 663 MB. It does not yet fix the remaining Claude Daytona transfer timeout. That recovery test never reached Claude, so recovery remains unverified. Use a matching image with the verified provider package preinstalled for the next recovery test; retain cold-upload coverage separately. - The report preserves diagnostic runs with missing source metadata and marks them as such. It does not claim a new full-suite pass. - This PR adds no new prompt policy or historical status reconciliation. ## Model Used OpenAI GPT-6 through Codex performed the primary implementation and review. The exact primary backend model ID is not exposed in this session. OpenAI `gpt-5.6-luna` assisted with bounded infrastructure work and verification. The agents used repository tools, code execution, and browser tests. The exact backend revision and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: OpenAI GPT-6 <noreply@openai.com> |
||
|
DottaandPaperclip
|
9d19f98b50 |
fix: harden native chat recovery and add coordination evals (#13741)
## Thinking Path > - Paperclip helps people manage AI agents and their work. > - Agent chat uses native runner sessions to plan, delegate, and track that work. > - A user can press Stop while the native session is still starting. > - The server can acknowledge that Stop without dispatching it, then let the session submit a turn. > - This leaves chat recovery waiting for an execution that the user expected to stop. > - This PR waits for the startup handle, dispatches cancellation, and prevents a late startup from submitting a turn. > - New full-stack evals check the resulting records and outputs across Claude and Codex. > - Those evals also exposed missing ACPX readiness fields, unbounded polling, and an old-run identity check that rejected valid warm handoffs. ## Linked Issues or Issue Description **What happened?** Stop during native startup could record an acknowledged cancellation with `dispatched: false`. The provider could then begin work. A subsequent `/new` stayed queued. A remote Claude follow-up also exhausted the command journal while probing warm-session readiness: ACPX never returned the readiness fields required by the shared transport. Once readiness worked, attachment incorrectly compared the next run descriptor against the old run ID. The 25 ms polling loop could issue 4,800 commands during its two-minute wait, beyond the 500-command bound. The existing chat eval treated lifecycle logs as proof of an active provider turn, so it did not distinguish startup cancellation from active-turn cancellation. **Expected behavior** A Stop during startup must reach the pending session. A late session must not submit a prompt after Stop. Recovery must retain control when startup exceeds the bounded wait. Chat evals must check saved task state, document contents, worker identity, account binding, and duplicate effects. **Steps to reproduce** 1. Start a native Claude or Codex chat turn. 2. Press Stop after process startup is requested but before the provider turn starts. 3. Send `/new`, then send a fresh message. 4. On the affected base, cancellation can be acknowledged without dispatch and the reset stays queued. **Paperclip version or commit** The live Claude baseline reproduced this on `29d6b3509`. The branch also includes master commit `0f5fafe16`. Related work: #13678, #13686, #13693, #13291, #13738. A separate runner reliability branch also contains a startup-wait fix. Its overlap must be reconciled before merging; this branch additionally prevents prompt submission after a late startup. ## What Changed - Wait for a pending native startup before acknowledging a run-scoped Stop. Preserve the existing recovery error when that wait expires. - Keep a Stop guard on startup. Cancel a late handle before it can submit a provider turn. - Add regression tests for normal handle publication and publication after the Stop deadline. - Back off blocked warm-attachment probes. Keep the fast two-snapshot barrier, fail closed, and record changed blockers. - Add red/green tests for delayed readiness, persistent blockers, alternating readiness, and readiness near the deadline. - Publish ACPX readiness and blockers. Preserve the old authority’s event acknowledgement barrier; only settled sessions can proceed to attachment. - Bind warm ACPX descriptors to the validated next authority while retaining old-run event correlation until activation. Preserve session identity and provider profile checks. - Exercise two consecutive run rotations through a qualified fake sidecar, verifying checkpointing, provider identity, pre-activation rejection, and new-run work admission. - Separate startup and active-turn cancellation checkpoints in the browser eval. - Add 18 explicit native chat eval cells: 12 local and 6 Daytona cells across Claude and Codex. - Cover hiring and reuse through managed AI accounts, source-based review, current blocked-task status, request replay after a lost HTTP acknowledgement, server restart continuity, and Stop/reset continuity. - Use ordinary production agent instructions. Enable API tools only for the two coordination cases that need them. - Calibrate the matchers with invalid records and outputs. Require remembered context after restart and a structured status snapshot that distinguishes the current blocker from history and task status from active execution. Compare the public issue mutation contract and relationships during read-only reporting. Preserve before/after source records in failed eval evidence. - Fix the lost-ack browser harness and verify it against a real HTTP server. Check the chat composer after restart instead of waiting for an unrelated document lifecycle event. - Document the scope and limits of each case. ## Verification - The startup regression failed on the unfixed executor and passed after the fix. - `pnpm test:e2e:runner:typecheck` passed. - `pnpm test:e2e:runner:unit` passed: 424 tests in 37 files. - `pnpm exec vitest run server/src/services/native-runtime/native-session-executor.test.ts` passed: 385 tests. - [Baseline live campaign](https://github.com/paperclipai/paperclip/actions/runs/35608208868): Claude Stop reproduced the bug. Codex Stop and Claude hire/reuse passed. Codex delegation was blocked by provider capacity. - [Eval-only startup campaign](https://github.com/paperclipai/paperclip/actions/runs/35609479786): both providers failed as expected. Both persisted `dispatched: false` and left `/new` queued. - [First fixed campaign](https://github.com/paperclipai/paperclip/actions/runs/35610533706) on `c9e95797d`: 10/18 cells passed. Startup Stop passed for both providers. Failed cases exposed eval harness defects and remote continuity failures. All attempts remain available. - [Original workflows and stronger memory checks](https://github.com/paperclipai/paperclip/actions/runs/35611896649) on `c04324fab`: 9/12 passed. Reassignment, local restart memory, and startup Stop passed for both providers; Codex remote restart passed. Claude remote restart exposed the missing readiness contract. Two Codex planning cells hit provider capacity. - [Unchanged-model retry](https://github.com/paperclipai/paperclip/actions/runs/35613854548): Codex planning and backlog creation both passed. - [18-cell campaign with ACPX readiness](https://github.com/paperclipai/paperclip/actions/runs/35614586963) on `6a98ef743`: 16/18 passed, including all local/remote Stop and committed-send cases. Claude remote continuity exposed the next-authority check, now fixed. Codex hiring produced its checklist, but the runner redacted the requested marker after it appeared as “Tracking token: …”. That content-redaction policy is unchanged and remains an explicit limitation. - [Structured status grading](https://github.com/paperclipai/paperclip/actions/runs/35614954725) on `50448c228`: both providers passed on their first attempt, including cleanup. - [Complete read-only state grading](https://github.com/paperclipai/paperclip/actions/runs/35616089011) on `551e13892`: both providers passed. - [Final ACPX handoff and hiring retry](https://github.com/paperclipai/paperclip/actions/runs/35617045456) on `cbd637587`: all three Claude Daytona cases passed (restart continuity, active Stop/reset, and lost-ack replay). Codex hiring reproduced the content-redaction failure: the saved checklist contained `Tracking token: [REDACTED]` instead of the required business marker. All four cases completed cleanup successfully. [Published report](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/gha-35617045456-1/). The only subsequent commit adds the qualified-sidecar integration test; production code is identical to this live proof. - `pnpm test:e2e:runner:browser-support` passed: 5 browser tests without paid models. - Runner TypeScript typecheck passed. All 5 warm-readiness tests pass; two failed with the prior fixed-rate loop, and the late-readiness test failed before the pacing correction. - ACPX readiness and warm-identity regressions each failed before their fixes. All 292 runner-core Rust library tests passed. The qualified-sidecar integration test passes. Rust formatting is checked. - Status-grader regressions for misleading historical mentions and previously unchecked mutations each failed before tightening the oracle and pass now. - [Latest-head CI](https://github.com/paperclipai/paperclip/actions/runs/35617522307) passed on `a4093c8f1`: full build, type checks, test partitions, browser E2E, and native runner checks. Two unrelated tests initially failed (Sentry fixture release attribution and local-service fixture readiness); both passed locally together (35 passed, 5 optional SDK tests skipped) and on the failed-job retry. No changes were made to those tests. - Greptile reviewed `a4093c8f1` at 5/5; both earlier findings are fixed and all review threads are resolved. - The paid live suite is not fully green: the reproducible content-redaction case remains red. This is separate from the passing PR merge checks. No production content-redaction, prompt, model, or completion-policy change is included. - Managed-account hiring and review cases explicitly enable API tools; these do not qualify default new-user onboarding. ## Risks - Stop can wait up to 30 seconds for startup, then use the existing pending-recovery path. This does not prove that remote cleanup has finished. - Blocked warm readiness adds up to 750 ms between later probes with the two-minute remote budget, or about 32 ms with the default five-second budget. Ready sessions retain the short second barrier. - Paid evals can fail because of provider capacity or agent decisions. Each failure needs evidence-based classification. - The HTTP request replay case checks comment idempotency and duplicate effects. It does not prove replay safety for an ambiguous provider tool call. - The new suite is opt-in. It does not increase the default paid campaign. - No production prompts or model selection change. Review-handoff behavior and content-redaction policy remain separate product decisions. The latter can remove harmless business content that looks like credential syntax; the failing attempt is retained. ## Model Used OpenAI Codex, GPT-6, with repository tools and code execution. The exact deployment model ID and context window are not exposed in this task. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
0f5fafe16b |
fix(runner): preserve task replies and warm process continuity (#13738)
## Thinking Path > - Paperclip lets people manage AI agents and their work. > - The Runner connects provider sessions to task state, replies, and delegated work. > - Full-stack tests found lost final replies, rejected helper calls that stopped the parent, and unnecessary process restarts. > - A completed child could also receive a new assignment wake that the scheduler then cancelled. > - This change fixes those boundaries and gives agents clearer teammate instructions. > - The tests retain strict completion and process-continuity requirements. ## Linked Issues or Issue Description **What happened?** A generated attachment comment could suppress an agent's final reply. A known Codex helper could stop its parent when it requested a Paperclip tool. Native Daytona processes restarted between turns because Paperclip minted an unused GitHub broker token. Reassigning a completed child queued a run that immediately cancelled. Revision instructions also allowed agents to do work assigned to a named teammate themselves. **Expected behavior** Keep the final reply. Reject helper tool requests without borrowing parent authority or stopping the parent. Keep an unconfigured sandbox process alive between turns. Treat assignment-only changes to completed tasks as metadata changes. Preserve explicit teammate assignments during revisions. **Steps to reproduce** Run the retained Runner E2E cases for file handoff, teammate reuse, Daytona warm continuity, and Legacy Claude interview/plan acceptance. The focused regression tests reproduce the reply, helper, process-lifetime, and assignment-wake defects without provider calls. **Paperclip version or commit** The live lifetime and completion campaign used `db3857807`. This PR replays the changes on master `c65fc9e3c`. See Verification for the limits of that evidence. **Deployment mode** Isolated local instances and native Runner sessions in Daytona sandboxes. Related work: #13546 handles a different queued-run issue after an issue-lock compare-and-set failure. This PR prevents the unnecessary assignment wake earlier. #13410 covers retained user services; this PR covers the provider process. No duplicate fix was found. ## What Changed - Exclude generated deliverable-binding comments from final-reply deduplication. Preserve the attachment and explicit user-facing replies. - Reject Paperclip tool and input requests from known Codex helper threads without terminating the parent. Keep unknown-thread rejection intact. - Explain how to hire or reuse a persistent teammate and preserve named delegation on revisions. Update generated protocol fixtures. - Use stable, token-free GitHub wrappers for unconfigured native sandboxes. Preserve credential isolation, configured-account rotation, and cleanup after partial staging failures. - Do not queue assignment-only wakes for done or cancelled tasks. Keep explicit reopening behavior. - Make warm-continuity fixtures create real review cards. Read the persisted final response selected by production presentation logic. Missing selected evidence still fails. ## Verification - Before rebase: 560 focused route, native-executor, and launcher tests passed. The new regressions were reproduced before their fixes. - Live E2E: Legacy Claude interview/plan acceptance passed 3/3 repetitions. Daytona warm continuity passed 2/3 full repetitions. Each successful run retained one process and provider session for all three turns. - The remaining Daytona repetition stopped after a same-URL browser reload left the page blank. Both completed turns retained the same process. Its failed verdict remains unchanged; this PR does not claim the blank-page cause is fixed. - Reports: https://pages.paperclip.ing/runner-e2e-lifetime-race-20260920/investigation.html and https://pages.paperclip.ing/runner-e2e-behavior-followups-20260919-results/investigation.html - Post-rebase `pnpm build` and `pnpm -r typecheck` passed. All 414 Runner E2E harness unit tests and its typecheck passed. Codex protocol tests: 88 passed, 2 ignored. - Latest-head CI: 55 successful checks and 2 intentional skips. Greptile: 5/5 with no review threads. The unchanged workspace exposure tests hit a fixed-port collision on the first CI attempt; their local suite passed (25 tests, 3 platform skips), and the CI shard passed on one retry. - The duplicate local `pnpm test:run` was stopped after the full hosted general and serialized test shards passed. It did not finish locally and is not counted as a local full-suite pass. ## Risks Configured GitHub accounts retain run-scoped credential rotation and can still restart warm processes. That limitation requires a separate design. Known provider helpers cannot use Paperclip coordination tools directly; they must return findings to the parent. The delegation prompt is an instruction, not an enforced guarantee; Codex Mini hiring/reuse failures remain open. No schema or workflow changes are included. ## Model Used OpenAI GPT-6 through Codex, with repository tools, code execution, and parallel coding agents. The exact deployment suffix and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub issues and shared reports) - [x] My branch name describes the change and contains no internal Paperclip ticket id or instance-derived details - [x] I have run focused tests locally and they pass; full hosted test shards also pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
9f30eb10dd |
fix: reduce chat latency and preserve managed session reuse (#13710)
## Thinking Path > - Paperclip manages AI agents and keeps their work attached to tasks. > - Chat connectors carry user messages and agent replies between a provider and those tasks. > - Each extra startup and context reset delays a reply. > - Managed account metadata was lost during adapter decoding, so compatible follow-ups started fresh. > - This branch fixes the reset and measures the remaining preparation, execution, and delivery costs. > - The changes must preserve account isolation, authorization, durable output, and recovery ownership. ## Linked Issues or Issue Description Refs #13699. The related service lifecycle work in #13410 and #13408 is separate; this branch focuses on task-bound chat response latency. **What happened?** Managed AI follow-ups started new provider sessions even after their configuration fingerprint stayed stable. The Codex codec removes unknown fields. The resume check then read the removed credential identity and treated it as a credential change. **Expected behavior** Compatible follow-ups resume the correct provider session. Changes to credentials, responsible users, permissions, or task configuration retain their reset behavior. **Steps to reproduce** 1. Use a Slack connector with a managed AI connection. 2. Send a message, then send a same-thread follow-up. 3. Inspect the configuration reset reason and the provider session identity. **Paperclip version or commit** Reproduced on `2a99de80ec52db01eead901f28323926ceaf3c1d`. **Deployment mode** Cloud staging with a native Codex runner. ## What Changed - Read saved credential identity before adapter decoding discards it. - Remove the internal credential identity from adapter-facing session params. - Test the real Codex codec and missing, changed, or unmanaged identity cases. - Preserve configured warm Codex runners and flush refreshed credentials after every turn. - Fence detached or closing session handles from successor credential ownership. - Stage current Codex launch credentials after restoring durable session history, uploading launch assets only once. - Reuse a runner binary already in the retained sandbox only when its SHA-256 matches the controller-owned artifact; still verify required capabilities before launch. - Lock the task before the run when saving results, preventing deadlocks with task updates. - Scope reusable projectless sandboxes to the company, environment, task, agent, and runtime configuration; verify Daytona sentinels for that scope. - Admit a new authorized chat message after a fully committed failed run and verified process cleanup. - Send compact deltas for verified plain-text Slack continuations. Match the actual prior run and current comment identity/body; exclude edited historical comments and prior agent output, preserve genuine brief edits and the full bootstrap fallback. - Keep attachments, omitted input, questions, approvals, recovery, and other providers on their existing framing. - Document managed session compatibility, credential lifecycle, and compact continuation boundaries. ## Verification - Workspace/session coverage: 156 tests passed. - Native session and credential ownership coverage: 390 tests passed, including exact artifact reuse, mismatches, failed probes, timeouts, and explicit artifact overrides. - Explicit continuation and durable chat authorization coverage: 172 tests passed. - Session resume and launch preparation coverage: 416 tests passed. - Result persistence coverage: 15 tests passed. The new concurrency test reproduced a PostgreSQL deadlock before the lock-order fix. - Environment lifecycle coverage: 92 tests passed, including projectless reuse and task/agent isolation at both selection and atomic handoff. - Daytona plugin coverage: 237 tests passed; 6 gated tests skipped. Standalone plugin build passed. - Compact Slack continuation and native resume coverage: 69 tests passed, including full-bootstrap retention, matching message authors/bodies, current-delivery selection, rejection of duplicate identities and historical comments, brief edits, and attachment/recovery fallbacks. - Final frozen-head `pnpm test:run` on repository-supported Node 26: 668 suites passed, 3 skipped, 1 failed; 12,797 tests passed and 82 skipped. The sole failure was a local `socket hang up` in `issue-recovery-actions.test.ts`, not an authorization assertion mismatch. All 57 tests in that suite passed three fresh reruns, and the suite passed latest-head CI. The full local invocation is therefore not claimed green. - An earlier Node 24 full run exposed an unrelated macOS symlink-cleanup failure; that 11-test catalog suite passes on Node 26 and in CI. No test behavior or timeout was relaxed. - Full local typecheck and build passed. Latest-head CI is green; Greptile is 5/5 with no unresolved review threads. - Two real Slack baseline replies took 25.1 and 24.6 seconds (24.9-second mean). Three same-thread signed probes on this head took 23.8, 23.9, and 22.7 seconds (23.5-second mean). This is a small sample and a modest wall-clock improvement, not a large or statistically established speedup. - In that same thread, uncached provider input fell from 8,514 tokens before compact input to 694–765 tokens afterward. The current delivery uses a 362-character delta; the full 19–21k-character bootstrap remains available for failed resume. Verified runner artifact preparation fell from about 1.2 seconds to 0.6 seconds. - A fresh thread created a separate task, sandbox, and provider session with full bootstrap (24.4 seconds). Its follow-up reused its own sandbox/session and compact input (28.3 seconds, including 16 seconds of model execution). Model variability and process startup remain substantial. - A signed duplicate webhook produced exactly one user comment, one successful run, and one final Slack reply. Slack's API independently confirmed the actual replies and a public task URL without an internal or pool hostname. - Earlier signed probes verified recovery after a failed run and reuse across a server deployment. The final idle test observed Daytona report the sandbox as stopped, then delivered a new reply in 19.9 seconds using the same sandbox/provider-session identity and compact input. Slack’s API confirmed that reply. - Live probes use signed synthetic inbound webhooks and real outbound Slack delivery, read back through Slack’s API. The final browser recheck found the Mac locked and the Slack tab blocked by another extension, so this is not claimed as full UI E2E proof. - This is a review branch. Do not merge until the maintainer reviews it. ## Risks - Incorrect session reuse could mix account or task context. Missing or changed identities continue to reset, and existing authorization checks remain in place. - Warm mode remains opt-in. Remote warm mode requires a reusable sandbox lease. Retained processes keep credentials until they close, so idle expiry and ownership fences are required. - A fresh user message may continue after a committed provider failure. Approval, current authorization, process termination, and prior-result checks remain required. - Projectless sandbox reuse is task- and agent-scoped. Missing or mismatched ownership cannot replace an existing lease; existing workspace-scoped leases keep their scope. Opt-in reuse retains a sandbox per task/agent, so provider auto-stop and deletion policies still determine idle compute and storage costs. Fleet defaults are unchanged. - Compact prompts apply only after proven resume and a matching prior-run delta. Missing or specialized context falls back to full input; fresh sessions always receive the full bootstrap. - No schema or migration changes. ## Model Used OpenAI GPT-6 through Codex, with code editing, tool use, and test execution. The exact serving model ID and context-window size are not exposed by this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [ ] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
45c99a0d06 |
fix(adapters): default legacy harnesses and connected tools to full auto (#13693)
## Thinking Path > - Paperclip lets people manage AI agents and their work. > - Legacy adapters launch provider CLIs and expose connected tools. > - Existing defaults did not consistently grant full automatic permission. > - Remote Claude used a fixed tool list that omitted MCP tools and future tools. > - Direct Codex launches and OpenCode configuration also used narrower defaults. > - This change gives all these paths the same full-auto default as native runners. > - Explicit restrictive settings continue to work. ## Linked Issues or Issue Description Refs #13686. This PR is stacked on that native-runner and task-reassignment PR. Merge #13686 first. Related: #831 (constructed Claude agents), #1935 (adapter-switching permission defaults). ## What Changed - Use actual Claude permission bypass for local and remote runs and probes. Remove the fixed tool list so MCP and future provider tools are included. - Identify actual managed sandbox targets to Claude with `IS_SANDBOX=1`. Do not mark ordinary host execution as a sandbox. - Default direct Codex execution to approval and sandbox bypass, matching agent creation. Preserve explicit false, CLI profiles, sandbox modes, approval policy, and network restrictions. - Set OpenCode's full-auto runtime permission to `allow` for every tool and connection. Preserve the existing explicit opt-out. - Default Gemini probes to the same YOLO mode as execution. Make the legacy ACP `default` alias use `approve-all` for fresh and resumed sessions. - Add default, opt-out, remote, probe, connected-tool, and resume regression tests. Update adapter configuration documentation. - Other adapter paths already request full automatic permission or have no provider approval gate. ## Verification - Full workspace `pnpm -r typecheck` and `pnpm build` passed locally after rebasing onto current master. Targeted adapter/server and legacy ACP tests passed, including defaults, explicit opt-outs, remote launches, connected tools, and fresh/resumed sessions. - Greptile reviewed current head `8ca135eaffcf9cfdba6f1368e896a781a0891d50` at **5/5**. The security reviewer acknowledged the documented full-auto requirement. Acknowledged discussions are resolved. - Current head has **54 passing checks**. [PR checks](https://github.com/paperclipai/paperclip/pull/13693/checks). The process-adapter signoff browser shard passed on one retry after its first attempt exceeded a three-second issue-run wait. - **Six native Claude/Codex real-provider cases passed on their first attempt, with cleanup passing**, against the combined branch: plans, reassignment, and backlog creation/status. [Campaign and downloadable evidence](https://github.com/paperclipai/paperclip/actions/runs/35469926548). This does not claim a real-provider run of every legacy adapter. - The live-tested revision is `a37881c824dcd7170380fc4b788732fc743e5da7`. The current head differs only in the corrected heartbeat test expectation; application code is identical. - The campaign result-enforcement job passed. The separate report publisher failed during frozen dependency installation because the trusted workflow's patched-dependency configuration does not match its lockfile. Passing case evidence remains downloadable from the workflow. - Full-suite coverage comes from CI partitions. The separate unsharded local run was stopped after the corresponding CI partitions passed; it is not counted as a completed local run. ## Risks - Missing permission settings now grant all provider operations, including connected tools. OpenCode full-auto also overrides ambient provider permission rules. An explicit Paperclip permission opt-out preserves restrictive behavior. - Claude refuses full bypass as root outside an identified sandbox. Ordinary host deployments must run Claude as a non-root user. Managed sandbox launches include the required marker. - These defaults do not grant additional Paperclip roles, connections, or company access. Existing controller authorization and governance still apply. - This PR depends on #13686. Retarget it to master after that PR merges. ## Model Used OpenAI Codex, based on GPT-6, with code execution and repository tools. The exact deployment model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
7bc03e0acd |
feat(runner): default harnesses to full auto and support task reassignment (#13686)
## Thinking Path > - Paperclip lets people manage AI agents and their work. > - Agent Chat uses native runners to save plans and coordinate tasks. > - Provider defaults differed across harnesses and could stop unattended work at a second permission gate. > - Agents also lacked a dedicated tool to move existing work to another agent safely. > - This change defaults native providers to full automatic permission for provider tools and connected tools. > - A guarded reassignment tool preserves task identity, stops the previous run, and schedules the new owner once. > - Codex and Claude chat acceptance tests now use production permission defaults. ## Linked Issues or Issue Description **Subsystem affected** Native runner, ACPX Claude permission policy, task authority, and Agent Chat acceptance tests. **Problem or motivation** A user can authorize an agent to save a plan or create a task, but Claude's default provider gate can still stop that action. Reassignment needs a dedicated operation that preserves context and avoids concurrent owners or unintended recovery runs. **Proposed solution** Default Claude/ACPX to `approve-all`, OpenCode to `allow`, and Codex to `never`. Apply the defaults at configuration, execution, fresh-session, resume, driver, and proxy boundaries. Keep explicit permission settings and server-side company, claim, task-mode, and approval checks. Add `reassign_task` with version checks, durable idempotency, audited cancellation, and guarded successor scheduling. **Alternatives considered** A Paperclip-only allowlist still blocks provider tools and other connections during unattended work. Full automatic permission is the requested product default. Recreating a task discards its identity and history. Updating assignment without stopping the previous run can leave two agents working on the same task. **Roadmap alignment** This extends the existing planning, delegated work, governed tool access, and recovery features. It adds no new service or schema migration. Recent related tasks and open PRs were checked for duplicate work. **Additional context** Related: #13678 (Agent Chat tools and recovery), #13677 (remote runner startup). The stacked legacy-adapter companion is #13693. This also fixes the deployed-server artifact fallback needed to stage the current runner binary. ## What Changed - Default Claude/ACPX to `approve-all`, OpenCode to `allow`, and Codex to `never`, including missing settings at direct driver and proxy entry points. These defaults cover provider tools and connected tools. Preserve explicitly configured restrictive modes. - Include assigned approval reads using canonical side-effect classifications, so verifying a recorded approval does not trigger another provider gate. Paperclip approval decisions still enforce controller authority. - Carry the new permission mode through server configuration, execution contracts, recovery identity, TypeScript, and Rust. Keep `approve-paperclip` as an optional restricted mode, with exact SDK rules and closed unknown requests. It is not a default. - Add `reassign_task` to the semantic catalog, controller, mock authority, and generated contracts. - Guard reassignment with company authorization, expected owner and version, protected-state checks, and durable retry receipts. - Honor explicit backlog task creation atomically with the initial plan, without scheduling a wake. Preserve backlog holds regardless of dependency readiness. - Stop active work before changing ownership. Restore the prior owner through a guarded, idempotent wake if final handoff validation fails. Keep intentional reassignment stops out of failure recovery. Preserve backlog and blocked states without waking them early. - Add authorization, concurrency, replay, stop, and permission boundary regressions. Add Codex and Claude chat reassignment cases and run native chat cases with production defaults. - Clarify shared runner guidance: save plans and Paperclip documents directly with `write_document`; create and register a local file only when a downloadable file is requested. - Document provider defaults and the operator choices for existing agents. ## Verification - Current head `d82fbb0f03546d27cecf072250e4172e0b1ee662`: **55 checks passed**, with two intentional skips. [PR checks](https://github.com/paperclipai/paperclip/pull/13686/checks). - Greptile reviewed that exact head at **5/5**. The security reviewer acknowledged the intended full-auto default, and the acknowledged discussions are resolved. - Full workspace `pnpm -r typecheck` and `pnpm build` passed locally after rebasing onto current master. Targeted adapter/server, runner, API, default/resume, and heartbeat configuration tests passed. - **All six real-provider acceptance cases passed on their first attempt, with cleanup passing:** plan handoff, task reassignment, and backlog creation/status, each on native Claude and Codex. Evidence records Claude's effective `approve-all` mode. [Campaign and downloadable evidence](https://github.com/paperclipai/paperclip/actions/runs/35469926548). - The live campaign tested combined revision `a37881c824dcd7170380fc4b788732fc743e5da7`. The final PR heads add only a heartbeat test expectation correction; application code is unchanged from that live-tested revision. - The campaign's result-enforcement job passed. Its separate report publisher failed because the trusted workflow's `patchedDependencies` configuration differs from its frozen lockfile. All six results and screenshots remain available as GitHub artifacts. The overall manual workflow is red for this publishing failure. - Full-suite coverage is supplied by the passing CI partitions. The separate unsharded local run was stopped after the corresponding CI partitions passed; it is not counted as a completed local run. - Reassignment tests cover stale state, cross-company access, denied authority, cancellation failure, compensating wake, and idempotent retries. Backlog tests verify the original creation audit, saved plan, exact task count, and absence of task-bound runs. ## Risks - Agents with no explicit permission mode now receive full provider tool permission, including connected tools. This is a deliberate broad default. Existing explicit restrictive modes still apply. Controller authorization, company isolation, workspace boundaries, and Paperclip governance remain in force. - Reassignment crosses run cancellation and task ownership transactions. Durable stop intent, revalidation, audit receipts, and guarded queue dispatch cover interruptions and retries. - The new permission enum requires a current runner artifact. The remote artifact fallback uses the same resolved controller binary for upload and execution. - Live provider behavior remains subject to the selected model. Targeted live results do not qualify the full catalog. ## Model Used OpenAI Codex, based on GPT-6, with code execution and repository tools. The exact deployment model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Nicky LeachandClaude Opus 5
|
aeef493f4a |
chore(db): keep only the newest 5 drizzle snapshots and stop shipping them (#13687)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - `@paperclipai/db` owns the Drizzle schema and the migration history > - Drizzle writes a full copy of the schema as a snapshot for each generated migration. Each snapshot is now about 1.3 MB. > - The `meta/` folder is about 103 MB. That is about half of each checkout and each worktree. The build also copies it into `dist`, so the published `@paperclipai/db` package is 112.8 MB unpacked. > - `drizzle-kit generate` reads only the newest snapshot. The runtime migrator reads only the `.sql` files and `_journal.json`. > - This pull request keeps the newest 5 snapshots and removes snapshots from `dist`. > - The benefit is a checkout that is about 100 MB smaller, and a published package that is about 1.5 MB instead of 113 MB. ## Linked Issues or Issue Description Refs #11240, #11254, #12333 (earlier snapshot work: diff collapse, binary diffs, drift repair) **What existing behavior does this improve?** The size of the Drizzle migration snapshots in the repository and in the published `@paperclipai/db` package. **Subsystem affected** `packages/db`: migrations and the build. **Current behavior** `packages/db/src/migrations/meta/` holds 141 snapshots (102.8 MB). The size grows faster than the number of migrations, because each snapshot is a full copy of the schema. `build` runs `cp -r src/migrations dist/migrations`. `@paperclipai/db@2026.916.0` contains 147 snapshot files. It is 112.8 MB unpacked and 5.3 MB as a tarball. **Proposed behavior** Keep the newest 5 snapshots. `generate` deletes older snapshots after it runs. `dist` gets only `*.sql` and `meta/_journal.json`. **Reason and benefit** - In drizzle-kit 0.31.10, `generate` sorts `meta/*` and diffs against the last snapshot only (`bin.cjs`, `preparePrevSnapshot`). The [generate docs](https://orm.drizzle.team/docs/drizzle-kit-generate) also say it compares against "the most recent" snapshot. - Gaps in the snapshot history already work. 138 of the 280 migrations never had a snapshot, because they were written by hand before `doc/DATABASE.md` required `generate`. - We keep 5 snapshots instead of 1. This lets a developer undo the latest generated migration, and it keeps the `prevId` chain for recent branches. - Git keeps the history cheaply. The 631 snapshot versions use only 2.2 MB of the pack, because git stores each version as a delta of the previous one. The cost is in the checked-out files, not the clone download. Therefore this change does not use git-lfs and does not rewrite history. Old snapshots stay available with `git show <rev>:<path>`. ## What Changed - `packages/db/package.json`: a new `prune:snapshots` script keeps the newest 5 `*_snapshot.json` files. It is `ls | sort -r | tail -n +6 | xargs rm -f`, which works with the BSD tools on macOS and the GNU tools on Linux. `generate` runs this script after `drizzle-kit generate`. - `packages/db/package.json`: `build` copies only `src/migrations/*.sql` and `meta/_journal.json` into `dist/migrations`. - Deleted 137 older snapshots. `0277`–`0281` remain. - `chat-identity-migration-reconciliation.test.ts`: removed the walk over snapshots `0254`–`0268`. Those files do not change after merge, and the walk would fail after pruning. The journal-order assertions in the same test remain. `migration-snapshot-drift.test.ts` still makes sure that the newest snapshot matches the schema. - `doc/DATABASE.md`: documented the retention rule. - The snapshots were already marked `linguist-generated=true -diff -merge` by `packages/db/.gitattributes` (#11240). No change there. `git check-attr` confirms it. ## Verification - `pnpm --filter @paperclipai/db exec vitest run`: 43 files, 160 tests pass. - `pnpm --filter @paperclipai/db build`: `dist/migrations` contains 280 `.sql` files and `meta/_journal.json`. It is 1.5 MB, compared with about 105 MB before. - `prune:snapshots` was run on macOS (BSD) and in `debian:stable-slim` (GNU findutils 4.10). With 7 fixture snapshots, it keeps the newest 5. When 5 or fewer are present, it deletes nothing and exits 0 on both. ## Risks - Low risk. Runtime migration does not read snapshots. The only commands that read older snapshots are `drizzle-kit check` and `drizzle-kit drop`. No script or CI job calls them, and they still have the newest 5 snapshots. - A branch that is open now can still add its own snapshot. If the branch conflicts, the rule is the same as today: renumber the migration and run `generate` again. - When we upgrade to drizzle-kit v1 (folder per migration), check whether its new cross-branch "commutativity" checks need a longer snapshot history. ## Model Used - Claude Opus 5 (`claude-opus-5`) in Claude Code, with tool use (shell, file edits, web fetch). ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Devin FoleyandPaperclip
|
b70641f23f |
feat(plugins): support image catalogs and persistent application overlays (#13646)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Plugins extend the application without adding each integration to
Core.
> - A downstream image needs a way to supply prebuilt plugins.
> - Some plugin UI must stay mounted as users move between pages.
> - This change adds an image catalog and a persistent application slot.
> - Operators can upgrade or remove these plugins through their image
and configuration.
## Linked Issues or Issue Description
**Subsystem affected**
Plugin packaging, activation and application UI.
**Problem or motivation**
The built-in plugin catalog is fixed in Core source. Downstream images
cannot add entries through an explicit catalog. Existing page slots also
cannot preserve a small application overlay across route changes.
**Proposed solution**
Read a bounded catalog of prebuilt plugins from the image. Verify its
files before importing manifests. Use the existing managed selection and
plugin lifecycle. Add an `appShellOverlay` slot with account and company
cleanup.
**Alternatives considered**
A downstream fork adds merge work. Script injection provides no plugin
lifecycle. A separate runtime download system adds a second distribution
channel.
**Roadmap alignment**
This extends the existing plugin system. Related PR #9006 covers runtime
install replication; this change covers immutable image contents. PR
#12555 covers CLI scaffolding. Neither provides this catalog or
application slot. The maintainer requested this work directly.
## What Changed
- Validate catalog identities, confined paths, package versions and
bundle hashes before importing code.
- Apply image selection to persisted plugin installs, including removal
and rollback. Adopt the verified image path from existing npm/local
installs and bind runtime worker/UI entrypoints to verified package
declarations.
- Mount application overlays in both UI shells. Preserve route state and
clear it on account, company and onboarding changes.
- Restrict service-worker offline storage/fallback to hashed public
assets in a separate cache namespace; exclude application HTML and
extension/API data, including after worker restart.
- Document the packaging contract, trust model and rollback
requirements.
## Verification
- Passed `pnpm -r typecheck`, `pnpm build`, and `pnpm
check:token-gates`. Affected server/UI typechecks and builds, plus token
gates, passed again after rebasing onto current master; the 124 focused
tests also passed after rebase.
- Latest focused verification: 124 tests in nine files passed for
catalog/reconciliation/loader, overlay lifecycle, Layout and
service-worker policy. The broader UI/shared/SDK run passed 7,204 tests
in 690 files with canonical `TMPDIR`.
- Real disposable Core/PostgreSQL: catalog install, selection removal,
0.1.0→0.1.1→0.1.0, same-version npm/legacy-path adoption, and
preservation of disabled status passed. Added permissions entered
`upgrade_pending`, withheld UI across restart, and activated only after
explicit operator enable.
- Real Chromium: desktop/mobile layout, route draft retention and
Escape/focus passed with mocked extension responses. A persistent
browser restart retained public hashed-asset offline fallback while
refusing seeded legacy/current private entries and legacy HTML.
- Full `pnpm test:run`: 12,539 passed; 17 failed across six existing
files, stopping later phases. macOS read-only directory renames fail in
runtime-skill-cache and company-skills-service; email tests require an
absent local AgentMail fixture. Native runner/comment-redaction passed
in isolation after temporary Rust setup; agent-conversations also passed
in isolation. No unrelated source was changed to hide failures.
- After rebase, two unchanged chat timing tests failed in CI and passed
locally in isolation. Their CI shard passed on its single retry. All
other current-head CI jobs passed on the initial run; review is 5/5 with
no unresolved threads.
- No live deployment or external plugin service was used.
## Risks
- Plugins are trusted code. The catalog detects packaging errors; it
does not authenticate an untrusted image builder.
- Invalid catalogs fail startup. Images must contain the catalog and
bundles together, with stable directories.
- A host older than this contract lacks the activation guard. Disable
added plugins and remove their configuration keys before reverting to
it.
- Offline navigation now returns 503 instead of replaying cached
application HTML. Only public build assets have offline fallback.
- Rolling back an unapproved permission change retains the approval
gate; review the current manifest and explicitly enable it. A reduced
permission set cannot establish prior approval or prior enabled status.
- Plugin data migrations need their own rollback policy. This change
retains installed records and does not reverse migrations.
## Model Used
- OpenAI GPT-6 (Codex), model ID `gpt-6`, with repository inspection,
code execution and browser verification. The runtime does not expose an
exact context-window size.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (relevant suites; broad
macOS server-run exceptions are documented above)
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green (fresh run on 488b3754ae; chat
shard passed its single retry)
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(fresh review on
|
||
|
DottaandPaperclip
|
64895f187b |
fix(runner): clarify completion errors and restart test failures (#13682)
## Thinking Path > - Paperclip manages work across persistent agent sessions. > - The runner validates completion calls before accepting their results. > - Generic validation errors can leave the agent unable to repair a rejected call. > - Restart tests also exempted every later failure on an intentionally interrupted run. > - This change gives bounded schema feedback and limits the test exemption to expected interruption outcomes. > - Failures become easier to repair and diagnose without changing authorization or task prompts. ## Linked Issues or Issue Description Refs #13674 and #13676. Related environment and Agent Chat fixes landed in #13677 and #13678. Those changes do not cover these diagnostics. **What happened?** A malformed completion call received a general field list without the failed schema location. The everyday restart test hid later adapter errors on an intentionally interrupted run until its deadline. A clean pnpm install also broke the shutdown test because it resolved an undeclared Playwright package. **Expected behavior** Return enough schema information to repair completion calls without returning submitted values. Fail promptly on an unexpected recovery error. Resolve the declared test package's CLI. **Steps to reproduce** Run the new completion-validation and everyday lifecycle regressions against the parent commit. The new assertions fail there. Run the shutdown test in a clean workspace installation. **Paperclip version or commit** Based on master |
||
|
DottaandPaperclip
|
f589660ec0 |
feat(routines): add safe webhook setup and in-routine run management (#13637)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Routines turn scheduled work and external events into tasks for an assigned agent. > - Webhook setup was disabled, and actor authentication rejected valid webhook bearer keys. > - Operators need to connect and test a sending app before events can start work. > - This pull request adds a guided setup with durable connection tests that cannot dispatch a task. > - It keeps trigger management, execution tasks, and activity within the routine. > - The benefit is a webhook that can be configured, verified, and operated from one place. ## Linked Issues or Issue Description Fixes #11937. Related: #13216 adds provider-specific Sentry support. This PR addresses general routine setup and ingress. #6841 addresses legacy secret bindings; this PR retains the existing secret service. **Current behavior** Webhook creation is disabled. Bearer deliveries can fail in agent authentication before the routine checks its key. Setup has no safe connection test. Runs and Activity send the operator away from the routine. **Proposed behavior** Choose a schedule or a webhook. Follow the setup steps, copy credentials or complete agent instructions, and test delivery without creating work. Finish setup to allow future events to start tasks. Edit or remove compact trigger cards, undo removal, and inspect tasks and activity inside the routine. **Reason and benefit** An operator can verify credentials and delivery before enabling automatic work. Durable setup state survives refreshes and restarts. Retry receipts prevent an old test event from starting work after activation. ## What Changed - Add a production trigger wizard using reusable Slack setup navigation and footer components. - Add schedule and webhook choices, one-time credentials, agent instructions, and live connection feedback. - Persist pending setup, test delivery receipts, connection status, and reversible trigger removal. - Keep setup checks free of routine runs, tasks, and agent wakeups. Preserve delivery idempotency after activation. - Add compact trigger cards, inline editing, key rotation, pause controls, removal, and Undo. - Keep Runs and Activity in the routine. Use the shared task list and compact activity rows. - Permit only exact public delivery POSTs through actor authentication. Retain webhook authentication, JSON-object validation, and log redaction. - Add production-backed Storybook states and focused server, database, and UI coverage. - Document signing modes, setup checks, retries, rotation, HTTPS ingress, and navigation. ## Verification - Full workspace typecheck, build, and token gates passed on the rebased branch. Storybook also builds. - Focused routine, middleware, logging, shared wizard, and UI coverage passes on the rebased branch: 195 tests across 14 files. The migration passed on a fresh PostgreSQL database and on two repeated applications. - Browser testing used the real app, database, and a deterministic process worker through Tailscale HTTPS and the current Cloud proxy code. - Verified rejected keys, safe setup deliveries, persisted state after restart, activation, retry deduplication, key rotation, schedule editing, removal, and Undo. - Fresh bearer and GitHub-signed deliveries created tasks that the worker checked out and completed. Runs and Activity stayed within the routine. - Current Cloud ingress tests passed. Public delivery POSTs passed through without a browser session; management routes remained gated. - All 54 current-head PR checks pass, including general and serialized tests, all eight browser E2E shards, typecheck, build, runner checks, security checks, and the canary dry run. Two optional Storybook jobs are skipped by workflow conditions. - Greptile is 5/5 on commit `7ea63a61e`, with no unresolved review threads. The stale connection-status finding is fixed and covered by a regression test. - No production deployment was performed. ## Risks - Migration 0281 adds three trigger columns and a test-receipt table. It is additive and safe to reapply. Apply it before running the new server. Existing triggers remain live by default. - Requests without delivery IDs are new events after activation. Senders must reuse an event's delivery ID for retries. - Completed webhooks keep normal dispatch behavior. Their management connection check can start work; the UI states this. - Removing a trigger archives it. Undo restores the URL and credentials. Permanent deletion remains available through the existing API. - Public ingress must remain restricted to the delivery POST route. The tenant verifies credentials. Cloud sleeping-stack behavior is unchanged. - Shared setup components also serve Slack. Existing setup contracts and navigation tests cover that integration. - Senders must use application/json with an object. Other media types receive 415. ## Model Used OpenAI Codex, based on GPT-6, with reasoning, repository tools, shell execution, and browser testing. The exact deployment model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
c1f6c3310a |
fix(runner): repair catalog runtime and grading boundaries (#13676)
## Thinking Path > - Paperclip manages tasks across persistent agent sessions. > - The full Runner E2E catalog exposed failures in session restoration, tool validation, and test controls. > - These failures prevented valid work from resuming or made a valid interaction fail the test. > - Invalid completion reports also reached finalization before the provider received useful feedback. > - This pull request repairs those boundaries without changing production prompts or approval policy. > - Focused regressions and fresh paid cases verify each fix. ## Linked Issues or Issue Description Follow-up to #13655. Stacked on the trusted worker prerequisite fix in #13674. **What happened?** Read-only skill uploads failed in resumed Daytona sandboxes. Invalid criterion IDs escaped tool validation. A progress event could park a run before its tool response settled. Partial question forms hid required answers. Two test assumptions rejected valid plan keys or failed to navigate an optional question page. **What did you expect to happen?** Resume identical skill bundles, give repairable feedback for malformed completion calls, preserve in-flight tool responses, show all required questions, and test the rendered workflow accurately. **Steps to reproduce** Inspect the failed cases in https://github.com/paperclipai/paperclip/actions/runs/35417932353. Fresh campaigns: https://github.com/paperclipai/paperclip/actions/runs/35444497313 and https://github.com/paperclipai/paperclip/actions/runs/35445327618. The later backup cleanup is tested in https://github.com/paperclipai/paperclip/actions/runs/35446477285. Combined report: https://pages.paperclip.ing/runner-e2e-operational-35444497313/investigation.html. ## What Changed - Compare immutable archives before reusing read-only Daytona bundles. Reject corrupted content and preserve unrelated files. - Validate exact criterion IDs before accepting completion. OpenCode returns a tool error instead of emitting a result that terminates runnerd. - Complete the activity item for rejected OpenCode calls. - Remove retired read-only harness backups without altering live files or following symlinks. A fresh paid rerun exposed this later checkpoint-cleanup failure. - Exclude progress messages from the governed-wait completion boundary. - Reject newly created question forms that omit questions or contradict their stored answer semantics. Keep historical rows readable. - Navigate all rendered question pages and recognize revision-bound descriptive plan keys in the continuation suite. ## Verification - Harness unit suite: 383 tests pass. Harness typecheck passes. - Native session executor and status corpus: 381 tests pass. - Shared question and interaction-service tests: 42 pass; native question bridge and executor: 360 pass. Daytona sync: 21 pass, including foreign-owner archives and corrupted immutable content. - OpenCode driver: 29 tests pass, including wrong, missing, and duplicate criterion IDs followed by a valid retry. - Repository typecheck and build pass. The later OpenCode activity fix also passes its package build. - The latest commit passes all 52 PR checks and Greptile 5/5. The backup-cleanup fix also passes 351 related local tests and server typecheck. Local full-suite coverage completed across runs. adapter-auth-signal-routes and pipelines-routes encountered transient socket resets; both pass on retry, and all remaining 24 serialized files pass. Paid reruns are complete: 27 of 29 unique cases pass using the latest recording per case. Both Daytona controller-restart cases still fail with runner_state_identity_mismatch; the report describes this remaining runtime issue. Eight affected cells need #13674 on master before their rerun. ## Risks Creation rejects inconsistent dual question representations but does not change historical records. Immutable bundle comparison must verify bytes before skipping extraction. Completion feedback must use the contract bound to the current run. Durable suspension and approval checks remain enforced. Production prompts are unchanged. ## Model Used OpenAI GPT-6 via Codex, with repository inspection, code editing, and test execution. The exact API model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
36dbb7ed1c |
fix: harden agent chat runner tools and recovery (#13678)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agent Chat turns discussion into plans, tasks, reviews, and hires. > - These workflows need reliable tool results and task context on the native runner. > - Live Claude and Codex tests exposed lost retry requests, invalid project inputs, and a child startup crash. > - Recovery also exposed a misleading retry action and missing child task context. > - This pull request fixes those paths and adds regression coverage. > - Agents can continue the original request and operators can inspect a stopped run. ## Linked Issues or Issue Description **What happened?** A failed Agent Chat retry could lose the user's question. Project creation accepted unsupported icons in its tool schema. Codex could stop when a helper's MCP startup event arrived before its thread lineage. A stopped task offered Retry even when the server required execution reconciliation. Resumed agents could miss existing delegated tasks. Hiring and review instructions did not describe the native runner's available tools and source requirements. **Expected behavior** Retries retain the selected request. Tool schemas match the API. Child startup information does not gain authority over the parent or stop it. Recovery actions match the server's requirements. Task context exposes existing child work. Handoffs contain the material the assignee needs. **Steps to reproduce** 1. Enable experimental Agent Chat in an isolated development instance. 2. Configure native Codex and ACPX Claude agents on Paperclip Runner. 3. Ask for a plan, revise it, approve task creation, and request a hire and status report. 4. Retry a failed chat turn and check that it answers the original request. 5. Start a Codex helper before its thread lineage arrives. 6. Resume a delegated task and inspect its existing children and saved output. **Paperclip version or commit** The live failures were found at `f2c5e54dc`. This branch is rebased onto `86b7ee992`. **Deployment mode** Isolated local development instance with native Codex and ACPX Claude. No database migration or default permission change. Related work: Refs #13284 for Agent Chat. Refs #13438 for the server-side API receipt fix, which this branch preserves. The transport also accepts the earlier HTTP receipt format. Refs #13655 for the current Codex continuation and helper lineage handling, which this branch also preserves. ## What Changed - Preserve failed Agent Chat wake-comment IDs and session generation from the authorized source run. Reject pre-reset retries. - Wrap API receipts with the correct semantic call identity. Test current and earlier receipt formats through real HTTP and runnerd. - Classify early child MCP startup notifications as information. Keep foreign completion and result events rejected. - Constrain project icons on both tool surfaces and regenerate the protocol contracts. - Include bounded, company-scoped visible direct child tasks in task context. Filter hidden tasks before applying the limit. - Replace the rejected Retry action with Inspect run for native continuation reconciliation. - Update hiring, review handoff, status reporting, and development guidance. ## Verification - Live tests covered Claude and Codex questions, plan revisions, approval, task creation, hiring, status, chat reset, failures, and recovery. - The recovered task produced its saved checklist and example. A later follow-up read the existing child tasks and document without creating more work. - Full build, repository type checks, token gates, 142 focused tests, 188 runner TypeScript tests, and the Rust notification/descendant regressions passed after rebase. The separate local full-suite run was stopped after the complete CI suite passed. - Review fixes passed the updated route, tool-authority, and icon regression tests plus server type checking. - Required commands: `pnpm build`, `pnpm -r typecheck`, `PAPERCLIP_IN_WORKTREE=false pnpm test:run`, and `pnpm check:token-gates`. - At `4ce8047b0`, all 55 applicable GitHub checks pass (two Storybook checks are intentionally skipped), including the complete general/serialized test matrix, runner tests, browser tests, build, type checks, Docker checks, and canary dry run. - Fresh Greptile review is 5/5 on `4ce8047b0`; all three findings were fixed with regressions and there are no unresolved review threads. - Two initial CI service-startup timeouts passed unchanged in local reproductions and in the latest CI run. ## Risks - The new event classification is limited to MCP startup information. It does not authorize foreign task completion, results, or tool requests. - Task context returns at most 100 direct child tasks and reports truncation. It excludes hidden tasks and other companies. This improves delegation context but does not enforce semantic duplicate detection. - Native reconciliation still requires an operator to inspect and record prior outcomes. The new link does not replace the recovery API. - API tools remain opt-in. Claude permission choices remain explicit. No default permission, schema, or workflow changes. ## Model Used OpenAI GPT-6 in Codex, with reasoning, repository editing, code execution, API tools, and browser testing. The exact deployment identifier and context-window size are not exposed in this session. Live acceptance agents used OpenAI `gpt-5.6-sol` and Anthropic `claude-sonnet-4-6`. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
9335b7db10 |
fix(runner): validate inherited environments and replace stale sandbox binaries (#13677)
Resolve the effective environment for account adoption and adapter tests. Preserve saved-agent overrides when the request omits environmentId, and treat explicit null as inheritance from the instance. Reject sandbox runners that lack unlimited-runtime and connection-lease-renewal capabilities. Stage the bundled runner before launch when the image binary is stale. Add regression coverage for environment precedence, fail-closed validation, adapter switches, API-key reverification, and runner artifact fallback. Document the operational workaround for older controllers. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
|
86b7ee992c |
feat(onboarding): ClipLab sleepy-to-wake hero and step hand-offs (#13629)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents have a persistent visual identity (#13171): a ClipLab character in one of 17 palettes, rendered as cached PNGs in lists and as a live character in larger placements. > - The onboarding wizard is where a person meets that identity first, and it showed a stock ClipLab expression on the previous engine while the rest of the app would show a different character on a newer one. > - The wizard's steps also cut from one screen to the next, so the arc read as separate pages rather than one walk. > - This pull request puts one character on one engine everywhere, gives the wizard's hero the studio's sleepy → wink → idle sequence on Review, and hands the steps over inside one presence. > - The benefit is that what wakes on Review is exactly what the agent looks like on the dashboard afterwards, and the walk to it reads as one screen changing. ## Linked Issues or Issue Description Refs #13171, now merged into master. This PR contains the onboarding and ClipLab update on top of that foundation. Original feature work by @tonio-alucema; merge preparation preserves the original commits. **Problem or motivation** The onboarding hero and the app's avatars were two different characters on two different ClipLab engines. Steps 1 → 4 of the wizard cut between screens, and the wizard mounted cold when a cloud-managed workspace arrived from Cloud's naming screen. **Proposed solution** Vendor ClipLab v0.2.0 as the shared engine and render one studio-exported character from it in every palette, for every pose and size. Play the export's one-shot wake on Review with the palette fading in over the gray dormant loop. Hand steps over inside one presence so the footer slides instead of jumping, and play the arrival half of that hand-off when the wizard opens directly on the agent step. **Alternatives considered** Exporting mp4/webm loops per size: no cursor following, no clean alpha, and the palette "colour in" is a runtime blend. Minting a `cap-v2` character version: nothing had shipped `cap-v1`, so the artwork is regenerated in place instead of migrated. Keeping the separately vendored runtime bundle for the hero: two engines and two characters in one app. ## What Changed - `packages/shared/src/cliplab`: re-vendored from ClipLab v0.2.0 (`987b6db0`) with the Paperclip adaptations replayed (optional graphics backend for the Node SVG snapshot path, supersampled live textures, character framing, deterministic SVG id prefixes); new upstream `particles.ts`. - `packages/shared/src/cliplab/character.ts`: the studio export, mirrored from `ui/src/assets/cliplab/onboarding.character.json` by `scripts/sync-cliplab-character.mjs` (drift caught by `check:token-gates`). `characterDefinition` builds every palette from it; the resting portrait is its idle beat. - `OnboardingCharacter`: gray `sleepy` loop through the agent and connect steps; on Review the one-shot sleepy → wink → idle plays on two lock-step canvases while the palette fades in, then the `idle` loop. Body-follows the pointer, page-scoped. 160px in the wizard. - `OnboardingWizard`: steps 1 → 2 → 3 → 4 hand over inside one `AnimatePresence` (departing content fades and gives its room back; arriving content opens its room then fills); the hero has a room that opens on the walk into the agent step; opening directly on the agent step plays the arrival half; the self-hosted naming step uses the arc's label and field. - Motion vocabulary in `onboarding-motion.ts` (`stepContentMotion`, `ledeMotion`, `heroRoomMotion`, `heroRoomArrival`, `titleSwapMotion`). - Storybook: `Onboarding / Character` (Wake Up), `Onboarding / Agent arc` walkable from the naming step plus `Arrive From Cloud`; the companies fixture answers the wizard's create call with a company. - Uses the shared runtime for onboarding; `doc/agent-personas.md` documents the shared character. - Releases both onboarding canvases after partial startup or transition failure. Registers each canvas before seeking so synchronous render errors can release it. Six component tests cover these failures and palette changes before or during wake. - Refreshes both sleeping canvases when the palette changes, including a palette change in the same render as wake. - Moves choreography values into the CSS token layer and preserves the shared motion catalog drift check across the imported stylesheet. - Repairs the static Storybook avatar route and uses accessible heading names/current button labels in the wizard play functions. - Closes the lazy avatar worker pool during application shutdown. ## Verification - Merge-preparation checks: `pnpm -r typecheck`, `pnpm build`, `pnpm build-storybook`, and `pnpm check:token-gates` pass. The final UI typecheck and 123 focused onboarding, lifecycle, and token catalog tests pass. All 55 checks on final head `b4f5e201a1564083d163abc6f93f5b3da06ccefd` pass, including the full sharded test suite, runner verification, and all eight browser shards ([CI run](https://github.com/paperclipai/paperclip/actions/runs/35445430535)). The duplicate monolithic local `pnpm test:run` was stopped after CI completed; it is not claimed as a separate completed local run. - Chromium walkthrough: palette change, wake, return to sleep, WebGL failure fallback, Review step hand-offs and cloud arrival pass with normal and reduced motion; no browser errors. The signoff happy-path browser test also passes against a disposable instance. - The final CI run confirms the catalog fix and a passing signoff browser shard. The earlier signoff failure was a heartbeat-run availability timeout; the focused local reproduction and final CI passed without signoff code changes. - Original author verification: - `pnpm check:token-gates` (includes the new character sync check); shared, server avatar/persona (17) and UI onboarding/persona (137) suites pass; `pnpm build-storybook` packages all 3,564 avatar PNGs through the worker pipeline. - Storybook: `Agents / Personas` Sizes, Expressions and Palettes render the studio character at every size and pose; `Onboarding / Character → Wake Up` plays the wake on the shared engine; `Onboarding / Agent arc` walks 1 → 4 with the hand-offs, and `Arrive From Cloud` plays the arrival (measured: content room 6 → 65px over 320ms, fade to 1.0 by ~560ms, footer travel continuous). - The original author walked the agent → connect → review flow and wake after a real sign-in on staging. - Not done here: the Linux Storybook visual baselines (`tests/storybook-visual/agent-personas.spec.ts`) need re-baselining for the new engine, hero size and naming-step changes. ## Risks - Every avatar's pixels change (new engine, new character) under the unchanged `cap-v1` name. Stacks that rendered avatars on the previous engine keep those PNGs in their cache (`generated-agent-avatars/cap-v1/...`, served immutable) until cleared; only the two pinned staging stacks ever did. - The one-shot handoff to the idle loop is timed from the sequence's authored duration (the engine reports completion by continuing into idle itself); presentation only, nothing in the wizard's state waits on it. - Reduced motion skips the wake and the hand-offs; jsdom is treated the same way, so the wizard tests see the next step's content immediately. - The committed export differs from the studio by one animation (Loop off, leading idle step removed); a re-export without that fix would play a 5.6s idle before the wake. ## Model Used Original feature: Anthropic Claude Fable 5.1 (`claude-fable-5-1`) in Claude Code, with shell, browser, and file tools. The original context window was not recorded. Merge preparation and lifecycle regression fixes: OpenAI GPT-6 in Codex, with reasoning, shell execution, file editing, GitHub CLI, and automated tests. The session does not expose an exact runtime model ID or context-window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Dotta <bippadotta@protonmail.com> Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com> |
||
|
|
1ef3b08714 |
feat(ui): integrate agent personas across the app (#13171)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - A stable agent persona is useful only when the same identity appears across the app. > - Lists, task messages, selectors, and activity feeds need inexpensive static avatars. > - Onboarding and agent headers need a larger character with expressions and pointer tracking. > - This pull request connects the persona foundation to those existing views and preserves onboarding draft assignments. > - Full-page stories and Linux checks make the placements and performance contract reviewable. ## Linked Issues or Issue Description **Problem or motivation** Agents need a stable visual identity in lists, tasks, onboarding, and configuration. External tools also need an image URL for that identity. **Proposed solution** Assign each agent a permanent palette from a fixed ClipLab character library. Store the assignment on the agent. Render and cache preset PNG URLs on demand. Use static images in dense views and one animated character in larger placements. **Alternatives considered** A generated image bundle requires a separate asset build. A live renderer in every avatar adds unnecessary work in large lists. Arbitrary uploaded images do not provide the requested shared character system. **Roadmap alignment** This improves agent identity across existing control-plane views. It preserves agent permissions, company boundaries, and status labels. ROADMAP.md has no separate ClipLab persona milestone. Related approaches: #2422 adds configurable image URLs and DiceBear generation; #5578 adds optional uploaded avatars. This work uses a fixed, versioned character library and preset URLs. ## What Changed - Replace agent icons with static persona images across lists, the sidebar, org charts, tasks, comments, selectors, activity, and dashboard views. - Put one animated character in the agent header. Let it follow the pointer across the page, with reduced-motion and touch fallbacks. - Add larger padded characters to agent creation. Keep the palette stable across draft refreshes and connection retries, then reveal it after success. - Pass appearance through shared projections rather than fetching each agent separately. - Add real full-page Storybook examples for the agent list, overview, task, dashboard, new-agent dialog, and connection page. - Add Linux screenshot, clipping, density, and 500-avatar performance checks. ## Verification - `pnpm -r typecheck`, `pnpm build`, and token gates pass on the rebased tree. Persona lifecycle tests pass. - The rebased feature passes 38 Linux screenshot/performance checks, including both display densities, corner pointer positions, and the no-WebGL/no-live-download contract for 500 avatars. - The final Linux persona suite passes all 38 visual, lifecycle, density, and full-page checks using the standard Storybook configuration and real on-demand avatar endpoint. - Final local focused verification: 45 avatar/native-recovery tests pass; UI identity/routine tests, typecheck/build, token gates, and Storybook build pass. - Current-head CI passes: full workspace/server tests, all serialized server groups, typecheck/release checks, build, canary validation, and end-to-end shards. The build passed after retrying a native-runner concurrency-test failure; its three targeted cases also pass locally. - Manual inspection covered stable identities in the app, header placement, full-page mouse tracking, onboarding size, and task/dashboard placements. ### Screenshots Linux captures use synthetic Storybook fixtures. Full-page captures use reduced motion. The live character, mouse tracking, and disposal are checked separately. <details> <summary>Agent overview with the character in its header</summary> <img src="https://raw.githubusercontent.com/paperclipai/paperclip/8c68f42b268ada22b67d79d0fe1bb0a2f84ec25c/screenshots/full-page-agent-overview.png" width="900" alt="Agent overview with the character in its header" /> </details> <details> <summary>Task messages and assignee identity</summary> <img src="https://raw.githubusercontent.com/paperclipai/paperclip/8c68f42b268ada22b67d79d0fe1bb0a2f84ec25c/screenshots/full-page-task.png" width="900" alt="Task messages and assignee identity" /> </details> <details> <summary>Larger onboarding character with room for expressions</summary> <img src="https://raw.githubusercontent.com/paperclipai/paperclip/8c68f42b268ada22b67d79d0fe1bb0a2f84ec25c/screenshots/full-page-meet-your-next-agent.png" width="900" alt="Larger onboarding character with room for expressions" /> </details> <details> <summary>Dashboard agent activity</summary> <img src="https://raw.githubusercontent.com/paperclipai/paperclip/8c68f42b268ada22b67d79d0fe1bb0a2f84ec25c/screenshots/full-page-company-dashboard.png" width="900" alt="Dashboard agent activity" /> </details> ## Risks - This PR depends on #13170, the persona foundation. Merge the foundation first, then retarget this PR to master. - Many placements change from icons to character silhouettes. Human avatars and authoritative agent status labels retain their existing behavior. - Only one character can render live per view. Reduced motion, hidden/offscreen content, touch input, and renderer failures use the defined fallbacks. - The full-page stories use fixture data. They do not contact a real company or complete real provider sign-in. ## Model Used OpenAI Codex, GPT-6 family. The exact model identifier and context window are not exposed in this session. Used code editing, shell execution, browser inspection, and Linux visual testing. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: Tonio <tonework@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com> |
||
|
DottaandPaperclip
|
43acbcc398 |
fix(runner): preserve sessions and complete question and approval continuations (#13655)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The native runner connects task state to provider sessions. > - Follow-up turns must retain provider memory and carry new user direction. > - Lost session IDs caused repeated context and extra input tokens. > - Native question answers and approval races could leave valid work blocked. > - This pull request repairs those paths and adds regression coverage. > - Agents can continue accepted work without repeating the conversation or losing the user's answer. ## Linked Issues or Issue Description Refs #13574. That merged PR shortened continuation prompts and moved question instructions into tool documentation. This change preserves sessions and fixes failures exposed by broader testing. Related runtime work: #13408 and #13410. **What happened?** Native follow-up turns could lose the provider session ID. Completion guidance could replace the original task with its latest comment. Claude native questions could remain pending after the user answered. Approval during a running tool call could suspend the run before the tool response arrived. Onboarding and chat handoff instructions also caused repeated planning or missing plan documents. **Expected behavior** Reuse a valid provider session. Send only new events when that session already has the history. Preserve the task requirements and apply later user direction. Store the question answer and deliver it to the waiting run. Finish governed tool responses before suspending. Execute the accepted plan without asking for the same approval again. **Steps to reproduce** Run the continuation, local-session-integrity, first-task, and agent-chat suites with native Codex and Claude. Include provider-question-bridge, accept-while-running, and plan-handoff. **Paperclip version or commit** This branch is based on master |
||
|
Devin FoleyandPaperclip
|
d54b750111 |
Preserve Claude ACP quota classification and reset time (#13651)
Typed Claude ACP quota failures lost their recovery classification and reset time when the runtime reduced provider metadata to a generic category error. Inspect terminal metadata in memory and retain only safe recovery labels and a parsed reset timestamp. Preserve the existing handling of other limits. Verified real child processes on both pinned ACPX runtimes, adapter and server recovery regressions, all PR CI gates, and Greptile 5/5. Also isolate a pre-existing chat regression from unrelated fixtures’ retry work. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
Devin FoleyandPaperclip
|
685d4faba3 |
Fix PostgreSQL recovery after a transaction connection closes (#13643)
Reject queued and late work from disconnected transaction and reservation scopes. Keep closed reservations out of the open pool, and clear old connection buffers and responses so new requests can reconnect safely. Twelve real-PostgreSQL regression cases cover crash prevention, recovery, and transaction isolation in both ESM and CommonJS. Database checks and all PR CI checks pass. Greptile: 5/5, no unresolved comments. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
924f07be8c |
feat(chat): simplify Slack onboarding and account linking (#13638)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Chat connections let people start and continue that work from Slack. > - Setup mixed app creation, credentials, URL verification, account linking, and testing on the same screens. > - People also needed a safe way to link their own Slack identity after the first operator finished setup. > - This pull request gives each step a clear place and keeps membership approval separate from identity linking. > - It also makes connection details easier to use and fixes misleading callback health behind HTTPS proxies. ## Linked Issues or Issue Description **Subsystem affected** Cross-cutting: chat routes and services, shared contracts, and the Apps board UI. **Problem or motivation** Slack onboarding made users find settings without enough guidance. A second user needed operator help to link their account. Activity stopped at 100 records, and TLS termination could mark working callbacks as stale. **Proposed solution** Use six setup steps with editable app names, a generated manifest, credential guidance, URL verification, account linking, and an optional message test. Send each Slack user a private, expiring confirmation link. Require company membership or an approved access request before linking. Add cursor pagination and tolerate the internal HTTP hop in callback diagnostics. **Roadmap alignment** This improves the existing connected-app surface and supports CEO Chat without changing the task-and-comments model. The maintainer requested and reviewed the flow during a live Slack test drive. **Additional context** Related work: #7, #3349, #13000, and #13620. Those cover broader chat capabilities, older webhook paths, or plugins. This PR improves the existing native connector's setup and account-linking flow. HTTPS documentation was published separately in paperclipai/paperclip-docs#128. ## What Changed - Split Slack onboarding into six clickable sidebar steps. Keep secondary and primary actions on one row. - Generate the Slack creation link and read-only manifest from editable app, bot, and command names. Add credential prefix validation and direct instructions. - Add live account-link status and an optional mention-based message test. - Add private, single-use Slack account invitations and membership access requests. Retain cloud authentication/bootstrap checks and enforce the chat rollout flag in all identity APIs. Default new Slack connections to linked users only. - Put Settings, Access, Conversations, and Activity in the sidebar. Simplify conversation rows and remove active header badges. - Add 25-item activity pages, stable timestamp/ID cursors, and replay safety across pages. Preserve the legacy array API for clients without pagination parameters. - Fix false callback warnings when HTTPS terminates at a proxy. Keep host, port, and path drift detection. - Document the setup flow, pagination, callback diagnostics, and shared wizard footer rule. ## Verification - Passed: `pnpm -r typecheck`, `pnpm build`, and `pnpm check:token-gates`. - Passed: focused Slack callback and pagination integration tests; UI clipboard, wizard, pagination, and activity tests; OpenAPI route tests. The final access-gate fix also passes 27 focused tests covering cloud authentication/bootstrap, nonmember invitations, token validity, and the server-enforced rollout flag. - Passed: all 1,002 chat integration tests, 6,356 UI tests, and all 11 provider browser scenarios (including mobile light/dark navigation). After rebase, the identity route, sidebar, and 25 clipboard tests pass. - The full local `pnpm test:run` was attempted. The first run found 14 Slack fixtures that needed explicit guest access; those are fixed and the complete chat suite passes. Unrelated embedded PostgreSQL startup/resource failures and timeouts prevented a clean full local run. All CI checks pass on `2d858b036`, including the full chat, server, workspace, build, typecheck, and browser suites. - Live test drive: Slack app creation, credential setup, URL verification, private account confirmation, mention messages, and thread replies. Verified the callback warning clears for the existing proxied connection. - Review: create a Slack connection, follow the six steps, link a second user's account, and browse older activity with Next and Previous. ## Risks - Identity invitations carry a temporary capability. Tokens are hashed, expire after 15 minutes, work once, and require explicit confirmation by a company member. Access requests do not grant membership. - New Slack connections reject unlinked people by default. Existing connection settings remain intact. - Activity is a live ledger. Updated action rows can move forward in time. Older pages do not poll. - Proxy tolerance affects health display only. Slack signature checks and proxy authentication settings remain unchanged. - No database migration or package-lock changes. ## Model Used OpenAI Codex, based on GPT-6, with reasoning, repository tools, code execution, and browser verification. The runtime does not expose an exact model build ID or context-window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (targeted suites; full local-run limitations documented above) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
352153b5ed |
fix: run the cargo-building native-runner CI suite in the Rust-cached vitest lane (#13586)
## Thinking Path Post-merge of #13557, the slowest check on the freshest fully-green PR run ([35246999382](https://github.com/paperclipai/paperclip/actions/runs/35246999382)) was `ci / General tests (server (1/12))` at **339s**. The cause is one suite: `server/src/services/native-runtime/native-codex-runner.integration.test.ts` runs 1 test in **277s of a 291s vitest step (95%)** because its `beforeAll` cargo-builds the Runner release binaries, and the general-server shards carry no Rust cache — every PR run cold-compiles the full third-party crate graph. The other 19 suites in that shard finish in under 70ms each. The obvious fix (a dedicated Rust-cached matrix lane) requires editing workflow files, which the available GitHub App credentials cannot push (`workflows` permission). But the `Verify Paperclip Runner` lanes **already restore the shared `release-runner-v1` Rust cache read-only**, and their commands are `pnpm --filter @paperclipai/paperclip-runner <package script>` — so the suite can move into a Rust-cached lane purely through script changes. ## What Changed - `scripts/run-vitest-stable.mjs`: new `general-server-native-runner` group carrying exactly that suite. Under the PR workflow (`GITHUB_WORKFLOW == "PR"`, inherited from `pr.yml` by the reusable `pr-trusted.yml`) the without-chat server shards exclude it and rebalance to ~211s of tests each. Every other caller — local runs, `release-verify.yml` under the Release / Cloud readiness workflows — keeps the suite in the shards, so a renamed or unknown workflow degrades to today's slower-but-covered behavior instead of dropping coverage. - `packages/paperclip-runner`: `test:typescript:vitest` now routes through `scripts/run-pr-vitest-lane.mjs` — the identical `ensure:eval-build-deps && build:rust && vitest run` chain (shard flags passed through), plus the native-runner group on the **final PR shard only** (`--shard=N/M` with `N == M`, i.e. today's `vitest 2/2`, the 122s lane). With the restored cache the suite's cargo build becomes an incremental rebuild. - `scripts/__tests__/run-vitest-stable-shard.test.mjs`: guards pin the whole contract — PR 12-shard coverage (shards + chat + native-runner = full server group exactly), Release/local 10-shard runs keep the suite, `pr.yml` is named `PR`, the vitest lanes partition with exactly one final shard, the package-script wiring, and the wrapper's shard/workflow gating via its `--dry-run` plan output. No workflow files change. `.github/workflows/*` are untouched. ## Verification - `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs scripts/__tests__/release-verify-workflow.test.mjs`: **36/36 pass** locally on this branch (includes the new coverage, wiring, and wrapper-gating guards). Both files run in CI's `Test general-server shard partition` / `Test release verify workflow wiring` steps. - Wrapper `--dry-run` plan matrix verified for all six shard/workflow combinations plus malformed-shard rejection (pinned as a guard test). - The executing proof is this PR's own CI: `ci / Verify Paperclip Runner (vitest 2/2)` must go green while running the native-runner suite (its log will show the `general-server-native-runner` group after the package vitest shard), and the 12 `ci / General tests (server (x/12))` shards must go green without it. ## Risks - The exclusion keys on `GITHUB_WORKFLOW == "PR"`. Failure mode of a rename is safe (suite falls back into the server shards, slower but covered) and the guard test on `pr.yml`'s name makes it loud. - `vitest 2/2` grows from ~122s to an expected ~210–260s — still well under the ~306s `vitest 1/2` and ~326s e2e shards, and inside the 20-minute lane timeout. If the cache misses (key drift), the lane pays a cold compile like the server shard does today; a miss is slow, never wrong. - Double-run/coverage-loss combinations are enumerated in the wrapper header and pinned by tests: each caller runs the suite exactly once. ## Model Used Claude (Bender agent, Paperclip) — Fable 5. --- Expected savings once merged: the 339s `server (1/12)` check drops to ~265s-equivalent shard levels (~211s of tests), the slowest `ci /` check becomes the ~326s e2e shard (~13–33s off PR wall time), and every PR run stops paying ~4.5 min of billed cold Rust compile. For the merger (squash): please keep the trailer below in the squash body to preserve authorship. `Co-Authored-By: Bender (Fable) <Paperclip-Paperclip@users.noreply.github.com>` ## Related PRs Searched the GitHub PR list for prior work on this surface — related groundwork, none duplicate this change: - #13457 — restored master's Rust dependency cache on the PR runner lane (the read-only cache this PR relies on) - #13500 — made that cache key image-toolchain-independent so GitHub-hosted PR runners actually hit it - #13521 — rebalanced PR shards and split the Verify Paperclip Runner lanes this PR extends - #13557 — previous health-check iteration (split the runnerd transport suite); this PR targets the next slowest check ## Checklist - [x] I have searched GitHub for duplicate or related PRs and linked them above Co-authored-by: Bender (Fable) <Paperclip-Paperclip@users.noreply.github.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com> |
||
|
Devin FoleyandPaperclip
|
5442f2d869 |
fix: repair managed Git launchers in sandbox projects (#13588)
## Thinking Path > - Paperclip runs agents in local and remote execution environments. > - Managed GitHub launchers select credentials for each Git operation. > - Remote launchers are written inside the project checkout as extensionless CommonJS scripts. > - An ES module project makes Node interpret those launchers as ESM, so they crash before credential resolution. > - When the launcher can start, empty identity variables also override valid repository and command-line Git configuration. > - This change gives the launchers their own CommonJS scope and clears empty identity overrides while preserving managed credential isolation. ## Linked Issues or Issue Description **What happened?** In a repository with `"type": "module"`, the managed `git` and `gh` launchers fail immediately with `ReferenceError: require is not defined in ES module scope`. The launchers use CommonJS but inherited the enclosing project's module type. Sandbox agents also report empty `GIT_AUTHOR_NAME` and `GIT_COMMITTER_NAME` variables and try to unset them for each command. With no managed identity available, the Git launcher recreated those empty values. `git commit` failed with `fatal: empty ident name`, even with explicit `user.name` and `user.email` configuration. **Expected behavior** Managed `git` and `gh` start in both ES module and CommonJS projects. Local commits with an explicitly configured identity work without manual environment cleanup. Managed credentials and captured identity continue to take precedence. Missing identity does not silently select the host user's details. **Steps to reproduce** 1. Create a sandbox project whose `package.json` contains `"type": "module"`. 2. Stage the managed GitHub launchers and run `git --version` or `gh --version`. Before this fix, the launcher fails at its first `require()`. 3. In a CommonJS project with no available managed identity, configure repository `user.name` and `user.email`, or supply them with `git -c`. 4. Run `git commit --allow-empty -m test`. Before this fix, both identity configuration forms fail with empty identity. **Paperclip version or commit** Reproduced from master commit `165b10bd9`. **Deployment mode** Sandbox execution. The shared launcher is also used for managed local and SSH execution. Related work: #13094 introduced the local-operation fallback; #13053 changes launcher discovery on Windows. Neither fixes empty identity overrides. Related identity work in #8945 and #8946 configures worktree authorship and does not remove these environment overrides. ## What Changed - Stage `package.json` with `"type": "commonjs"` in the launcher directory before the Node scripts. Keep the project's package configuration unchanged. - Leave inherited author and committer variables unset in the real Git process. When credentials are absent, require explicit Git identity configuration with `user.useConfigOnly`. - Clear empty identity merge overrides in staged shell profiles after environment merging. Preserve nonempty captured identity values. - Exercise real Git commits with repository and command-line identity, broker failures, and managed-user switching. Verify startup in ES module and CommonJS projects, shell cleanup, and captured identity preservation. - Document launcher module scope and local identity behavior in the execution GitHub identity contract. ## Verification - Confirmed both new local-commit regression cases fail before the fix with `fatal: empty ident name`. - Confirmed the new ES module project regression fails before the fix with `require is not defined in ES module scope`. - Focused launcher and shell tests: 28 passed. - `pnpm exec vitest run --project @paperclipai/adapter-utils --exclude '**/dist/**'`: 1,216 passed, 11 skipped across 58 files. - `pnpm --filter @paperclipai/adapter-utils typecheck` and `pnpm --filter @paperclipai/adapter-utils build`: passed. - `pnpm -r typecheck` and `pnpm build`: attempted; both stop in the unchanged native runner because Cargo is not installed on this machine. - Full `pnpm test:run`: started locally; stopped the duplicate run after the complete CI suite passed. No local full-suite success is claimed. - CI on `99ea8050e`: all 53 checks passed (2 skipped), including full tests, typecheck, build, native runner checks, and browser checks. - Greptile reviewed `99ea8050e`: 5/5 with no findings or unresolved comments. GitHub reports no merge conflicts with master. - No live sandbox or GitHub push probe performed. ## Risks - The new package scope is confined to the run-specific launcher directory. It does not change the project's module type, launcher names, or credential selection. - Without a managed identity, an explicitly configured repository author can now create local commits. GitHub access remains subject to the existing credential broker. Global/system Git configuration, ambient credentials, and SSH identity remain isolated. - Managed identity still wins over repository settings. Missing local identity still fails instead of guessing host details. - New or resumed executions must stage the updated launcher and shell profiles. Existing processes retain their prior files and environment until refreshed. No database migration or sandbox image rebuild is required. - Revert this change to restore the prior behavior. ## Model Used - OpenAI GPT-6 via Codex, with code inspection, implementation, and local test execution. The hosted model variant and context window were not exposed. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub references) - [x] My branch name describes the change and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (the affected adapter-utils package) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
84fe89906d |
fix: complete native agent review handoffs (#13581)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Native execution uses durable runs, issue locks, wake requests, and typed tool authority > - A child can finish with a native agent review request while its original assignee stays responsible for the work > - The reviewer then needs a bounded execution path that can inspect the child, record one decision, and finish safely > - Before this change, assignee-only gates rejected the reviewer or left the parent waiting after the child review ended > - This pull request adds typed reviewer admission, scoped reviewer tools, durable wake and recovery handling, and parent continuation evidence > - The benefit is that native review handoffs complete without changing child ownership or granting broad mutation access ## Linked Issues or Issue Description Refs: #13314 Refs: #13574 **What happened?** A native child run could report `needs_review` for an agent reviewer. The reviewer wake then failed assignee and execution-lock checks. The child remained in review and the parent remained waiting. **Expected behavior** The named reviewer should receive one durable wake. The reviewer should inspect the child and resolve the exact review card. The child assignee should stay unchanged. The parent should receive the recorded review outcome after the child reaches its terminal state. **Steps to reproduce** 1. Run a native task with a different named agent reviewer. 2. Keep the child assigned to its original worker. 3. Let the worker finish with a native completion review request. 4. Start the durable reviewer wake. 5. Resolve the review and finish the reviewer run. 6. Observe the child and parent state. **Paperclip version or commit** Base: `e926b1301`. PR head: `b31ad9ab8`. Live reviewer verification source: `eea171aae`. **Deployment mode** Built from source. **Installation method** Built from source (pnpm build). **Agent adapter(s) involved** Not adapter-specific (core bug). **Access context** Both. **Database mode** Embedded PostgreSQL in the isolated live test fixtures. ## What Changed - Add server-validated native review assignment facts. - Admit only the exact company, issue, source run, decision, revision, addressee, and resolver policy. - Give reviewer runs a narrow set of Paperclip read and resolve tools. File and shell access follow the configured agent and environment policy, so reviewers can run tests. - Separate server-owned reviewer instructions from untrusted persisted review data. Escape the data boundary; retain server-enforced authorization. - Keep the child assignee unchanged. Atomically claim the reviewer run, wake request, and issue execution lock. A competing lock prevents provider startup. - Require the exact running reviewer session and current issue lock to resolve its assigned card. Reject missing, unrelated, or terminal reviewer runs. - Add durable reviewer wake, lock, stale-card, and abandoned-run recovery handling. - Prevent duplicate native wake dispatches during deferred admission and recovery. - Carry accepted or rejected child review outcomes into parent task context and continuation evidence. - Add focused server, runner, and native protocol coverage. - Preserve upstream continuation rules. Add child review decisions as separate evidence, while keeping real human answers in their own field. - Return actionable completion validation feedback to both providers. Permit a corrected completion after rejection. Keep strict terminal acknowledgment validation. - Apply exclusive shared-workspace locks to sandbox environments. Local and SSH folders can run concurrently, including when old settings request serialization. - Repair test timing, native event parsing, and the review artifact assertion. Allow a valid reject, correct, and accept review sequence. Check the accepted card against its reviewer run and decision. Keep polling within the existing deadline when review acceptance precedes the parent wake projection; report a specific missing-continuation error at timeout. - Apply the ACPX pending-call limit to reserved finish/block calls, with capacity-release and cancellation tests. ## Verification - `pnpm build`: passed on `eea171aae`. - `pnpm -r typecheck`: passed on `eea171aae`. - `pnpm test:e2e:runner:unit`: 359 tests passed in 30 files on `b31ad9ab8`; runner E2E typecheck also passed. - `pnpm check:token-gates`: passed. - Focused DB review, reviewer authority, and prompt-boundary checks: 31 tests passed. They cover invalid reviewer runs, competing locks, atomic admission, duplicate claims, and valid resolution. - Heartbeat, workspace, and recovery checks: 30 tests passed. - ACPX sidecar suite: 27 tests passed. Moving the capacity guard back below reserved handling makes both new regression cases fail. - Four focused live continuation checks passed on their first attempt at `f15f55e0a`: answer updates scope (6/6 each on Codex and Claude) and question tool guidance (12/12 each). These cases do not use the reviewer prompt path changed afterward. - Fresh Codex and Claude review-handoff checks passed all 29 native checks each on their first attempt at `eea171aae`. Both runs received the expected fixed prompt and completed cleanup. Only the six selected live flows were tested; no full paid provider catalog run. - The final commit only extracts the existing test-harness timeout diagnostic into a shared helper and adds positive and negative coverage. Removing the accepted-review guard makes two regression assertions fail; restoring it passes all six timeout tests. Production runtime code, prompts, deadlines, and grading criteria are unchanged by this final commit. - Deadline regressions: a valid continuation delayed 20 seconds succeeds within its 30-second unit-test deadline; an absent wake returns a specific candidate-failure diagnostic at that same deadline. Both assertions failed before the fix. Production E2E deadlines remain unchanged. - Historical native failures remain recorded: Docker availability failures; a valid reject/correct/accept sequence that the first-card grader misread; and a test that rejected the gap between accepted child review and parent wake projection. No failed result was regraded. The latest tests use a protected reference to the pinned Docker image and the unchanged artifact oracle and time limits. - Full repository verification runs in GitHub CI. Local verification uses the focused suites above, full build, and full typecheck. An unchanged Codex shutdown timing test failed once in CI, passed in isolation, and its full shard passed on the final commit without changes to that test or its causal code path. The original failure is retained in the verification record. Greptile reviewed `b31ad9ab8` at 5/5 with no outstanding actionable findings. All review threads are resolved. All current-head CI gates passed, including the isolated native runner Docker build (55 successful checks; two skipped by the workflow). ## Risks - Reviewer admission depends on exact persisted decision and interaction bindings. A stale or changed card is rejected. - Paperclip control-plane tools are limited to inspection and review resolution. This is not a filesystem permission boundary; provider file and shell access retain the configured policy. - Deferred wake recovery changes dispatch receipt coalescing. A scheduler regression could delay a continuation if the receipt state is wrong. - Parent review outcomes are evidence for the model. They do not grant tool authority or change issue ownership. - This change does not address legacy lease-hold handoff behavior. > Roadmap review: native execution, review gates, and durable recovery are existing roadmap capabilities. This PR completes a narrow reliability path for those capabilities. ## Model Used OpenAI `gpt-6-astra` with reasoning, tool use, and code execution. OpenAI `gpt-5.6-luna` assisted with bounded implementation, review, and journal work. Context window size is not exposed by this session. Live test subjects use `gpt-5.6-sol` and `claude-sonnet-5`; they are not the PR authors. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Nicky LeachandPaperclip
|
e926b13017 |
fix: keep sandbox termination progressing after bridge loss (#13287)
## Thinking Path > - Paperclip must stop remote execution after losing its controller. > - A bridge can remain blocked while the sandbox still incurs costs or performs actions. > - Waiting forever for that bridge prevents provider termination. > - A temporary provider outage can also exhaust cleanup attempts permanently. > - This pull request bounds bridge drain and persists cleanup retries with backoff. > - Cleanup ends only after provider confirmation, without a user accepting uncertain side effects. ## Linked Issues or Issue Description Builds on merged #13285. Related #13254 added exact provider termination receipts; merged #13272 adds explicit user retry. Merged #13352 stops active sandbox startup before waiting for setup. This PR preserves that immediate cancellation path and extends bounded teardown to ordinary release and destroy. Cleanup continues automatically after repeated provider failures. Refs #12953 for provider failures blocking execution. **What happened?** Daytona release waits for in-flight bridge activity before stop/delete. A dead bridge can prevent that wait from finishing. The host also stops cleanup after five failed attempts. **Expected behavior** Provider termination proceeds after a bounded bridge drain. Cleanup retries survive service restarts and provider outages. **Steps to reproduce** Start a sandbox command whose bridge promise never resolves, then release its lease. Separately, persist a pending-cleanup lease with five failed attempts and recover the provider. **Deployment mode** Hosted Paperclip with a Daytona provider; rebased onto master at `728f7185f` on September 14. ## What Changed - Bound bridge drain and provider lifecycle calls. Prefer stop for reusable sandboxes, with delete fallback. - Persist cleanup attempt identity, renewable in-flight deadline, and cooldown. Fence completion writes against superseded attempts. - Preserve scoped explicit Retry and its activity log. Explicit Retry can skip cooldown, but cannot take over a live cleanup attempt. - Continue cleanup after five failures with slower retries and an operator warning. - Exclude leases in cooldown before paging so they do not starve due work. - Add hung-bridge, restart, provider-recovery, and concurrent-cleanup regressions. ## Verification - Rebased onto master at `728f7185f`. The outstanding diff contains only cleanup changes; the merged controller-ownership prerequisite is excluded. - Daytona plugin suite: 160 passed, including immediate startup cancellation, graceful release, hung activity, and teardown regressions. - `pnpm exec vitest run server/src/__tests__/heartbeat-pending-cleanup-sweep.test.ts`: 31 passed. Two added integration cases verify explicit Retry during cooldown and while another cleanup owns the lease. They also verify run scoping and the activity log. - Targeted cleanup and cancellation cases in `environment-runtime.test.ts`: 20 passed. - Earlier live disposable Daytona test: provider stop ended background work, resume preserved files without restarting the old process, a new command succeeded, and the sandbox was deleted. This verifies provider behavior; it was not repeated for this rebase. - Latest-head CI and automated review are pending. Broad local tests, typecheck, and build were not rerun for this focused rebase; CI supplies those checks. ## Risks - Timing out bridge drain permits provider termination; it never supplies a stop receipt. - A crashed cleanup attempt remains protected for 15 minutes, then becomes eligible again. Repeated failures retry every 30 minutes after escalation. - The existing counter saturates at the escalation threshold; the new attempt identity and deadline prevent overlapping claims. - No schema, UI, telemetry, lockfile, or workflow change. ## Model Used OpenAI GPT-6 through Codex, using reasoning, repository inspection, code execution, and test tools. The precise backend revision and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
mouse-value-add
|
fcdb3f2499 |
feat: add optional you.com search integration (#13555)
<!-- Simplified Technical English (ASD-STE100). --> ## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Agents that do research work need live information from the web > - Paperclip reaches external systems through governed, catalog-based MCP connections > - The Apps catalog is data-driven: a researched provider with a hosted remote MCP server becomes a connectable app with no runtime code change > - You.com operates a hosted remote MCP server for web search, content extraction, and research tools > - The server supports OAuth 2.1 with dynamic client registration, an API key in a bearer header, and a keyless free profile at a separate endpoint > - This pull request adds You.com to the self-serve MCP research ledger and generates its catalog entry with three connection methods: browser sign-in, API key, and the keyless free profile > - The benefit is that an operator can give agents live web search through the normal connection governance, and the free profile needs no account at all ## Linked Issues or Issue Description No public issue exists for this provider. The problem description follows the new-adapter issue template. **Agent or provider** You.com — web search and research tools over a hosted remote MCP server. **Why this adapter is useful** Agents that do research, monitoring, or fact-finding tasks need current web results. You.com exposes web search (`you-search`), live page extraction (`you-contents`), citation-backed research (`you-research`), and finance research (`you-finance`) as MCP tools. Any Paperclip company can connect it in a few clicks. The free profile offers `you-search` without an account, so a new company can try agent web search at zero cost and zero setup. **How the agent is invoked** Hosted remote MCP server (Streamable HTTP) at `https://api.you.com/mcp`. Three supported access paths, verified against the live server on 2026-09-16: - OAuth 2.1 browser sign-in. The server returns a `WWW-Authenticate` challenge with RFC 9728 protected-resource metadata and advertises a dynamic client registration endpoint, so Paperclip's automatic DCR path applies. - API key. Sent as an `Authorization: Bearer` header per the provider's official server manifest and docs. Keys come from you.com/platform and unlock higher rate limits plus the full tool set. - Keyless free profile at `https://api.you.com/mcp?profile=free`. Provides a reduced, read-only tool set. Official docs: https://you.com/docs/build-with-agents/mcp-server **Are you willing to implement it?** Yes. Implemented in this pull request. **Additional context** Research evidence collected 2026-09-16, from live protocol probes and official provider sources only: - Unauthenticated `POST https://api.you.com/mcp` returns HTTP 401 with `WWW-Authenticate: Bearer resource_metadata="https://api.you.com/mcp/.well-known/oauth-protected-resource" scope="Tools offline_access"`. - RFC 9728 metadata lists one authorization server with scopes `Tools` and `offline_access`. - The authorization-server metadata (RFC 8414) publishes authorization, token, and revocation endpoints, and advertises a `registration_endpoint`, so DCR is available. No registration was performed during research, per the runbook's non-registering preflight rule. - The keyless free profile answers `initialize` (server `You.com`, version `4.0.1`), lists the tools `you-search` and `you-discover`, and executed both tools successfully during the probe. - The API-key placement matches the provider's official `server.json` in the youdotcom-oss/mcp repository: header `Authorization`, value `Bearer <key>`. ## What Changed - Added You.com (slug `youcom`, wave 4, risk tier S2) to the self-serve MCP research ledger in `packages/shared/src/self-serve-mcp-research.json`, and refreshed the ledger verification date. - Added the You.com category (`ai`) and API-key header spec to `scripts/ingest-app-definitions.mjs`. - Added a You.com case to `specialMethodsFor` that emits three methods: browser sign-in (`mcp-oauth`, DCR), API key (`mcp-api-key`, bearer header), and keyless free profile (`mcp-free`, no auth). - Regenerated `packages/shared/src/app-definitions/youcom.json` and the generated registry via the ingestion script (`--definitions-only` mode; no unrelated provider churn). - Added the official You.com wordmark artwork (light and dark theme variants, taken from the provider's docs site) under `ui/public/brands/apps/`, with a manifest entry. - Updated `packages/shared/src/app-definitions.test.ts`: ledger counts (47 providers, 44 candidates), store count (48), verification date, and assertions for the three You.com methods and their endpoints. ## Verification - `node scripts/ingest-app-definitions.mjs --definitions-only` — passed. Generated the new definition and registry import only; no other provider JSON changed. - `node scripts/check-app-brand-assets.mjs` — passed (71 identities). - `node --test scripts/app-brand-validation.test.mjs` — passed. - `pnpm exec vitest run packages/shared/src/app-definitions.test.ts ui/src/lib/app-brand-assets.test.ts ui/src/pages/apps/AppLogo.brand-assets.test.tsx` — passed (39 tests). - `pnpm exec vitest run packages/shared/src/app-definitions.test.ts server/src/__tests__/tool-access-service.test.ts server/src/__tests__/generic-mcp-connection.test.ts server/src/__tests__/tool-connection-removal.test.ts ui/src/pages/apps/AppsConnect.test.tsx ui/src/pages/apps/Browse.test.tsx` — passed (181 tests). Two server suites that require embedded Postgres skipped on this machine by their own environment gate; the gate is unrelated to this change. - `pnpm --filter @paperclipai/shared typecheck` — passed. - `pnpm --filter @paperclipai/ui typecheck` — passed. - `pnpm --filter @paperclipai/plugin-sdk ensure-build-deps` — passed; builds `@paperclipai/shared` with the new definition. - `pnpm test:run` (full Vitest suite) — 7,930 passed, 18 failed, 4,592 skipped. Every failure is environmental on this container: the embedded-Postgres suites refuse to start because the machine runs as root, the native runtime suites need the Rust runner binary that this container cannot build, and one media suite needs a native HEIC binary. No failure touches the app-catalog, connection, branding, or shared-package surface; those suites pass locally. CI is the authoritative gate for the full suite. - `pnpm --filter @paperclipai/server typecheck` — not completed: the script's `prepare:runner-vendor` prelude builds the Rust runner, which cannot build on this container. A direct `tsc --noEmit` reports only pre-existing errors from the missing vendored runner types; no error touches this change. No server code is changed. - Live You.com proof on 2026-09-16 (keyless free profile, real network calls): preflight 401 challenge with RFC 9728/8414 metadata and DCR endpoint ✓, `initialize` ✓, `tools/list` ✓, `you-search` call returned results ✓, `you-discover` call returned results ✓. - Live proof NOT run: an authenticated OAuth connect and an API-key call against the full server. This environment has no You.com account or API key. Per the runbook, this proof stays outstanding and must not be assumed from the keyless probe. Both paths match the reviewed `mcp_remote` patterns (DCR and bearer header) used by existing providers. - Browser e2e suites not run: opt-in per `AGENTS.md`, and this change adds catalog data only, with no UI code. ## Risks - Low risk. The change is catalog data plus generated output. It adds no runtime code and touches no existing provider. - The free-profile method is a fixed keyless endpoint. If You.com changes or removes `?profile=free`, that method breaks and the entry needs a ledger update. The OAuth and API-key methods do not depend on it. - The authenticated tool catalog is discovered live at connect time, so provider-side tool changes appear through the normal catalog refresh and quarantine flow, not through this definition. - Rollback is a single revert; no migration and no state are involved. ## Model Used - Provider: Zhipu AI, via OpenRouter - Model: GLM-5.3 (`z-ai/glm-5.3`) - Context window: 200K tokens - Capabilities used: tool use (shell, file edits, live HTTP probes), long-context repository reading - The change was produced with AI assistance and reviewed by a human before submission. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes: #` / `Refs: #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge |
||
|
Devin FoleyandBender
|
ec40bd8bf6 |
ci(runner): split retained-settlement suite out of runnerd-codex-transport.test.ts (#13557)
Moves the 19-case 'settles only retained control authority' family (~156s) plus its two private helpers from runnerd-codex-transport.test.ts (8,895 lines, 398s sequential) into a new runnerd-codex-transport-settlement.test.ts so vitest can schedule the two files onto separate workers. Pure code move, no test-logic changes; vitest collects the identical 182 test names. Measured on the PR's own CI run: 'ci / Verify Paperclip Runner (vitest 1/2)' dropped from 531s to 340s and the end-to-end PR workflow from ~545s to ~415s. Co-Authored-By: Bender (Fable) <Paperclip-Paperclip@users.noreply.github.com> |
||
|
DottaandPaperclip
|
e26d787928 |
Shorten continuation prompts and verify question tool guidance (#13574)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents must continue tasks using user answers without losing earlier requirements or approval gates. > - The wake prompt mixed human decisions with prior tool evidence and repeated detailed question instructions. > - Those instructions belong with the question tool, with a short routing hint in the wake. > - The Runner evals need to prove that answers, approvals, and completed work survive later turns. > - This PR shortens the prompts, separates authenticated answers, and adds continuation tests with useful screenshots. ## Linked Issues or Issue Description Refs #13517. This is a follow-up to the merged onboarding skill and Runner E2E work. Related #13539 covers responses received while a run is active; this PR preserves its cases and adds continuation coverage. Existing continuation/recovery and question PRs were searched; none covers this prompt/documentation and eval change. **What existing behavior does this improve?** The instructions sent when an agent continues a task, the native human-input tool documentation, and the evidence captured by Runner full-stack E2E. **Current behavior** The wake repeats a long question-tool guide. Human answers appear alongside untrusted prior results. Screenshot capture can finish at DOM load while the task still shows a spinner, even when backend behavior checks pass. **Proposed behavior** Keep earlier requirements unless the user changes them. Treat clarification as distinct from approval. Give authenticated human responses a scoped field. Keep tool and agent results as evidence. Put detailed question behavior in the tool descriptor and retain one routing sentence in the native wake. Wait for the correct task and loaded conversation before taking screenshots. **Reason and benefit** Reduce repeated prompt text and make authority boundaries clear. Test that real question cards, later answers, approval gates, and completed child tasks still work. Make screenshots useful for human review. ## What Changed - Shorten shared continuation instructions for legacy and native runners. Separate authenticated user responses from tool results and agent summaries. - Remove the detailed question guide from native wake prompts. Keep its behavior in the canonical `request_human_input` descriptor and existing payload schema. Regenerate semantic contracts and fixture hashes. - Add five continuation cases across four local profiles. Add a dedicated choice-then-text case for native Codex and native Claude. All 22 cells join the shared full E2E campaign. - Cover revised scope, clarification without approval, hostile instructions in a handoff file, and reuse of a completed child after restart. Keep production instructions and fixed user facts. - Capture continuation screenshots only when the intended task and conversation have rendered. Add provider-free browser regressions for loaders and wrong-task capture. - Preserve current master’s extra tool and onboarding cases. The default campaign now contains 166 cells; 35 manual everyday cells remain separate. ## Verification - `pnpm -r typecheck`: passed after replay on current master. - `pnpm test:e2e:runner:unit`: 340 passed. Harness typecheck passed. - `PAPERCLIP_PLAYWRIGHT_CHANNEL=chrome pnpm test:e2e:runner:browser-support`: 4 passed. These tests failed against immediate screenshot capture and passed after the fix. - Focused continuation and native-input tests: 36 passed locally. The tool-authority suite could not initialize embedded PostgreSQL locally, including one isolated retry; its 17 assertions did not run locally. The full remote server shards passed on this PR commit. - `pnpm build`: passed after replay on current master. `pnpm test:run` was attempted locally but hit the same embedded PostgreSQL initialization failure; the remaining local run was stopped after complete remote CI passed. This is not claimed as a full local test pass. - [Full PR CI](https://github.com/paperclipai/paperclip/actions/runs/35232755685): passed on `6a22128c14f4552d0613a6d9a25955db4a1ed02f`. All server/chat/workspace/serialized shards, browser shards, Runner checks, typecheck, build, canary and policy checks passed. The isolated native Runner build and security checks also passed: 57 successful checks, with two expected Storybook skips. - Greptile reviewed the exact PR head at 5/5, with no findings or unresolved review threads. The PR has no merge conflicts. - [Live question-docs report](https://pages.paperclip.ing/runner-e2e-question-docs-35227647794/): 3/3 passed at source `83dd132f2` before replay on master. Native Codex and Claude each asked a choice, waited, asked a text question, and saved both answers. Claude also passed a completed-child restart case. All three native turns are checked for absence of the old question block. - [Earlier continuation report](https://pages.paperclip.ing/runner-e2e-continuation-35154943615/): all five continuation cases passed on native Claude. The report retains campaign and revision provenance and separately shows two unresolved onboarding behavior failures. - [Before/after prompt report](https://pages.paperclip.ing/runner-prompt-comparison-20260917/): full text, current recorded Claude inputs, and reproducible reference-token counts. The controlled wake comparison removes 401 reference tokens; the net counted input reduction is 339 after charging the larger tool description. These are text-size estimates, not measured billing savings. ## Risks - Prompt wording affects model behavior. Live results cover the stated cases, not every provider or conversation. Legacy profiles are registered but were not rerun for this change. - The optional continuation field changes prompt data only; there is no database migration or new production API. - Authenticated answer projection excludes generated summaries and agent-resolved interactions. It preserves the answer’s question or approval scope. - The screenshot guard can expose UI loading failures that earlier runs hid. Backend grading alone no longer makes those captures valid. - The two prior onboarding failures remain separate product issues: work before acceptance and a missing saved plan. This PR does not claim the entire onboarding suite passes. ## Model Used OpenAI Codex, GPT-6, with reasoning, repository tools, code execution, and browser verification. The exact deployed model identifier and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass — targeted tests above; the full local database-startup limit is documented - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
327ab2fe38 |
fix(grok-local): do not pin empty GROK_HOME over host login (#13570)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Local adapters such as `grok_local` invoke a host CLI (`grok`) for each heartbeat > - Grok authenticates from `GROK_HOME/auth.json` when that env var is set, otherwise from `~/.grok` > - Recent work (#12469, #12618) isolated subscription credentials into a company-scoped Grok home filled only by sandbox device login > - Local_trusted instances have only a Local environment, so that login never runs, the company home stays empty, and execute still sets `GROK_HOME` to it > - This pull request stops pinning `GROK_HOME` on local subscription runs unless the company home already has usable auth, a managed AI connection supplied a home, or the run is remote/sandbox > - The benefit is that `grok login` on the host works again for local Grok agents, without leaking host credentials into sandboxes ## Linked Issues or Issue Description Fixes: #13568 Related PRs (predecessors, not duplicates): - Refs #12469 - Refs #12618 - Refs #12696 I searched GitHub for `GROK_HOME`, `grok login`, `not signed in`, and `device-code`. No existing PR restores host-login fallback for local `grok_local` runs. ## What Changed - Local subscription execute no longer sets `GROK_HOME` when the company Grok home has no usable `auth.json` - Remote/sandbox runs and managed AI connections still pin `GROK_HOME` so they cannot fall through to the host login - Local runs still pin `GROK_HOME` once a company home has a usable credential (completed device login) - Adapter configuration notes document the host-login vs company-home split - Subscription detection respects an explicit empty `XAI_API_KEY` that clears an inherited host key. This keeps a valid company login selected. - Tests cover a real child process reading fixture host credentials, custom host homes, malformed company credentials, API-key overrides, managed connections, and empty remote homes. - Original fix by @hawikk. The follow-up preserves the contributor commit and adds independent regression coverage. ## Verification - `pnpm exec vitest run packages/adapters/grok-local`: 131 tests pass in 12 files. - `pnpm --filter @paperclipai/adapter-grok-local typecheck`: passes. - The new subprocess host-login regression fails against `master` and passes with this fix. It uses disposable fixture credentials and makes no provider request. - The explicit-empty-key regression fails against the contributed commit and passes with the follow-up. - `pnpm -r typecheck` and `pnpm build`: pass locally. - `pnpm test:run`: attempted locally, then stopped after embedded PostgreSQL startup failures. A focused retry of `ai-legacy-compatibility.test.ts` reproduced the same startup failure after five attempts. - All CI checks pass on `f4a380fec`: 54 successful checks and two intentional Storybook skips. This includes all general tests, serialized server suites, runner checks, browser shards, typecheck, build, and canary packaging. - Greptile reviewed `f4a380fec` at 5/5 with no actionable findings. GitHub reports no merge conflicts. - Grok CLI 1.0.13 is installed on the verification host, but it has no signed-in account. A live authenticated inference run was not performed. ## Risks - Low. Behavior changes only local subscription runs whose company Grok home has no usable `auth.json`. - Remote/sandbox isolation is unchanged: those runs still pin `GROK_HOME` and never use host `~/.grok`. - Managed AI connections still pin even with an empty home (fail closed rather than using the host account). - Operators who previously copied `auth.json` into the company home keep the pinned-home path. > For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and discuss it in `#dev` before opening the PR. Feature PRs that overlap with planned core work may need to be redirected — check the roadmap first. See `CONTRIBUTING.md`. ## Model Used - Provider: xAI Grok - Model: Grok 4.6 (`grok-4.6`) - Tool use: yes (repository search, local tests, GitHub issue/PR) - Human-authored: no — AI-assisted implementation - Follow-up review, code, and tests: OpenAI GPT-6 in Codex, with reasoning, repository tools, and code execution. Exact backend model ID and context window are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: Dotta <bippadotta@protonmail.com> |
||
|
Devin FoleyandPaperclip
|
fae6980310 |
revert(apps): restore Google connector visibility (#13552)
## Thinking Path > - Paperclip helps people manage AI agents for work. > - The Connectors catalog lists services that agents can use. > - PR #13551 temporarily hid Google connectors. > - We now want to restore their catalog visibility. > - This PR reverts that change and restores the previous catalog behavior. ## Linked Issues or Issue Description Refs: #13551 Revert the temporary removal of Google connectors from the UI. ## What Changed - Restore Gmail and eight Google Workspace entries to the catalog. - Restore the matching branding flags and original catalog and service tests. - Remove the temporary-hiding documentation note. This is an exact revert of commit `cf1e873ab24277d55ffd3ab06074f77014dc4015`. ## Verification - Passed: 507 catalog, UI, and connection service tests. - Passed: `pnpm check:token-gates` and `node scripts/check-app-brand-assets.mjs`. - Passed: `pnpm --filter @paperclipai/ui... build` and `pnpm --filter @paperclipai/ui... typecheck`. - Full local build and typecheck stop at the Rust runner because `cargo` is not installed. - Full local Vitest was not repeated because the unchanged base has confirmed macOS skill-cache permission failures. The full CI suites passed. - Passed: all GitHub CI gates; Greptile 5/5 on commit `4e3dddef0ebfef1f99001e7735822ed4cba852ab`, with no review threads. - Reviewer check: open Connectors and confirm that Gmail and Google Workspace entries appear again. ## Risks Low risk. This restores the previous catalog visibility and setup entry points. Connector implementations and saved connection data are retained. ## Model Used OpenAI GPT-6 through Codex, with reasoning, tool use, and code execution. The exact deployment ID and context window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Devin FoleyandPaperclip
|
cf1e873ab2 |
fix(apps): temporarily hide Google connectors (#13551)
## Thinking Path > - Paperclip helps people manage AI agents for work. > - The Connectors catalog lists services that agents can use. > - We need to temporarily remove Google connectors from the UI. > - The catalog already separates visibility from retained definitions. > - This PR uses that setting so Google can return with a small change. ## Linked Issues or Issue Description **What existing behavior does this improve?** The Connectors catalog and its setup entry points. **Current behavior** The catalog shows Gmail and eight Google Workspace connectors. **Proposed behavior** Temporarily hide those nine entries. Keep their definitions and existing connections. **Reason and benefit** Make the temporary UI removal easy to reverse. **Breaking changes** Fresh catalog setup no longer offers Google. Saved connections keep the existing management and reconnect paths. ## What Changed - Add the nine Google connector slugs to the existing hidden list. - Match the branding manifest visibility flags. - Update existing catalog and service tests. Keep backend Google connection coverage and document how to restore visibility. ## Verification - Passed: 507 targeted tests covering catalog definitions, URL matching, setup routing, connector UI, branding, and the connection service. - Passed: `pnpm --filter @paperclipai/ui... build` and `pnpm --filter @paperclipai/ui... typecheck`. - Passed: `pnpm check:token-gates` and `node scripts/check-app-brand-assets.mjs`. - Full local build and typecheck stop at the Rust runner because `cargo` is not installed. - Stopped the full local Vitest run after skill-cache permission failures. Three failures in `company-skills-service.test.ts` also reproduce on the unchanged base branch. The final connector service suite passes all 319 tests. - Greptile: 5/5 on the current commit, with no open review threads. CI is retrying one unrelated preview-server readiness timeout. That test file passes all seven tests locally. - Reviewer check: open Connectors in a company with no Google connections. Gmail and Google Workspace entries should be absent. Existing saved connections remain manageable. ## Risks Low risk. This uses the existing catalog visibility mechanism. No connector implementation, credential, or database schema is removed. Restoring visibility requires updating both the hidden list and branding manifest. ## Model Used OpenAI GPT-6 through Codex, with reasoning, tool use, and code execution. The exact deployment ID and context window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Devin FoleyandPaperclip
|
6fe8e30625 |
feat(apps): add Railway connection and governed deployment tools (#13415)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Apps gives agents governed access to external resources. > - Operators need to inspect Railway services, read logs, deploy code, and run container commands. > - Railway offers hosted MCP with OAuth, but broad remote actions hide their internal operations. > - This PR adds a branded connection and fixed direct operations through the existing gateway. > - Separate SSH keys enable container commands under the same grants and policies. > - Operators can require approval for an action and inspect the resulting audit record. ## Linked Issues or Issue Description **Subsystem affected** Apps catalog, connection setup, gateway execution, and connection documentation. **Problem or motivation** Agents need Railway access through Paperclip. Operators need to grant and revoke that access, inspect available actions, and govern deployment and container operations without giving agents provider credentials. **Proposed solution** Reuse hosted MCP OAuth, vault storage, catalog discovery, grants, and the gateway. Probe the actual credential before enabling fixed GraphQL operations. Use a dedicated grant-owned SSH key for bounded container commands. **Alternatives considered** A catalog entry alone cannot execute the missing operations. The hosted general agent has opaque internal effects. An unrestricted CLI runtime can bypass action policy and inherit ambient credentials. **Roadmap alignment** This extends the existing MCP Tool Gateway & Apps path and the Connected Apps direction in ROADMAP.md. It does not add a plugin or parallel connection service. Related PRs #311, #939, and #7861 concern hosting Paperclip on Railway. They do not add this outbound Apps connection. The separate shared agent-picker fix is #13414 and is not included here. ## What Changed - Add the generated Railway catalog entry, official marks, provenance, and OAuth setup guidance. - Add fixed service/deployment status, bounded logs, and redeploy/restart/rollback tools. Block source deployment until the provider can atomically bind the approved repository and commit. - Verify API access with an explicit workspace before exposing direct tools. - Add grant-owned SSH key setup and a bounded runner with host verification, target checks, isolated state, and cleanup. - Block the opaque hosted railway-agent and accept-deploy actions. Preserve normal Allowed defaults and Ask-first policies for other actions. - Quarantine new or changed Railway schemas after initial discovery, including reconnect. - Add provider, lifecycle, gateway, SSH, UI, and browser fixtures. Document setup, limitations, and the release checklist. ## Verification - Security follow-up: removed the unsafe source-deployment mutation. Direct calls and old active catalog entries are denied before any upstream request, including normalized aliases. Refresh marks retired entries disabled. All 386 focused Railway, catalog and gateway tests passed, and server TypeScript checking passed. Full [GitHub CI](https://github.com/paperclipai/paperclip/actions/runs/35139421144) passed on |
||
|
DottaandPaperclip
|
d0b67bfe71 |
feat: queue approvals and answers during active runs (#13539)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Users guide running agents through messages, questions, and approval cards. > - Messages already wait in a queue when an agent is running. > - Card responses did not appear in that queue. Some question answers also steered a later run without a user click. > - A fast approval could invalidate the agent's review handoff and cause it to stop its own run. > - This pull request gives card responses the same queue controls and preserves the exact response during delivery. > - Users can wait for completion or explicitly send the response with Interrupt or Steer. ## Linked Issues or Issue Description Refs #13517, which is merged. This PR targets master and adds queued interaction responses on top of the onboarding changes. Related continuation work: #10519 and #12866. **What happened?** Accepting a proposal while its source run was active left a saved response outside the message queue. The agent could then lose its review path, reassign the task, and cancel itself. Answers to older questions could also steer another active turn without a click. **Expected behavior** Save the response immediately. Queue its continuation behind the active run. Deliver it after completion, or when the user explicitly chooses Interrupt or Steer. Preserve approval revisions and answer choices. **Steps to reproduce** 1. Let an agent publish a confirmation card while its run is still active. 2. Accept the card before the agent finishes its review handoff. 3. Inspect the message queue and the task's next run. **Paperclip version or commit** Reproduced on da8a3876c with the onboarding changes from #13517. **Deployment mode** Local development from source. The fix covers legacy adapters and native Runner turns. ## What Changed - Project resolved cards into the existing queue as immutable responses. Keep answers and exact approval revisions. - Require an explicit click to steer a response into a compatible native turn. Use Interrupt when a fresh session is required. - Preserve typed response context through interruption, cleanup waits, and normal queue promotion. Keep the direct answer channel for a provider blocked on its original question request. - Accept the source run's review handoff after its card resolves. Reject stale agent reassignment that would orphan a queued response. - Add deterministic regression tests and an `accept-while-running` case to the first-task suite. Require recorded timestamp overlap before that case can pass. - Keep the first-task skill name out of user-facing messages. ## Verification - Red-green: the original route failed the queue regression; the changed route passes it. - Focused server/UI tests: 139 passed, including 64 queue-route tests. - Runner harness unit tests: 314 passed. - Server, UI, and Runner E2E typechecks passed. UI token gates passed. - Full repository typecheck and build passed. Server typecheck passed again after review fixes. - Review regressions: 165 queue/reopen route tests, 53 wake admission tests, and 18 run identity tests passed. Approval acknowledgement recovery and both message/approval arrival orders are covered. - Full local test run: 12,401 passed; three new admission regressions ran against a cached pre-fix module. A fresh run of that entire suite passed (53 tests). The complete CI suite passed on the final commit. - Previous-head CI at `c28e2ef12`: 32 checks passed and 2 optional Storybook checks skipped. Every server/workspace/browser shard, Runner verification, build, typecheck/release registry, canary, policy, and security check passed. Greptile: 5/5, no unresolved threads. Earlier interrupted CI workers were replaced by this fresh complete run. - After integrating the updated parent: 314 harness tests, 119 queue/admission tests, 44 onboarding/question-delivery tests, and 13 native recovery tests passed locally. Full repository typecheck and build passed. - Clarified the skill wording preference: routine replies describe the action without announcing the internal skill; direct questions and permission/security/execution disclosures remain truthful. - The paid `accept-while-running` scenario is registered for all four local first-task profiles. It has not been run against a model in this change. - Rebased onto the merged parent at `11921075a`; the resulting tree exactly matches the locally verified integration tree. Final-head CI on `b53054807` passed: 54 successful checks, 2 optional Storybook checks skipped, no failed checks. Every new server/browser shard, aggregate verify/e2e gate, Runner, typecheck, build, canary, and security check passed on the first attempt. Greptile reviewed this exact head at 5/5 with no unresolved threads. ## Risks - Responses now wait instead of implicitly steering another active turn. A provider blocked on the original question still receives its answer directly. - Approval receipts cannot be edited, discarded, or reordered as comments. This preserves the recorded decision. - Interruption must still prove that the prior execution stopped. The tests cover cleanup waits and duplicate delivery. - The new paid overlap case can be unexercised if the model finishes before the click lands. It cannot pass without evidence of overlap. - No database migration is required. This repairs the existing approvals and execution controls; it does not implement the roadmap's work-stream queues. ## Model Used OpenAI GPT-6 through Codex. The exact deployed model ID and context-window size were not exposed in this session. Capabilities used: agentic reasoning, repository inspection, code editing, terminal commands, and test execution. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
11921075a4 |
Add first-task onboarding skill and Runner E2E coverage (#13517)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The first task helps a new user define and approve useful work. > - That workflow needs reusable instructions and tests against the production experience. > - Native Codex and Claude must load the assigned skill, including after resume. > - Maintainers need recorded conversations and precise failed checks to judge regressions. > - This pull request adds the first-task skill and a suite in the shared Runner E2E harness. > - It keeps behavior results separate from informational quality scores and incomplete recordings. ## Linked Issues or Issue Description **What existing behavior does this improve?** The first onboarding task and the Runner E2E report used to review it. **Current behavior** Onboarding embeds its policy in a hidden brief. Native Codex drops the skill-instructions setting at the Rust boundary. The shared E2E harness has no onboarding suite or full conversation view. **Proposed behavior** Assign and invoke `/first-task` for the onboarding task. Send selected Codex skills as structured protocol inputs. Run twelve scenarios across legacy Codex, legacy Claude, native Codex, and native ACPX Claude. Include all 48 cells in full campaigns. Show recorded chat, question and approval cards, exact checks, instructions, and billing in the shared dashboard. **Reason and benefit** Measure the real onboarding experience before changing prompts. Distinguish infrastructure failures, behavior failures, and unexercised journey steps. **Breaking changes** No database migration or production API change. First-task instructions now live in an assigned skill. The user-edited persona is preserved; the skill includes the maintainer-approved proposal-mode mapping and saved-plan requirement. Related: #11043 is earlier onboarding work. #13422 already fixes native Claude model pinning, context delivery, and read permissions on master; this branch includes those fixes through its base. The new Claude recovery test supplements them. ## What Changed - Extract and assign the first-task skill while retaining the production greeting and opening question. - Carry the Codex skill-instructions flag through thread start and resume. Resolve explicit task skill references only against assigned skills and send native skill inputs. - Invoke an unambiguously selected assigned skill through Claude ACPX’s native slash-command parser on initial and resumed turns, retaining the entire task/wake envelope as its argument. Do not carry that invocation into ordinary tasks. - Restore the saved single-task proposal modes: confirmation card, or saved plan with revision-targeted checkbox approval. Explicit plan requests also require a saved plan. - Add first-response and complete-journey cases with fixed user facts, acceptance checkpoints, durable outcome checks, and accounting for child runs. - Fail the eval when choice questions have fewer than two real options. Recognize planning documents without treating them as completed work. - Add optional, bounded quality judging as explicit post-processing. - Render full conversations and static interaction cards in the shared report. Conversations start folded. Show original and regraded results and incomplete journeys distinctly. - Keep credential-persistence scanning outside the first-task behavioral suite; retain public evidence redaction. - Refresh generated capability references after the API-reference edits. - Correct shared native question guidance and tool schemas: choices need at least two meaningful options; open-ended questions use canonical text fields with the required compatibility payload. Verify both formats through real tool-authority persistence. - Disable announcements automatically for every isolated Runner E2E process and label the gallery environment/provider/target explicitly. - Remove CI races in the GitHub connection browser test and native session recovery test by waiting for the actual async work before asserting its results. ## Verification - `pnpm exec vitest run server/src/services/onboarding-first-task-assets.test.ts server/src/__tests__/issue-onboarding-first-task-routes.test.ts`: 19 passed. - `pnpm --dir packages/paperclip-runner exec vitest run src/drivers/acpx/runtime-host.test.ts src/drivers/acpx/native-skill-prompt.test.ts src/cli/acpx-runtime-sidecar.test.ts`: 70 passed. Native command forwarding and the 1 MiB input boundary both failed before their fixes and passed afterward. Coverage includes changed skills on reopen, approval context, and an ordinary subsequent task. - Runner E2E unit suite: 306 passed. Harness typecheck passed. The 64 first-task fixture and grader tests also pass. - Full repository typecheck and build passed locally. Server typecheck and Runner build passed again after the native-command change. - Full GitHub Actions CI passed on `23e56447b`: all server/workspace/browser shards, Runner verification, typecheck/release registry, build, canary, policy, and Docker checks. Greptile reviewed this exact head at 5/5 with no unresolved threads. The earlier broad local run had database startup/timing failures that passed isolated retries; the complete remote suite is green. - Merge verification against current master: 312 harness tests and 13 native recovery tests passed. Regenerated semantic contracts and fixture hashes pass their consistency check. Full local typecheck and build also passed on the stacked queue branch. After merging the latest master and preserving the GitHub setup timing regression in the split browser suite, both focused GitHub browser tests passed. Three CI timing/startup flakes passed local verification and one remote retry; all latest-head checks are green. - Real pinned Claude SDK and Claude ACP JSON-RPC probes against a local mock API confirmed that `/skill-name` expands the assigned skill body before the model request and retains the task arguments. A prose mention does not. The probes made no paid model calls. The ACP probe used the current first-task skill body and retained the wake arguments. - [Full 48-case campaign and report](https://pages.paperclip.ing/runner-e2e-first-task-35053063880/): 44 passed after three interrupted Codex cases completed in targeted reruns. Original results, regrades, and all 51 executions remain in the report provenance. - [Claude campaign after the shared-question fix](https://pages.paperclip.ing/runner-e2e-first-task-claude-35099525201/): 10/12 passed with zero single-option failures. All 12 recorded the current assigned skill and corrected guidance. The failures exposed skipped skill invocation and a missing saved plan. This PR adds native command invocation and explicit saved-plan instructions; the subsequent report below still shows behavior failures. - [Fresh 12-case Claude report](https://pages.paperclip.ing/runner-e2e-first-task-claude-35102737804/) at `78452129e`: 10/12 pass after correcting two false proposal-matcher failures. The recordings said “Here is the task I will create and run/complete” in approval cards; the old matcher missed that word order. Regression tests failed before the fix and pass after it. Original results and offline regrade provenance remain linked. No agent rerun was needed. Zero single-option-question failures; two behavior failures remain: direct work before acceptance on a plain first message, and an explicit plan request without a saved plan. Neither check was relaxed. The follow-up `82087ac7e` fixes command-prefix size accounting; `94aefb1f3` fixes only that proposal matcher. - Report browser checks confirm folded conversations, rendered cards, explicit Local/Daytona labels, and no page errors. The published-object audit scanned 1,306 text files across 2,154 objects with no credential-format findings or prohibited files. Image pixels and unknown token formats are outside that scan. ## Risks - Model behavior is nondeterministic. One campaign is evidence, not a guarantee. The two remaining Claude behavior failures are visible in the report and require further product work; this PR does not claim all onboarding scenarios pass. - The suite checks persisted Paperclip effects. It cannot prove the absence of arbitrary external effects. - Historical recordings can miss later journey steps. These remain incomplete, never passes. - Native profiles switch runtime after the production onboarding wizard because it does not yet expose a native option. - Quality scores are informational and cannot override behavioral failures. ## Model Used OpenAI Codex, GPT-6, with reasoning, repository tools, and code execution. The exact deployed model identifier and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Devin Foley
|
dcb04a8062 |
fix(claude-local): read a macOS isolated login from its suffixed Keychain item (#13519)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Connecting a Claude subscription during onboarding uses an isolated
login: the wizard points `claude` at a per-connection
`CLAUDE_CONFIG_DIR` and then verifies the credential before saving the
connection
> - The verifier reads `.credentials.json` from that directory — but on
macOS, Claude Code does not write a credentials file at all: it stores
the OAuth credential for a custom config dir in a per-directory Keychain
item named `Claude Code-credentials-<first 8 hex chars of sha256(dir)>`
> - So on macOS the connect step can never verify a successful sign-in,
and onboarding dead-ends at "Could not verify the local subscription"
(Linux works because the CLI falls back to writing the file there, which
is why the Docker-based smokes pass)
> - This pull request teaches the credential readers to consult the
login home's own suffixed Keychain item when the file is missing
> - The benefit is that macOS self-hosted users can connect a Claude
subscription during onboarding, while the standing isolation invariant —
an isolated login must never fall through to the machine-level operator
login — is preserved, because only the per-directory suffixed item is
ever read
## Linked Issues or Issue Description
No existing issue. Description follows the bug-report template:
**What happened?**
On macOS, connecting a Claude subscription during onboarding (or from
Connections) always fails with "Could not verify the local subscription.
Run the sign-in command shown for this connection, finish signing in,
then try Connect again" — even after `claude auth login` completes
successfully in the isolated `CLAUDE_CONFIG_DIR`.
**Expected behavior**
After finishing the browser sign-in for the printed command, clicking
Connect verifies the subscription and saves the connection.
**Steps to reproduce**
1. On macOS, run onboarding on a fresh instance and reach "Connect a
model" → Claude → Subscription.
2. Run the printed `export CLAUDE_CONFIG_DIR=… && claude auth login`
command in a terminal on the same machine and complete the browser
sign-in.
3. Return and click Connect. Verification fails every time. Inspecting
the isolated directory shows `.claude.json` with a fully populated
`oauthAccount` but no `.credentials.json`; `security
find-generic-password -s "Claude Code-credentials-<suffix>"` shows the
credential landed in the Keychain, where the verifier never looks.
**Paperclip version or commit**
Reproduced on `2026.915.0-canary.11` (`dffc2b3ca`) with Claude Code
2.1.231.
**Deployment mode**
Self-hosted, authenticated instance on macOS.
**Installation method**
`npx paperclipai onboard` (also affects any macOS install; Linux is
unaffected).
## What Changed
- `packages/adapters/claude-local/src/server/quota.ts`:
- New exported helper `readIsolatedClaudeKeychainToken(loginHome)` —
computes the suffixed service name (`Claude Code-credentials-` + first 8
hex chars of `sha256(loginHome)`) and reads only that item via
`/usr/bin/security`; returns null off macOS
- `readClaudeToken` with a custom `CLAUDE_CONFIG_DIR` now consults that
directory's suffixed item after the file reads miss (previously it
refused the Keychain entirely for custom homes). The unsuffixed operator
item is still gated behind the explicit `allowKeychain` opt-in with no
custom home, unchanged
- `server/src/services/local-ai-credentials.ts`: for anthropic isolated
logins, fall back to the suffixed Keychain item after the hardened
credentials-file reads miss. The file path is untouched and still
preferred; the hardened file reader (`readLocalAiCredentialFile` with
its uid/mode/symlink checks) is not bypassed
- Tests: adapter keychain suite extended (suffixed lookup for custom
homes, no unsuffixed fallback when the suffixed item is absent,
off-macOS null); server verifier suite extended (keychain fallback when
the file is missing, file preferred over keychain, absent-login failure
still never touches the ambient reader)
Security note: the suffix binds each Keychain item to exactly one auth
home, so reading it can only surface the login performed inside that
home. The account-isolation invariant the old code enforced by refusing
the Keychain outright ("never substitute the server operator's login for
a user's isolated login") is preserved — the unsuffixed item is never
consulted for an isolated login, and a new test pins that.
The suffix derivation was confirmed against a live login on macOS: a
real `claude auth login` into an isolated home left no credentials file,
wrote the full `oauthAccount` to `.claude.json`, and created a Keychain
item whose suffix equals the first 8 sha256 hex chars of the exact
`CLAUDE_CONFIG_DIR` string; reading it back with the same `security`
invocation returned the live token, which the new code path then
verifies via the existing quota probe.
## Verification
- `pnpm exec vitest run src/server/quota-keychain.test.ts`
(claude-local): 10 tests pass; full claude-local suite: 287 passed, 1
skipped
- `pnpm exec vitest run src/__tests__/local-ai-credentials.test.ts`
(server): 11 tests pass
- Reverting only the verifier change makes the two new server tests fail
— the suite reproduces the live bug
- End-to-end on macOS: a dev server built from this branch, fresh data
dir, full onboarding walk with a real `claude auth login` into the
printed isolated dir — the connect step verifies and saves the
connection
## Risks
- Low. The change is additive and fail-closed: when the suffixed item is
absent (Linux, older Claude Code versions, no login performed), behavior
is byte-identical to today — the file reads run first and the failure
message is unchanged
- The `security` call runs with the existing 10s timeout and swallowed
errors, matching the established unsuffixed-item code path
- No migrations, no API surface changes
## Model Used
Claude Fable 5 (`claude-fable-5`), extended thinking with tool use
(Claude Code).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
|
||
|
Devin Foley
|
d08abcba15 |
ci: cut PR wall clock from ~16 to ~6 minutes (#13521)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Every pull request runs the Trusted PR CI workflow before merge > - The test suites roughly tripled in six weeks, and shard balance did not keep up, so PR runs crept from ~4 to ~17 minutes > - Slow CI delays every merge and every contributor > - This pull request rebalances the shards from fresh measurements, splits the largest test files, reuses the Rust build cache in three more jobs, and takes the policy job off the critical path > - The benefit is a PR wall clock near 6 minutes with the same coverage ## Linked Issues or Issue Description **What existing behavior does this improve?** PR CI wall clock. A typical green run took 16-17 minutes. Two months ago it took about 4 minutes. **Subsystem affected** The Trusted PR CI workflow (`.github/workflows/pr-trusted.yml`), the shard-duration manifests, the vitest shard runner scripts, the `paperclip-runner` package scripts, and the dry-run branch of `release.sh`. **Current behavior** The shard-duration manifests were stale. The general-server manifest had durations for ~400 of 649 suites. The e2e manifest was missing 14 of 29 specs. Stale median weights made shard steps range 417s-806s (server) and 277s-745s (e2e). Three jobs each paid a ~3m40s cold cargo release build. Every test lane waited ~60s for the policy job before it could start. **Proposed behavior** All lanes finish in a narrow ~200-290s band. The manifests carry fresh measured durations for every suite. The three largest test files are split so no single file caps a shard. The Rust cache restore runs in every job that builds the Runner binary. Test lanes start as soon as the gate resolves. **Reason and benefit** Merges stop waiting on CI. The projected wall clock is ~6 minutes for the same test coverage. ## What Changed - Rebuild `scripts/general-server-shard-durations.json` (646 suites) and `scripts/e2e-shard-durations.json` (all specs) from per-suite completion timestamps in runs 35036001734 and 35024948947. - Move the PR server lane to the release-verify shape: `general-server-without-chat` across twelve duration-balanced shards, plus the chat integration suite split by collected test location across three dedicated lanes. - Split `tests/e2e/chat-adapters-ui.spec.ts` into `-providers` and `-messaging` specs, and `tests/e2e/agent-chat.spec.ts` into `-sessions` and `-projects` specs. Each pair shares fixtures through a `.shared.ts` module. Playwright collects the same test sets (39 and 20 tests). - Raise e2e shards to eight and serialized shards to nine. - Run the runner package's `check:all` as four matrix lanes: `check:static`, `check:runner`, and two native vitest `--shard` halves. The union is exactly `check:all`. - Add the read-only Rust cache restore (toolchain pin, `save-if: false`) to the Canary Dry Run, Build, and Typecheck jobs. - Make release.sh preview publish payloads concurrently in batches of eight during `--dry-run`. The real publish path stays strictly serial. - Drop the policy-job lockfile artifact chain. Each lane installs with `--frozen-lockfile` and falls back to an inline `--resolution-only` regeneration. The policy job stays a required check through the `verify` and `e2e` aggregates. - Update the shard-count mirrors and workflow assertions in the partition and gate tests. ## Verification - `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs scripts/__tests__/e2e-shard.test.mjs` — 30 pass. - `node --test '.github/scripts/tests/'*.test.mjs` — 410 pass. - `node --test scripts/__tests__/release-verify-workflow.test.mjs scripts/cloud-source-verification.test.mjs scripts/__tests__/release-dry-run-notes.test.mjs` — 42 pass. - `playwright test --list` collects 39 tests across the chat-adapters split and 20 across the agent-chat split, equal to the original files. - A local vitest collection of the chat suite partitions 995 tests into 498/497 line shards. - Projected shard weights: server 230s x12, chat ~143s x3, e2e 207-242s x8, serialized ~216s x9. ## Risks - The split spec files reorder tests relative to the original files. Every describe seeds its own company, so the specs stay independent; a hidden cross-describe dependency would surface as a deterministic failure in one shard. - The inline lockfile fallback changes install behavior for manifest-changing and stacked PRs. The policy job still validates resolution as a required check. - `release.sh` changes are confined to the `--dry-run` preview branch. The publish loop is untouched. `bash -n` passes and the release dry-run tests pass. - One PR now schedules ~44 fleet runners. If the RunsOn fleet caps concurrency, queueing may absorb part of the gain; watch the first runs. ## Model Used - Claude Fable 5 (`claude-fable-5`, Anthropic), extended thinking, with tool use (shell, file edits) in Claude Code. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge |
||
|
DottaandPaperclip
|
9fd2e50310 |
feat: create company skills from runner tasks (#13538)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The Runner gives agents tools to change company resources. > - Users need agents to save reusable skills during a task. > - A saved skill needs a visible result that users can inspect and edit. > - This pull request adds `create_skill` and a task feed card linked to Skill Studio. > - Users can open the saved skill from the task and edit the same resource. ## Linked Issues or Issue Description **Subsystem affected** Runner tools, company skill storage, task feed, and Skill Studio. **Problem or motivation** The Runner has no dedicated tool to create a company skill. A user cannot follow a creation result from the task feed to the saved skill. **Proposed solution** Add a company-scoped `create_skill` tool. Save the skill with the existing company policy. Add one creation card to the task. Open a named sidebar tab from that card. Let the user open the same skill in Skill Studio. **Alternatives considered** An agent can write a local file, but that file is not a company skill. A second document copy in the task would become stale after a Studio edit. The sidebar therefore reads the saved skill directly. **Roadmap alignment** This extends the shipped Skills Manager, Skill Studio, and Skills Store milestone. The maintainer requested and approved this scope. Search found no duplicate `create_skill` PR or issue. Related UI validation work: #8715. This PR does not change that validation display. ## What Changed - Add the real Runner tool, its contract, and its mock implementation. - Validate the complete SKILL.md and derive company, task, agent, and run identity from authentication. - Apply the existing company skill policy. Do not assign the skill to an agent. - Make keyed retries return one skill and one creation event. Reject conflicting retries. - Make concurrent file creation safe. Never replace an existing published skill during creation. - Add a creation card, a named sidebar tab, and an Open in Skill Studio action. - Show saved Studio edits when the user returns to the task. - Add storage, policy, mode, retry, UI, and Product E2E tests. Document the tool. - Fix deleted-name reuse, onboarding panel persistence, immediate feed refresh, and mock validation parity from review. - Serialize Studio file edits and renames with skill deletion and recreation. Reject stale editor requests before they can change a replacement skill. - Generate the standalone mock parser and validator from the production contract. Use portable UUIDs so the browser scenario bundle builds. ## Verification - All latest-head PR checks pass on `145dd76a5`, including all server shards, browser E2E, Runner verification, build, typecheck, and release dry run. Greptile: 5/5 with no open findings. An interrupted CI runner was retried successfully. - `pnpm -r typecheck`: passed. - `pnpm build`: passed. - `pnpm check:token-gates`: passed. - Review regressions: 73 storage tests, 6 real API tests, 63 UI tests, and 61 semantic runtime tests passed. Parser synchronization passed. - CI exposed existing fire-and-forget Sentry test races. Reproduced the resumption race locally, then synchronized the related sweep and finalizer assertions on the actual report; all 27 tests across the three affected files pass. - Runner scenario browser build and strict content-security-policy check: passed. - Runner suite: 2,012 tests passed; 10 skipped. - `pnpm test:run`: the general-server batch had 12,416 passes and two failures. The old tool-count assertion was fixed; all 16 authority tests then passed. The chat webhook test had a socket error; it passed four isolated reruns. - Both workspace test groups passed. The isolated route suites completed. Two socket failures in the initial route batches passed on individual reruns; all remaining 61 files passed. - Product E2E `create-skill-studio`: passed with local Codex and local ACPX Claude. - Manual browser test: submit a task, observe the real tool call and creation card, open the sidebar, edit in Studio, save, and return. The task reached Done. The saved second revision and sidebar tab survived a server restart. - The new companion headless Runner Eval passed. Companion coverage PR: https://github.com/paperclipai/paperclip-evals/pull/23. Daytona was not run because no immutable runner image was configured. ## Risks - Database writes and local file writes cannot share one transaction. Recovery accepts only an exact file-for-file retry after a database rollback. Conflicting files remain untouched. - The sidebar displays the current skill. The feed card remains the historical creation receipt. - No database migration, dependency, or workflow change is included. - Remote Daytona behavior still needs a run with a configured immutable image. ## Model Used OpenAI GPT-6 (`gpt-6-astra`) handled design, integration, review, and browser verification. OpenAI `gpt-5.6-luna` assisted with bounded implementation and eval work. Both used code execution and tool access. The host did not expose the context window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
18989a9e73 |
docs: add eval guide, authoring skills, and public history hub (#13535)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Its evaluations test both the Runner and complete product workflows. > - The guides and run histories are in separate places. > - The shared Evalbook viewer can make the test boundary unclear. > - This pull request names the two families and adds a guide, authoring skills, and a public hub. > - Contributors can choose the correct test and inspect its history. ## Linked Issues or Issue Description **Issue type** Missing documentation. **Where is the issue?** Runner and Product E2E evaluation guides, case-authoring procedures, and public result navigation. **What's wrong?** There is no single entry point. A report format can be mistaken for an execution boundary. There are no dedicated case-authoring skills for these two families. **Suggested fix** Add a guide and three skills. Link both existing histories from a public hub. Keep existing campaign URLs and grading unchanged. ## What Changed - Add `doc/evals.md` and links from existing guides. - Add the `paperclip-evals`, `add-runner-eval`, and `add-product-e2e-eval` skills. Install copies in `~/paperclipai/.agents/skills`. - Add a static hub builder that reads the existing public history feeds. - Show a dated snapshot for each family. Label partial campaigns and preserve measurement dates across report refreshes. - Document publication and refresh commands for https://pages.paperclip.ing/evals/. ## Verification - Seven Python summary tests pass: `python3 -m unittest discover -s scripts/evals-hub -p 'test_*.py'`. Run these checks directly; this PR does not modify package scripts. - All three skills pass the skill-creator `quick_validate.py` check with `/usr/bin/python3`. - Build tested with saved history fixtures and the live public feeds. - Desktop and mobile browser checks pass. The mobile page has no horizontal overflow. - Published https://pages.paperclip.ing/evals/. Browser check: HTTP 200, no page errors, all eight links return HTTP 200, no mobile overflow. - Independent skill exercises found the existing Notion-decline case and a direct Runner permission-denial case. Roster validation with an explicit run ID passes. - Missing refresh measurement date: regression fails before the fix and passes after it. - `git diff --check` passes. - No paid evals were run for this documentation and reporting change. The preceding head passed typecheck, build, server/workspace tests, runner verification, browser E2E, and the canary dry run. Checks for the latest commit are pending. Local repository-wide typecheck, test, and build were not repeated because no product code changed. ## Risks The hub is a dated static snapshot. It can lag behind the linked histories until an operator refreshes it. A changed history schema stops the build. Existing archives and grades are not modified. The published guide link is pinned to the reviewed commit so branch deletion cannot break it. Later builds can use master. ## Model Used OpenAI gpt-6-astra for implementation and review. OpenAI gpt-5.6-luna for documentation and independent skill checks. Both used repository tools and code execution. Context window sizes are not exposed by this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green (latest commit pending) - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups (preceding head was 5/5; latest commit pending) - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
a8d32e5e61 |
feat(sandbox-providers): add CreateOS sandbox provider (#13434)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Agent work runs in sandboxes that provider plugins supply > - Operators can choose a provider to run agent work > - CreateOS adds another provider with workspace-preserving pause and resume > - This pull request adds a CreateOS provider plugin > - The benefit is that operators can preserve a workspace between runs without keeping its compute active ## Linked Issues or Issue Description Refs #13203 and the earlier closed #13096. This continues the CreateOS contribution from @bhautikchudasama and @ashwaq06. The branch preserves the original implementation commit. Thank you to both contributors. When squash-merging, preserve the original author's credit in the squash commit body: ```text Co-Authored-By: bhautikchudasama <BhautikChudasama@users.noreply.github.com> ``` The original fork rejects maintainer pushes. This branch includes the merge-conflict resolution and review fixes. The request is described below using `adapter_request.yml`. **Agent or provider** CreateOS sandbox API (https://api.sb.createos.sh). **Why this adapter is useful** CreateOS can pause a sandbox and resume it by ID. The workspace survives the pause. This adds a reusable-lease option to the existing sandbox provider system. **How the agent is invoked** Build and install the local plugin as described in its README. Open Instance Settings, then Environments. Select the `createos` driver. Supply an API key and shape. The driver then supplies sandbox leases for agent runs. **Are you willing to implement it?** Yes. This pull request is the implementation. ## What Changed - Adds the `createos` sandbox provider under `packages/plugins/sandbox-providers/createos`. - Calls the CreateOS HTTP API directly. The package adds no vendor SDK. - Implements the environment lifecycle hooks, incremental process output, and binary workspace sync. - Registers the optional bundled provider and its trusted host credential fallback. The fallback is limited to the official API origin; custom endpoints require an explicit key. - Lists the package in the release manifest with `publishFromCi: false` until its first npm publish is bootstrapped. - Waits through delayed pause/resume state updates without duplicate action requests. - Cancels queued API requests promptly while preserving request spacing. - Uses direct CLI invocation in the setup guide so paths and IDs are passed without an extra shell expansion. - Includes current master and retains its existing Git-subfolder containment fix. ## Demo Fresh setup and a run against a CreateOS sandbox. https://github.com/user-attachments/assets/e71b9e06-c006-4fb9-b847-52dfd68f6110 https://github.com/user-attachments/assets/43b5ac75-66bd-4f76-8563-67e4c7759084 ## Verification All 25 jobs in [CI run 34884260542](https://github.com/paperclipai/paperclip/actions/runs/34884260542) passed at commit `f8d0997677024b784fdadf9d44a84c01cb4e813c`, including typecheck, build, native runner verification, server and workspace tests, browser tests, and the canary release dry run. Greptile reviewed the same commit at 5/5 with no unresolved review threads. GitHub reports no merge conflicts. The remaining merge gate is code-owner approval for the new `package.json`, as required by `.github/CODEOWNERS` and the `master` ruleset. Reviewers have been requested automatically. Local checks passed: - Provider: `pnpm typecheck`, `pnpm test` (52 passed, one live smoke skipped), and `pnpm build`. - Host: focused credential and bundled-plugin tests (17 passed), plus CLI invocation safety (39 passed). - Release: package manifest check and release policy tests (18 passed). The full local `pnpm test:run` attempt caught the README command issue; its focused rerun now passes. The full local run stopped after its general-server group: 7,804 tests passed, with unrelated embedded PostgreSQL startup failures and 10 failures in unchanged runtime-skill-cache tests (`EACCES` on directory rename on macOS). It did not reach the later test groups. Local `pnpm -r typecheck` and `pnpm build` reach the runner package and stop because this machine has no Rust/Cargo installation. The corresponding CI checks passed on provisioned runners, as linked above. The live CreateOS smoke requires explicit provider credentials and was not run during this review. It is available with `CREATEOS_LIVE_TEST=1 pnpm test` in the provider directory. The author supplied the demo links above. ## Risks The provider is opt-in and is not installed by default. It is available through a local-path install or explicit image inclusion. npm publication remains disabled until a maintainer bootstraps the package and enables publishing. Sandbox creation has no idempotency key. An ambiguous create response can leave a resource that requires provider-account inspection. Process tracking is in memory; durable lease recovery belongs to the host. The provider does not advertise guaranteed expiry, interactive login, snapshots, duplex channels, or ingress. Live native-runner qualification remains outside this PR's tested claims. ## Model Used Original provider implementation: human-authored by @bhautikchudasama, as reported in #13203. The original description reports Claude Opus 5 assistance. Review and follow-up fixes: OpenAI GPT-6 via Codex, with code review, editing, and tool execution. The precise runtime model variant and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (focused checks; full-suite environment limits documented above) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: bhautikchudasama <bhautikrchudasama@gmail.com> Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
scotttongandClaude Fable 5.1
|
dffc2b3ca1 |
fix(claude-local): skip expired credentials file when reading the Claude token (#13505)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Agents on the Claude Local adapter run on a local Claude Code
subscription. A user connects that subscription from Agent → Harness /
Runtime → "Connect account".
> - The server verifies the login with `readClaudeToken` in
`packages/adapters/claude-local/src/server/quota.ts`. It reads
`~/.claude/.credentials.json` first and consults the macOS Keychain only
when no file is present.
> - On macOS the Claude CLI refreshes the Keychain item, not the file. A
leftover credentials file keeps an expired token forever, and the reader
ignores `claudeAiOauth.expiresAt`.
> - The stale file shadows the live Keychain login. The usage check
fails and the user sees "Could not verify the local subscription"
although `claude auth status` reports a valid login. Running `claude
auth login` again does not help.
> - This pull request skips a credentials file whose token has expired,
so the reader falls through to the Keychain or returns null.
> - The benefit is that a valid local Claude login connects on the first
try, and a dead token is never sent upstream.
## Linked Issues or Issue Description
No public issue exists for this bug. Description follows
`bug_report.yml`.
### What happened?
Agent → Harness / Runtime → "Connect account" → Claude (Subscription) →
Connect failed with:
> Could not verify the local subscription. Run claude auth login in a
terminal on the machine running Paperclip, then try Connect again.
`claude auth status` on the same machine reported `loggedIn: true`,
`authMethod: claude.ai`, `subscriptionType: max`. The Keychain item
`Claude Code-credentials` held a fresh token. `POST
/api/companies/:id/ai-connections/local/check` returned
`{"status":"sign_in_required"}`.
A leftover `~/.claude/.credentials.json` (written weeks earlier) held an
access token that expired the same day it was written. `readClaudeToken`
returned that token. `fetchClaudeQuota` got a non-OK response from
`/api/oauth/usage`, and the route threw the generic 422.
### Expected behavior
An expired credentials file must not block a valid login. The reader
skips the dead token and falls through to the Keychain. Connect
succeeds.
### Steps to reproduce
1. On macOS, sign in with `claude auth login` (credentials land in the
Keychain).
2. Place a `~/.claude/.credentials.json` with
`claudeAiOauth.accessToken` set and `claudeAiOauth.expiresAt` in the
past.
3. Open an agent → Harness / Runtime → Connect account → Claude
(Subscription) → Connect.
4. Before this change: the "Could not verify the local subscription"
error appears. After: the connection is created.
### Agent adapter(s) involved
Claude Code (`@paperclipai/adapter-claude-local`)
### Operating system
macOS (Keychain-backed credentials). On Linux the file is the live
store; an expired file token now returns null instead of a failing
request, so the user-facing message is unchanged.
## What Changed
- `packages/adapters/claude-local/src/server/quota.ts`:
`parseClaudeCredential` now returns the token plus `expiresAt` (epoch
ms) when the file records one. `readClaudeTokenFromFile` returns `null`
for a token whose `expiresAt` is in the past, so `readClaudeToken` moves
on to the next candidate (second file name, then Keychain when
`allowKeychain` is set). Files without an `expiresAt` keep the old
behavior. `parseClaudeCredentialToken` (used for the Keychain payload)
is unchanged in behavior.
- `packages/adapters/claude-local/src/server/quota-keychain.test.ts`:
three new cases — expired file falls through to Keychain; expired file
with no Keychain access returns `null`; a file with no expiry is still
accepted.
## Verification
- `pnpm --filter @paperclipai/adapter-claude-local exec vitest run
src/server/quota-keychain.test.ts` → 7 passed (4 existing + 3 new).
- `pnpm --filter @paperclipai/adapter-claude-local exec tsc --noEmit` →
clean.
- `pnpm --filter @paperclipai/server exec vitest run
src/__tests__/local-ai-credentials.test.ts` → 9 passed.
- Manual, on the affected machine: with the stale file in place, `POST
/api/companies/:id/ai-connections/local/check` (provider `anthropic`,
method `subscription`) returned `sign_in_required`; with the stale file
removed it returned `ready`. This change makes the first case behave
like the second without touching the file.
## Risks
- Low risk. The only behavior change is for a credentials file that
carries a numeric `expiresAt` in the past. Such a token is already
rejected upstream, so the change removes a guaranteed failure rather
than a working path.
- Clock skew: a machine clock that runs ahead of real time could treat a
token as expired slightly early. The fall-through then reads the
Keychain (macOS) or returns null, which triggers the same "sign in"
message the user already sees for an expired token.
- Keychain payloads are not expiry-checked in this PR. The CLI refreshes
that item itself, and `getQuotaWindows` already falls back to the CLI
`/usage` probe when the OAuth call fails.
## Model Used
- Claude — `claude-fable-5-1` (Claude Fable 5.1) via Claude Code, with
extended thinking and tool use (shell, file edit). Root cause found by
reproducing the server's read path against the local credential file and
Keychain.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
|
||
|
DottaandPaperclip
|
d49f168381 |
fix: publish sandbox files on legacy and native runners (#13493)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents must publish generated files so users can inspect their results after a sandbox stops. > - Legacy sandbox bridges blocked attachment listing and could not carry multipart binary uploads through the queue transport. > - The native runner has a separate verified file registration path that needs the same durable result. > - This pull request repairs legacy binary transport, makes native download receipts explicit, and reveals new outputs in the task Artifacts tab. > - Users can open generated files from either runner without a transport flag change. ## Linked Issues or Issue Description Refs #13355 for the existing native file publication path. Related filename fixes: #2615 and #4788. Related sandbox persistence work: #13376. This change repairs attachment delivery through the existing API; it does not add workspace persistence. **What happened?** The upload helper first lists task attachments to avoid duplicates. Both legacy bridge allowlists rejected that GET request with 403. A direct multipart upload also failed: the queue bridge accepted only JSON, excluded attachment uploads, and converted bytes to UTF-8 text. Enabling HTTP/2 alone did not fix the missing listing route. These failures occurred before attachment storage. **Expected behavior** Both runners can publish a workspace file, register its work product, bind it to a response, and return a working download. The file stays accessible after sandbox deletion. A new output opens the task Artifacts tab. The agent receives accurate errors and decides how to retry or report a failure. **Steps to reproduce** 1. Run a legacy agent in Daytona with the duplex bridge disabled. 2. Invoke the bundled upload helper with Bash on a PNG or PDF. 3. Repeat with the duplex bridge enabled. 4. Register the same file through the native runner with generic API tools disabled. 5. Retry registration, delete the sandbox, and compare the downloaded bytes with the original file. **Paperclip version or commit** The failing baseline was `f2c5e54dc`. This branch is rebased onto `6cfe4acff`. **Deployment mode** Source checkout with a local API and real isolated Daytona sandboxes. ## What Changed - Allow authenticated attachment listing, upload, and content download through both legacy bridge transports. - Add optional base64 body encoding to queue envelopes. Preserve the existing UTF-8 contract when the encoding field is absent. Decode binary bodies before forwarding them. - Preserve multipart headers. Bound raw bytes, encoded envelopes, and in-flight reservations. Retain timeout and uncertain-write behavior. - Preserve helper deduplication and return structured uncertain-write failures. Document explicit Bash invocation in live skills. - Add attachment IDs and content/download paths to native registration receipts. Reuse verified local and remote file reads, attachment storage, work-product registration, and response binding. - Preserve Unicode upload filenames and provide a valid Content-Disposition header. - Open the task Artifacts tab when new stored outputs arrive, including a closed desktop panel or mobile drawer. Deduplicate upload and registration events by object ID. Preserve manual selection on refetches, edits, and panel remounts. - Remove task artifact filters, the company Artifacts footer link, and the unassigned group heading and timestamp. ### Screenshot  This is the local display fixture. The image was generated separately and published through the attachment and work-product APIs. ## Verification - Post-rebase `pnpm -r typecheck` and `pnpm build` pass. - The post-rebase local `pnpm test:run` passed 12,369 tests before one existing conversation reset test timed out; all 33 tests in that suite pass when rerun with isolated test configuration. The aggregate command stopped before its remaining groups. GitHub runs the complete suite in separate shards. - All [GitHub verification checks](https://github.com/paperclipai/paperclip/actions/runs/35017893350) pass on `b66ac276dd3d5fc738a22ecea783400106a494d4`: 32 successful checks and two configured skips. The native-session recovery assertion initially raced its fire-and-forget Sentry report; all 13 tests pass locally, and the same-commit CI rerun passes all 170 suites (3,079 tests). - Live post-rebase Daytona: all three file-delivery tests pass. They cover the real Bash helper with the queue bridge, the helper with HTTP/2, and native `register_deliverable` with generic API tools disabled. - Daytona cases cover PNG/PDF bytes, spaced and Unicode names, duplicate registration, response binding, authorization controls, and byte-for-byte downloads after sandbox deletion. - Local focused coverage includes transfer bounds, malformed encoding, interrupted transfers, remote path containment, and native file verification. The attachment route suite passes all 32 tests, including an eight-case filename-header matrix for Unicode and special characters, inline and forced downloads, and full and partial responses. - Browser verification confirms image previews, persisted downloads, automatic Artifacts selection, and preserved manual selection after edits and reloads. Desktop/mobile component coverage passes. The latest UI cleanup passes its 10 affected tests and token gates. - Coverage limit: the Daytona tests call the real helper and native registration path directly. They do not replay a complete model-led image-generation task through the browser. Live command (requires a configured Daytona credential): ```sh PAPERCLIP_FILE_DELIVERY_DAYTONA=1 pnpm exec vitest run server/src/__tests__/file-delivery-bridges.test.ts ``` ## Risks - Binary queue bodies use more memory because base64 adds encoding overhead. Transfer and process limits must remain aligned. - An interrupted write can have an unknown result. The bridge reports this state and preserves stable retry identities. - New artifacts intentionally change the active task tab. Existing history and repeated updates must not take focus again. - Transport flag defaults, server authorization, frozen skill snapshots, and completion policies remain unchanged. No schema migration is required. ## Model Used OpenAI Codex, based on GPT-6, with reasoning, repository tools, code execution, and browser testing. The runtime does not expose a more specific model ID or context-window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
f4cdc7b231 |
fix: recover transient workspace bootstrap scans (#13481)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The control plane prepares task workspaces before it starts an agent. > - Workspace preparation reads Git state so it can preserve edits and exclude private files. > - A failed scan was treated as a non-Git folder and lost its actual failure code. > - The resulting generic setup failure could not recover, even when the cause was temporary. > - This pull request keeps the cause and uses the existing bounded retry schedule before provider startup. > - Tasks can recover without human intervention, while permanent failures and exhausted retries stop with useful guidance. ## Linked Issues or Issue Description **What happened?** A Git scan error during managed repository preparation became `Configured repository folder is not a Git checkout`, followed by generic `setup_failed`. The agent never started. Generic recovery could not distinguish a temporary timeout from a bad workspace configuration. **Expected behavior** Keep the closed scan error code. Retry temporary timeouts and queue saturation under the existing shared budget. Preserve edits, exclusions, ownership, and pause gates. Stop permanent failures and exhausted retries with a specific explanation. Do not replay historical generic setup failures. **Steps to reproduce** 1. Configure a task project with a local Git source that must be copied into its managed repositories. 2. Make the ignored-file scan exceed its timeout before the agent starts. 3. Before this fix, the snapshot returns null and the run ends as non-retryable `setup_failed`. 4. Use the disposable browser fixture in `tests/e2e/workspace-bootstrap/README.md` to inject real timeouts and test the full recovery path. **Paperclip version or commit** Reproduced against `4510bf7c9e2fcbeb043445850928b5dcb79908ca`. **Deployment mode** Built from source. The defect is in core workspace setup, not a specific model provider. Related work: Refs #13442 (managed repository preparation), Refs #11572 (bounded Git scheduler), Refs #12997 (separate adapter startup retry work), Refs #13469 (separate terminal-workspace scan performance work). ## What Changed - Return the non-Git fallback only for repository discovery. Propagate failed scans of a confirmed repository. - Replace full ignored status output with an ignored-only directory listing. Preserve NUL-delimited paths and exclusions. - Preserve typed, sanitized scan errors through workspace preparation and persist pre-provider failure details. - Retry only timeouts and queue saturation, using the existing durable two-retry budget and issue gates. Prevent generic recovery from adding another budget. - Show workspace-specific failure copy and actionable exhausted-recovery notices. - Add red-green unit tests, real-database restart and retry-boundary tests, and opt-in browser acceptance fixtures with real Git subprocess timeouts. - Document the recovery contract and browser verification procedure. ## Verification - Red: injected scan failures returned null instead of rejecting; setup lost the timeout code; task-thread and recovery notices had generic copy. - Green: 119 focused adapter/backend tests, 20 recovery-boundary tests, and 136 task-thread tests. - `pnpm -r typecheck` — passed. - `pnpm build` — passed on the final production code. - `pnpm check:token-gates` — passed. - The initial local `pnpm test:run` overlapped source edits and was interrupted after two late-added assertions saw pre-fix behavior; it is not counted as a green full run. A fresh final-head run passed all 229 tests across the six affected adapter/backend/UI suites. The clean latest-head CI full test matrix passed: all five general-server shards, all five serialized-server shards, and all three general-workspace shards. - Latest-head CI also passed all three browser shards and their aggregate gate, typecheck and release registry, build, runner verification, canary dry run, policy, Docker context integrity, and security gates. Greptile: 5/5, with the review thread resolved. - Browser: created a task in a disposable instance. A real Git timeout scheduled recovery, the next run completed through the run-scoped API without manual Retry, and Done survived reload. The deterministic process worker checked preserved source edits and excluded private files; no model calls were made. - `WORKSPACE_BOOTSTRAP_TEST_URL=<disposable-instance-url> pnpm exec playwright test --config tests/e2e/workspace-bootstrap/playwright.config.ts` — 2 passed (3.6 minutes). The persistent case made exactly three failed attempts, never started the worker, showed the cause-specific notice, stayed stopped for another scheduler tick, and retained Blocked after reload. - Extra red-green coverage: 50 recovery tests passed after fixing an exhausted-bootstrap classification that incorrectly implied unknown provider actions. Missing or uncertain evidence still retains the safety hold. - Verified the documented Git executable override during repository seeding. ## Risks - A confirmed repository scan failure now fails closed instead of falling back to directory sync. This prevents unfiltered copying but makes previously hidden errors visible. - Temporary host problems can create up to two additional setup attempts, 30 seconds apart. Permanent scan errors do not auto-retry. Generic recovery cannot reset this budget. - The durable retry path still enforces ownership, pause, and work eligibility. Integration tests cover restart, duplicate promotion, pause, exhaustion, and non-retryable categories. - No schema migration, new runtime setting, new retry budget, production deployment, or historical task replay. ## Model Used OpenAI Codex, GPT-5-based coding agent, with reasoning, repository tools, shell execution, and browser testing. The exact deployment model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
cceeb0aa66 |
test(runner): add everyday workflow evaluation harness (#13474)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The native runner must support project work, delegation, hiring, and service access. > - Browser tests exposed lost connection access, rejected helper events, and stalled recovery. > - Some eval failures also came from incorrect fixtures and decision controls. > - This pull request fixes those paths and adds eight everyday workflow stories. > - The tests retain observed failures and verify delivered files independently. > - The benefit is repeatable evidence for common user tasks and their remaining gaps. ## Linked Issues or Issue Description Related work: #13404 contains earlier workflow fixes. #13300 and #13470 changed the CI contracts used by the harness security tests. Merged companion: [paperclip-evals#22](https://github.com/paperclipai/paperclip-evals/pull/22). **What happened?** Native ACPX sessions did not receive the assigned connection gateway. Codex helper events could arrive before their spawn receipt and fail thread validation. A parent continuation could take a shared workspace before its child retried. A failed native continuation could leave the task status without a clear recovery blocker. The eval harness also confused tool approvals with new connection requests and could reject a valid delegated download. **Expected behavior** Keep assigned gateway access and its approval checks. Verify helper lineage before accepting helper progress. Let a waiting child proceed before automatic parent recovery. Preserve a failed task's recovery ownership. Grade the actual requested workflow and its delivered files. **Steps to reproduce** Run the everyday workflow suite with the native Codex and Claude profiles. Exercise service approval, connection refusal, delegated project work, and teammate reuse. The commands and case requirements are in `tests/runner-e2e/EVERYDAY-WORKFLOWS.md`. Use `pnpm test:runner-recovery` for controlled crash and replacement cases. ## What Changed - Pass the scoped connection gateway binding through the native ACPX host and sidecar. - Recognize Codex helper lineage from parent metadata and spawn receipts. Verify early helper events with `thread/read`. Keep helper events separate from root completion authority. - Guide agents to use persistent hiring, child tasks, dependency records, and a blocked handoff while waiting for a child. - Defer automatic parent recovery while a child has an active execution path in the same shared workspace. Allow parent recovery when the child needs review. - Record Blocked status and recovery evidence when a failed native continuation needs reconciliation, including existing active or escalated incidents. Preserve their owner and retry budget. - Add eight browser-driven workflow cases. Use real decision controls, explicit child feedback delivery, managed hiring credentials, and independent ZIP checks inside a bounded Docker sandbox. Verify sandbox availability before task creation. Record screenshot SHA-256 at capture. - Keep runner crash probes in controlled recovery tests. Preserve the original failure when cleanup also fails. - Display missing accounting and replay revisions as unavailable. Align harness security assertions with the approved CI changes. - Make the channel-rejection browser fixture bind its file after the send captures its payload. This prevents live refresh from removing the file before the simulated race. ## Verification - Full workspace `pnpm -r typecheck` passed after merging current master. - Runner E2E typecheck passed. Harness unit tests passed: 216/216. - Wake-queue database tests passed: 55/55. The two added existing-incident tests failed before the fix and pass after it. - Docker artifact calibration passed: 12/12. Host-file and host-loopback isolation tests failed before the fix and pass after it. Read-only delivery and output limits are also verified. - Full `pnpm build` passed. Targeted recovery tests passed: 83/83. - The channel-rejection browser test passed five consecutive runs after fixing the fixture race found in CI. - Local general-server (12,351 tests), UI (6,250), CLI (485), and workspace package groups passed. The monolithic run stopped at an unchanged lock-heartbeat fixture race; the isolated workspace group passed on rerun (shared: 747/747). A separate local serialized run passed 97 files before two socket errors in the unchanged issue-list route suite; that suite passed 15/15 on isolated rerun. These local full commands did not finish uninterrupted; the complete CI matrix below covers the remaining suites. - Final head `0fb293733fe307be7e6667ae8f1364077d0c6455`: **34 successful checks, 2 expected skips**, including every server/workspace shard, browser shard, native runner verification, build, and typecheck. [Final CI run](https://github.com/paperclipai/paperclip/actions/runs/34989136700). - Greptile reviewed this exact head at **5/5**; all review threads are resolved. Both Superagent security checks are successful. - ACPX credential-boundary tests passed: 118/118. Superagent accepted the runner/sidecar versus provider-environment trace and cleared its finding. - The latest paid local campaign on source `f6a2fdf7ac2af859826a2ae627ff4125a5478529` passed 22/24 cases: Sol 8/8, Claude 7/8, Mini 7/8. These results predate the merge with current master. - The two remaining failures are in `hire-reuse`: Claude exceeded the attempt deadline during final review; Mini made invalid deliverable tool calls and remained Blocked. - Six Daytona cases were not run because the matching immutable runner image was unavailable. This PR does not claim new remote model results. ## Risks The changes affect connection admission, helper identity, and recovery scheduling. Assigned gateway grants and user approval still govern service calls. The workspace admission gate still exists; the broader folder-sync design is separate work. Provider behavior can still cause the two recorded hiring failures. No database migration is required. Paid cases are opt-in and have bounded attempt deadlines. Project stories now require Docker and the documented pinned Python image on the harness host. ## Model Used OpenAI `gpt-6-astra` performed implementation, diagnosis, and substantive review. OpenAI `gpt-5.6-luna` assisted with verification, PR preparation, and review tracking. Both used repository tools and code execution. Context-window sizes were not recorded. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (focused checks and isolated reruns; full-run limitations are documented above) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: OpenAI GPT-5.6 Luna <noreply@openai.com> Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Nicky LeachandPaperclip
|
9ed55f6931 |
fix: allow concurrent agent runs on one OpenAI or xAI subscription connection (#13452)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Paperclip starts local command-line sessions and stores provider credentials through managed connections > - One OpenAI Codex or xAI Grok subscription connection held a credential lease for the full agent run > - A second run then waited for the first run, and concurrent runs could overwrite a newer credential > - The write-back must compare fresh credentials while it holds the row lock > - This pull request removes the run lease, keeps the revocation guard, and bounds Codex timestamps against the host clock > - The benefit is safe concurrent use of one subscription connection with newest-credential selection ## Linked Issues or Issue Description **What happened?** A managed OpenAI Codex or xAI Grok subscription connection held a credential lease for the full agent run. A second run waited for the first run to finish. The write-back gate also rejected any row change before it compared credential freshness. **Expected behavior** Concurrent runs should start on one subscription connection. The server should keep the newest valid credential and reject a credential write after a person revokes the connection. **Steps to reproduce** 1. Start two runs that use one OpenAI or xAI subscription connection. 2. Let both provider tools refresh the credential. 3. Finish the runs in either order. 4. Confirm that the newest valid credential remains in the connection. **Paperclip version or commit** `ec25bf1e4a81d1729a6d7276e6a486587cff4a0b` **Deployment mode** Local dev (`pnpm dev`), built from source. **Agent adapter(s) involved** Codex. The server path also covers xAI Grok subscription connections. **Database mode** Embedded PGlite for local development, and external Postgres for deployments. **Access context** Both board and agent runs can use managed connections. **Additional context** The provider command-line tool refreshes credentials inside the sandbox. The server copies the result back after the run. Two long runs can still refresh one token hours apart, so the provider can reject the second refresh. The server cannot observe that provider call. ## What Changed - Remove the full-run credential lease for OpenAI Codex and xAI Grok subscription connections. - Lock and re-read the connection row before credential write-back. - Accept only a strictly newer credential, while keeping the connection revocation guard. - Reject Codex freshness timestamps more than five minutes ahead of the host clock. - Add tests for both completion orders, xAI cleanup, revocation, the Codex time bound, and agent hiring. - Update the connection and run-log documentation. ## Verification - `server/src/__tests__/ai-connections.test.ts` passes with 42 tests. - `packages/adapters/codex-local/src/server/codex-auth-merge-decision.test.ts` covers the five-minute boundary and the one-millisecond overflow. - `packages/adapters/codex-local/src/server/codex-auth-merge.test.ts` passes. - `server/src/__tests__/agent-hire-ai-connections.test.ts` covers OpenAI and Anthropic. - The project type check reports no new error in changed files. - GitHub Actions must pass on this pull request. ## Risks The write-back now permits concurrent runs, so the provider may reject a later refresh when both long runs use one token. The server keeps the revocation guard and rejects future-dated Codex timestamps. No schema change occurs. ## Model Used OpenAI Codex, GPT-5. Context window and exact deployment build are not exposed in this run. The model used tool calls, code inspection, and test verification. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes: #` / `Refs: #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Nicky LeachandClaude Opus 5
|
b64469e403 |
feat(workspaces): add an operator default for isolated execution workspaces (#13444)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The execution workspace subsystem decides if a task run uses the
shared project checkout or an isolated per-task git worktree
> - The mode comes from the project policy, then the task settings. A
project that stores no policy always falls back to the shared checkout
> - An operator who wants every project to use isolated workspaces must
therefore edit each project one at a time, and must repeat this for each
new project
> - There is no instance-level control, so a fleet operator cannot set
this default at all
> - This pull request adds a managed experimental flag that moves the
default for projects that store no policy of their own
> - The benefit is that an operator sets the workspace default one time,
and every current and future project follows it
## Linked Issues or Issue Description
No public issue exists. The description follows the feature request
template.
**Subsystem affected**
Execution workspaces. The files are
`server/src/services/execution-workspace-policy.ts` and the run dispatch
path in `server/src/services/heartbeat.ts`.
**Problem or motivation**
`resolveExecutionWorkspaceMode` reads only the project policy, the task
settings, and a legacy field. Its last statement returns
`shared_workspace`. A project that stores no policy always gets the
shared project checkout.
An operator has no way to change this default for many projects at the
same time. The operator must edit each project, and must edit each new
project again later. Tasks in one project therefore share one checkout,
and they run one at a time when the environment driver makes the
scheduler serialize them.
**Proposed solution**
Add the managed experimental flag `enableIsolatedWorkspacesByDefault`.
When the flag is on, a project that stores no policy of its own resolves
as if it selected isolated workspaces. A project that stores a policy
keeps that policy.
The new helper substitutes a project policy. It does not move the last
statement of `resolveExecutionWorkspaceMode`. Two behaviors make this
necessary:
- A task that has no project must keep its current behavior. An isolated
workspace needs a repository to cut a worktree from.
`isUnrunnableWorktreeCombo` blocks an isolated task that has no
`projectId` and no `projectWorkspaceId`. A moved fallback would resolve
isolated for project-less tasks, such as agent chat, and stop them
before dispatch.
- The mode and the strategy must agree.
`buildExecutionWorkspaceAdapterConfig` supplies the default
`git_worktree` strategy only when one layer asserts workspace control. A
moved fallback would leave isolated mode with a `project_primary`
strategy.
**Alternatives considered**
- Change the last statement of `resolveExecutionWorkspaceMode` to
`isolated_workspace`. This is one line, but it changes the default for
every deployment. It is also not gated, so it would apply where isolated
workspaces are off.
- Write the policy to each project row with a script. This does not
cover new projects, and it does not cover new instances.
- Add an instance defaults section to the managed-config document. This
needs a new document key, new validation, and new delivery code. A
boolean flag reuses the delivery machinery that exists today.
**Roadmap alignment**
`ROADMAP.md` does not list execution workspace defaults. This change
adds an operator control to an existing capability. It does not add a
new capability.
**Additional context**
The flag is `tier: "managed"`. A cloud operator can therefore deliver it
with the managed-config machinery that exists today. No new delivery
code is needed.
## What Changed
- Add `enableIsolatedWorkspacesByDefault` to the feature catalog with
`tier: "managed"`. Both defaults are off.
- Add the flag to the experimental settings schema, the type, and both
branches of `normalizeExperimentalSettings`.
- Add `applyDefaultIsolatedExecutionWorkspacePolicy` to
`execution-workspace-policy.ts`. It substitutes `{ enabled: true,
defaultMode: "isolated_workspace" }` only when the flag is on, the task
has a project, and the project stores no policy.
- Apply the helper in the run dispatch path in `heartbeat.ts`, after the
existing `gateProjectExecutionWorkspacePolicy` call. The `hasProject`
argument reads the resolved project row, not the raw `projectId` of the
task.
- Gate the new flag behind `enableIsolatedWorkspaces` at the call site.
The new flag does nothing on its own.
- Add a toggle card to the instance experimental settings page. The card
shows only when isolated workspaces are on.
- Add eight tests for the new helper.
## Verification
Commands:
```
pnpm --filter @paperclipai/shared typecheck
pnpm --filter @paperclipai/ui typecheck
cd server && ../node_modules/.bin/tsc --noEmit
```
The server typecheck script also builds a Rust binary. I ran `tsc`
directly because this machine has no `cargo`. The server package reports
no type errors.
Tests:
```
./node_modules/.bin/vitest run \
server/src/__tests__/execution-workspace-policy.test.ts \
server/src/__tests__/instance-settings-service.test.ts \
server/src/__tests__/instance-settings-cloud-defaults.test.ts \
server/src/__tests__/instance-settings-managed-overlay.test.ts \
server/src/__tests__/instance-settings-routes.test.ts \
server/src/__tests__/managed-config.test.ts \
server/src/__tests__/heartbeat-workspace-busy.test.ts \
server/src/__tests__/heartbeat-workspace-session.test.ts \
server/src/__tests__/heartbeat-workspace-ready-comment.test.ts \
server/src/__tests__/execution-workspaces-service.test.ts \
server/src/__tests__/issue-runtime-workspace-binding.test.ts \
server/src/__tests__/run-trust-preset.test.ts \
packages/shared/src/feature-catalog.test.ts \
packages/shared/src/settings-visibility.test.ts \
packages/shared/src/validators/instance.test.ts \
ui/src/pages/InstanceExperimentalSettings.test.tsx \
ui/src/components/Sidebar.test.tsx
```
All of these files pass. The new tests cover each of these cases:
- The helper substitutes an isolated policy for a project that stores
none.
- The helper changes nothing while the flag is off.
- The helper changes nothing for a task that has no project.
- The helper keeps a stored policy, including a policy with `enabled:
false`.
- The resolver returns `isolated_workspace` for an unpolicied project.
- An explicit task setting still wins over the operator default.
- The substituted policy produces the `git_worktree` strategy.
- A project-less task does not become an unrunnable worktree.
To confirm the behavior by hand:
1. Turn on Isolated Workspaces, then turn on Use Isolated Workspaces By
Default.
2. Open a project that has no execution workspace policy.
3. Start a task in that project.
4. The run gets its own worktree. Tasks in that project no longer wait
for each other.
## Risks
Low to medium. The details:
- The flag defaults to off, and it is inert unless
`enableIsolatedWorkspaces` is also on. An instance that does not turn on
both flags sees no change.
- A project that stores a policy keeps it. This includes a policy with
`enabled: false`, which the helper reads as a decision to stay on the
shared checkout.
- When an operator turns the flag on, the workspace configuration
fingerprint changes for projects that store no policy. Their next run
creates a new workspace. This is correct, because the mode did change,
but the first run after the change does more setup work.
- A task that is in flight when the flag changes resumes with a
different workspace path than the path its session remembers. An
operator should let current runs finish before turning the flag on.
- Isolated workspaces use more disk, because each task gets its own
worktree.
## Model Used
Claude Opus 5 (`claude-opus-5`) in Claude Code, with extended thinking
and tool use. The model read the repository, made the change, and ran
the typechecks and tests above.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
DottaandPaperclip
|
a2e7ffdc34 |
fix(runtime): validate sandbox paths and preserve live controller leases (#13432)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents can run in remote sandboxes. > - Connection checks must use the selected execution target. > - Recovery must respect the controller that owns an active run. > - A host path or PID does not describe a remote sandbox. > - This pull request checks sandbox paths on the target and preserves live controller leases. ## Linked Issues or Issue Description **What happened?** Selecting an AI account for a sandbox agent could fail because the Claude ACP environment check tried to create the sandbox directory on the Paperclip host. The recovery sweep could also interrupt a sandbox run while its controller lease was still valid. It treated a PID absent from the local host as proof that the run had stopped. **Expected behavior** ACP checks directories on the selected execution target. Recovery leaves a run with a live controller lease alone. Its final database write rejects a stale snapshot after renewal, a claim, a controller change, or a runtime change. **Steps to reproduce** 1. Test a Claude ACP sandbox agent with a directory that cannot be created on the host. The check fails before this fix. 2. Give a running sandbox task a valid controller lease and a PID absent from the host. Run the stale-lock sweep without an in-memory handle. The sweep interrupts the run before this fix. 3. Renew or replace the controller between the sweep's read and write. The old snapshot must not end that controller's run. **Paperclip version or commit** Rebased onto `origin/master` at `0e9b24c8216171c26c8358ba387d77858e02c7a9`. All seven regression cases still fail against this base. Refs #13438, which supplies the managed hiring and task-connection behavior, and #13433, which preserves non-assignee subscription comment wakes. This PR preserves both upstream changes and addresses the two remaining sandbox failures. ## What Changed - Resolve and create Claude ACP test directories through the execution-target helpers. - Preserve active legacy controller leases during stale-lock recovery, including finalization after a task becomes terminal. - Recheck the controller, lease, runtime mode, and native ownership in the terminal database write. - Add two sandbox-directory cases and five database-backed controller-lease cases. - Document the target used for ACP directory checks. ## Verification - Red: all seven new cases fail against `0e9b24c82` without these two implementation changes. - Before the final upstream sync, 192 focused tests passed. Full `pnpm test:run` coverage completed using the repository's group/shard runner: all general server and workspace groups passed, and all 147 serialized server suites passed across the initial run and isolated continuations. Five cold-import timeout suites passed with `--experimental.fsModuleCache`; their assertions and deadlines were unchanged. - The hiring routes, default-selection service, and upstream hiring tests match `origin/master` exactly. The two remaining fixes are unchanged by the final rebase. - On final head `bd28d5cefbdf7084acc3759199cd9661907e7a26`, all 372 focused tests pass across 16 suites covering both upstream changes and these fixes. Two timeouts in the combined run (database setup and an existing ACP case) pass in isolated reruns with fresh test homes and temporary directories. `pnpm -r typecheck` and `pnpm build` also pass on this head. - All 32 active checks pass on final head `bd28d5cef`, including the full test matrix and browser shards, in [CI run 34904204849](https://github.com/paperclipai/paperclip/actions/runs/34904204849). Two Storybook checks are skipped by path filters. - Greptile reviewed final head `bd28d5cef` at 5/5 with no findings or unresolved review threads. ## Risks - A failed remote directory check still blocks connection adoption. - A live controller retains finalization authority after its task becomes terminal. Cleanup waits for ownership to expire and must pass the final ownership check. - No schema or credential-storage changes. ## Model Used OpenAI GPT-6 in Codex, with reasoning, repository inspection, code execution, and API tools. The exact serving model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
8f1905d34d |
fix: provision all project repositories for local and sandbox tasks (#13442)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Projects can now attach several source repositories. > - Task preparation still treated these sources as alternative workspaces. > - Sandbox sync preserved Git history only for the selected repository. > - A task needs every attached repository to complete work across the project. > - This pull request prepares all distinct project repositories and preserves their separate Git histories through sandbox restore. ## Linked Issues or Issue Description **What happened?** A user reported that a project with two repositories received only the first repository in Daytona. Repository-only project rows also reached the agent with null local paths. Managed checkouts with matching repository names could resolve to the same directory. **Expected behavior** Local and sandbox tasks receive every distinct repository attached to their project. Repository-only sources work without preconfigured local folders. Each repository keeps its own Git history and working files. **Steps to reproduce** 1. Create a project with two repository sources and no local folder paths. 2. Assign a task to the project and run it in Daytona. 3. Inspect the task workspace and the repository paths exposed to the agent. 4. Observe that the original implementation supplies only the selected checkout. Related change: #13010 added multiple repository selection. The open repository-catalog proposals #11234 and #11228 cover a different data model. This fix uses the existing project workspaces. ## What Changed - Materialize each additional distinct repository as an editable checkout inside the task root. Seed configured local sources with their current working files and retain task edits across runs. - Pass materialized repository paths to local agents and native sandbox task prompts. Apply existing run-scoped Git credentials to each remote clone. - Preserve each repository's Git history, dirty files, and restore baseline during sandbox staging and durable recovery. Apply each repository's ignore rules and the operator's workspace exclusions. - Keep same-name managed repositories in separate directories. Report additional clone failures before the task starts. - Add task-level, checkout, sandbox round-trip, environment-hint, and recovery-descriptor regression coverage. Document checkout and restore behavior. ## Verification - Red: the original implementation fails the sandbox test because the second repository has no Git directory. It also fails the same-name checkout test and both real-database task tests because repository hints have no local path. - Green: focused tests pass for one and two repository-only sources, local source edits, clone failures, per-repository credentials, separate Git histories, ignored files, and recovery from remote or durable seed state. - Live Daytona smoke passed with two disposable repositories through the production provider sync functions. Both repositories arrived with Git history. Commits from both restored locally. Ignored files stayed excluded. The disposable sandbox was deleted. - Passed on final commit `93ab76763`: `pnpm -r typecheck` and `pnpm build`. - Final focused coverage: 254 assertions across the six changed test areas passed across the serial run and an isolated rerun of the existing process-kill timing test. The live Daytona smoke also passed. - The local `pnpm test:run` overlapped source edits and retained stale transformed code. Its first phase reported 12,240 passed assertions, nine failed assertions, three hook failures, and one worker error; later phases did not run locally. This run is not claimed as green. Fresh focused tests verify the changes, and every general/workspace and serialized-server CI shard passes on the final commit. - Final CI is green on `93ab76763`: all test shards, all three browser shards, typecheck, build, runner verification, canary dry run, and security checks. The initial unrelated chat-delivery browser timing failure passed in the final CI run. Optional Storybook visual checks were skipped. - Greptile is 5/5 on the final commit with no unresolved review threads. Its checkout-race finding was reproduced with a failing test, fixed, and rechecked. ## Risks - Additional repositories need disk space and clone time. Access failure for an attached repository stops preparation. - Additional checkouts live under `.paperclip-repositories/` and keep independent histories. Changes stay in those task copies; they do not overwrite configured source folders. - Detached or reconfigured repository copies are retained under `.paperclip-runtime/detached-repositories/`. Sandbox recovery retains per-repository merge baselines. - No database migration, UI contract change, or new credential delegation is required. Referenced projects retain their separate read-only behavior. ## Model Used OpenAI Codex, based on GPT-6, with repository inspection, code execution, and tool use. The runtime does not expose a more specific model deployment ID or context-window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
f912ecaacf |
fix: carry AI connections through hiring and unblock task execution (#13438)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents hire other agents and assign tasks to them. > - Managed AI connections must follow those hires across legacy and native runners. > - Missing accounts should pause task execution and let the user connect from the task. > - Subscription contention must wait without asking for new credentials. > - This pull request fixes these paths and the native tool and Daytona staging failures found during live tests. > - The result is a working hire, subtask, and connection setup flow on local and remote runners. ## Linked Issues or Issue Description **What happened?** A managed Claude or Codex agent could hire a teammate without a usable AI binding. Cross-provider hiring could fail before the user had a chance to connect the new provider. First-time task setup did not show the existing AI credential form inline. A busy subscription could request a new connection. Native API replies could stop the parent after a hire had already committed. Fresh Daytona sandboxes could fail to extract read-only skill directories created on macOS. **Expected behavior** Compatible hires inherit the managed connection choice. A hire for another provider uses the responsible user's default. If that account is missing, the hire succeeds and the task asks for a connection. Completing setup in the task resumes work automatically. Explicit child auth settings and existing unmanaged login paths keep precedence. Shared-account access checks remain in force. **Steps to reproduce** 1. Connect a Claude or Codex parent with a managed AI account. 2. Ask it to hire one agent of each provider and create a self-assigned subtask. 3. Assign work to both hires without connecting the second provider first. 4. Connect the missing provider from its task card. 5. Check that all tasks finish and same-provider work uses the original account. 6. Repeat with native runners and fresh Daytona sandboxes. The opt-in browser suite in `tests/hiring-ai-connections/README.md` performs these steps. **Paperclip version or commit** The live failures were reproduced from `f2c5e54dc`. The branch is rebased onto `5282cabde`. **Deployment mode** Isolated local development instance. Legacy CLI and native runners. Local execution and ephemeral Daytona sandboxes. Related work: Refs #13247 for managed AI connections. Refs #13268 for legacy credential-reference inheritance, which this branch preserves. Refs #13432 for a concurrent managed-inheritance fix. This PR also covers cross-provider task setup, subscription waits, native API replies, and Daytona extraction. It permits missing responsible-user defaults at hire time; restricted shared selections still fail. ## What Changed - Apply managed connection defaults to both agent creation routes. Preserve explicit auth choices and legacy credential-reference inheritance. - Allow hires before their responsible user connects the provider. Keep approval gates, company boundaries, and shared-account access checks. - Reuse the production AI credential form inside the pending task card. Resume the task after setup. - Retry subscription lease contention without consuming the provider-failure allowance or creating a connection request. - Require task execution-lock ownership when scheduling, promoting, and dispatching subscription retries. Recheck ownership under the issue row lock. - Rename the HTTP operation identity at the native tool boundary so it cannot override the runner's operation identity. - Delay directory permission restoration during Daytona extraction. Preserve the final read-only modes. - Add database-backed regressions, real browser acceptance tests, and Storybook states. Document setup and run-log behavior. ## Verification - Six real browser scenarios passed: both parent providers on legacy local and legacy Daytona; native Codex locally; native Claude on Daytona. Each scenario hires both providers, completes a self-subtask and assigned work, and connects the missing provider inline with automatic continuation. - Successful runs verify the account, responsible user, runner mode, and Daytona lease. All 18 test sandboxes were deleted. - Live authentication used API keys. Subscription inheritance, lease contention, and retry have integration coverage. Fresh subscription OAuth sign-in was not automated. - Red/green tests reproduced missing bindings, missing inline forms, subscription contention, native API reply failure, and GNU tar permission failure. - Seven Storybook browser checks passed. They cover both providers, method selection, narrow layout, completion, cancellation, and invalid credentials. - Full local suite coverage completed before rebase. Initial timing and fixture startup failures passed unchanged on isolated reruns. The first full command did not exit cleanly; the remaining workspace and serialized groups were completed separately. - After rebase, 107 hiring/auth/retry tests and 59 native API, task-card, and Daytona tests passed. The full workspace typecheck, production build, and token gates passed again. Storybook build passed before rebase. - Review fixes: 169 hiring/retry/dispatch tests, 37 adjacent tests, and four explicit cancellation-race cases passed. Eight cross-provider cases cover stale auth keys on both creation routes and both runner types. Server typecheck and build passed. - Final CI on `ee4890837a8a4913e07453392b9a75969580dae1`: 32 checks passed. Two optional Storybook jobs were skipped. The full server, workspace, browser, native runner, build, typecheck, and release checks passed. - Three unchanged tests initially failed on a busy port, a chat row-lock race, and preview-server readiness. Each affected job passed after one CI rerun. Isolated local checks also passed: 41 credential tests, the chat-concurrency case, and 25 preview-runtime tests. - Greptile reviewed the final commit at 5/5. Both review threads are resolved. GitHub reports no merge conflicts. ## Risks - A missing personal account now defers authentication to the first task. Explicit incompatible bindings and restricted shared accounts still fail at hire time. - An inherited personal default uses the responsible user's existing authorization to install access for the new agent. It never copies credentials or another user's identity. - Subscription contention retries after a delay and rechecks task eligibility. It does not consume the provider-failure budget. - Native hiring uses the existing managed API-tools opt-in. Remote native runners require a matching Linux binary and provider pack, as documented in the acceptance README. - No schema changes. Live tests make paid provider calls and remain opt-in. ## Model Used OpenAI GPT-6 (`gpt-6-astra`) in Codex, with reasoning, repository inspection, code execution, browser automation, and API tools. The context-window size is not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
Devin Foley
|
5054c9ef9b |
fix(grok): stage the environment test from a host directory, and survive an absent workspace (#13416)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - An agent runs through an adapter. Before you use an environment, the
adapter's environment test probes it and reports structured checks
> - The grok adapter's environment test stages the managed account into
a remote environment first. It gave the runtime its resolved `cwd` as
the host workspace directory
> - For a remote target that `cwd` is a path inside the sandbox. It does
not exist on the host that runs the test
> - A new workspace ignore scan reads that host directory. The scan
failed on the missing directory, and the error escaped the environment
test
> - The user got a server error instead of a result with checks
> - This pull request stages from a host temporary directory, sends the
remote path separately, and reports a staging failure as a check
> - The benefit is that the grok environment test always answers with
checks, and a workspace directory that does not exist no longer stops a
runtime preparation
## Linked Issues or Issue Description
No existing issue. The problem, in the bug report format:
**What happened**
The grok environment test failed with `Error: Workspace ignore scan
failed: git-ignore-scan-failed`. The route returned a server error, so
the user saw no checks at all.
**Expected behavior**
The environment test always returns `{status, checks}`. Every other
problem it finds (an invalid working directory, a command it cannot
resolve, a probe that times out) becomes a check with a level. A
credential staging problem must do the same.
**Steps to reproduce**
1. Configure a grok agent with a managed AI connection.
2. Point the agent at a remote (sandbox) environment.
3. Run the environment test for that environment.
**Paperclip version or commit**
Present on master. Both halves landed on 2026-09-12: the call site in
#13247, and the scan that it trips in #13353.
## What Changed
- `packages/adapters/grok-local/src/server/test.ts` stages the managed
account through a fresh host temporary directory and passes the remote
path as `workspaceRemoteDir`. This is the split `codex-local` and
`opencode-local` already use.
- A failure while staging becomes a `grok_environment_unprepared` check
with level `error`, instead of an exception that escapes the function.
The probe does not run after it, because there is no prepared
environment to probe.
- The temporary directory is removed in the existing `finally` block.
- `packages/adapter-utils/src/sandbox-managed-runtime.ts` treats a
`workspaceLocalDir` that does not exist as "nothing to sync". A
directory that is not there has no files for ignore rules to govern,
nothing to stage, and nothing to restore.
- Tests: the grok environment test now covers the managed-connection
remote branch, which had no coverage. `sandbox-file-sync.test.ts` covers
a preparation whose workspace directory does not exist.
## Verification
```
npx vitest run packages/adapters/grok-local/src/server/test.test.ts # 9 passed
npx vitest run packages/adapter-utils/src/sandbox-file-sync.test.ts \
packages/adapter-utils/src/sandbox-managed-runtime.test.ts # 90 passed
cd packages/adapter-utils && npx tsc --noEmit # clean
cd packages/adapters/grok-local && npx tsc --noEmit # clean
```
The two new grok tests fail against the old call site: the first asserts
the runtime never receives the remote path as its host workspace
directory, and the second asserts a staging failure becomes a check
instead of an exception.
## Risks
Low risk, and limited to environment preparation.
- The adapter change only affects the managed-connection remote branch
of one adapter's environment test.
- The `adapter-utils` change makes a preparation that used to throw now
continue with no workspace sync. A real workspace is unaffected, because
the directory exists in that case and the scan runs exactly as before.
- No migration. No API change.
## Model Used
- Claude Fable 5 (`claude-fable-5`), 1M context, extended thinking, run
through Claude Code with tool use and code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] I have updated relevant documentation to reflect my changes — no
documented behavior changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green — pending first run
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups —
pending first review
- [x] I will address all Greptile and reviewer comments before
requesting merge
|
||
|
Devin Foley
|
46bcc0863c |
test(runner): stop two CI-contention flakes that starve the canary lane (#13418)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - Each master commit is published as an npm canary, and the nightly release gate installs the newest canary and runs the onboarding smoke against it > - A canary only publishes after the Cloud readiness workflow passes for that exact commit > - Cloud readiness fails often, and almost every failure stops that commit from ever becoming a canary > - The nightly gate then keeps testing an old canary, so a fix that landed after it cannot reach the gate, and the gate stays red for reasons no longer in the code > - Most of these failures are short timeouts that only expire because the runner is loaded > - This pull request gives two such timeouts the budget the surrounding tests already use > - The benefit is that more master commits publish a canary, so the nightly gate tests current code ## Linked Issues or Issue Description No existing issue. The problem, in the bug report format: **What happened** Cloud readiness failed on 25 of 85 runs over three days (29%). 24 of those failures stopped the commit from publishing a canary. Two timeout clusters account for 10 of the 25: - `packages/paperclip-runner/src/control-plane/durable-prp-control-plane.test.ts` — `Error: Test timed out in 5000ms.` in "promptly fails the real transport request and notification paths on authenticated bad semantic input". Seen in runs 34652752663, 34654215468, 34656885170, 34657111560, 34661910120. - `packages/paperclip-runner/src/live/live-session.test.ts` — `Turn settled before the intentional SIGKILL: Error: Capability Codex turn timed out after 2000ms`. Seen in runs 34632264036, 34658026290, 34658106386, 34727991195, 34760283177. **Expected behavior** A commit that is good publishes its canary. A test fails because the behavior is wrong, not because the runner was busy. **Steps to reproduce** 1. Run the Cloud readiness workflow on master repeatedly. 2. Watch the Runner protocol and chaos jobs. 3. The two tests above fail intermittently, and the commit then publishes no canary. **Paperclip version or commit** Present on master. Measured over runs from 2026-09-11 to 2026-09-14, which is the workflow's whole history. ## What Changed - `durable-prp-control-plane.test.ts`: the "promptly fails the real transport request..." `it.each` now takes the same 15s budget as the test directly above it. It ran on the 5s default, although it builds a real control plane over a socket. The neighbouring test already asks for 15s, so this reads as an omission. - `live-session.test.ts`: the real-runner SIGKILL test gives the turn a 60s timeout instead of 2s. The test kills the runner process and then asserts the turn was still running at that moment, so the timeout is only a backstop. At 2s a loaded machine settles the turn first and the assertion fails. Neither change weakens an assertion. No test in this PR asserts that a timeout happens, and the tests that do assert one (`turnTimeoutMs: 500` and `turnTimeoutMs: 20`) are untouched. ## Verification ``` cd packages/paperclip-runner npx vitest run src/control-plane/durable-prp-control-plane.test.ts # 87 passed npx vitest run src/control-plane/durable-prp-control-plane.test.ts \ -t "promptly fails the real transport" # 2 passed (both it.each cases) ``` The live-session change is verified by the readiness lane itself, because that suite needs a real runner and a provider. ## Risks Low risk, and limited to test budgets. - A genuine hang in either test now takes longer to report: 15s instead of 5s, and 60s instead of 2s. - A real regression still fails, because the assertions are unchanged. - Timeouts are not a complete fix for this lane. Port races (`EADDRINUSE`), lease races, and lost runner VMs cause about another 40% of its failures. A retry on the `verify` matrix would cover all of those classes, and none of them reproduce on a second attempt. That is a larger policy change, so it is not in this PR. ## Model Used - Claude Fable 5 (`claude-fable-5`), 1M context, extended thinking, run through Claude Code with tool use and code execution. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [ ] I have updated relevant documentation to reflect my changes — no documented behavior changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green — pending first run - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups — pending first review - [x] I will address all Greptile and reviewer comments before requesting merge |
||
|
Devin Foley
|
e1d2e279a3 |
fix(db): replay a query whose socket write failed on a recycled pooled connection (#13417)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - A hosted instance keeps its data in Postgres, and it reaches that
database through a connection pooler
> - A pooler recycles the server side of a connection that sat idle. The
client still holds the socket and believes it is open
> - The next query fails while the driver writes it to that dead socket,
with `write CONNECTION_CLOSED host:5432`
> - The request that happened to draw the recycled connection fails,
although nothing is wrong with the query or the database
> - Only one code path guards against this today, so the same error
keeps reaching users and error reporting from every other path
> - This pull request replays a query when the write itself failed,
because those bytes never reached the server
> - The benefit is that a recycled connection costs one retry instead of
one failed request
## Linked Issues or Issue Description
No existing issue. The problem, in the bug report format:
**What happened**
Requests fail with `write CONNECTION_CLOSED <host>:5432` (driver code
`CONNECTION_CLOSED`). It is most visible on the first queries after an
idle period, when the pool holds connections the pooler already
recycled.
**Expected behavior**
A connection that the pooler recycled while it was idle is not a
user-visible failure. The client notices the dead socket and uses a live
connection.
**Steps to reproduce**
1. Run the server against a pooled Postgres endpoint (for example a Neon
`-pooler` host).
2. Leave the instance idle until the pooler recycles the server side of
the pooled connections.
3. Issue any request that queries the database.
**Paperclip version or commit**
Present on master.
## What Changed
- `packages/db/src/transient-write-retry.ts` (new) wraps the root
`postgres.js` client. When a query fails while the driver writes it to
the socket, the wrapper runs it again on a fresh connection. Three
attempts, 50 ms then 100 ms backoff.
- `packages/db/src/client.ts` gives Drizzle the wrapped client. The
teardown registry keeps the real client, because shutdown must end the
actual pool.
- The wrapper is deliberately narrow:
- It matches only the write phase (`code === "CONNECTION_CLOSED"` and a
message that starts with `write CONNECTION_CLOSED`). The write failed,
so the server never saw the query, and a replay cannot run anything
twice. That makes it safe for reads and writes alike.
- An error after the write propagates untouched, because the server may
have acted on the query.
- `CONNECTION_ENDED` and `CONNECTION_DESTROYED` propagate untouched,
because they mean a deliberate shutdown.
- Queries inside `db.transaction()` run on the scoped client that
`sql.begin()` returns, which the wrapper does not touch. A transaction
that loses its connection must abort, not replay.
- A query runs once however many handlers attach to it, and it still
executes lazily, like `postgres.js` itself.
## Verification
```
npx vitest run packages/db/src/transient-write-retry.test.ts # 7 passed
npx vitest run packages/db/src/client-teardown-registry.test.ts \
packages/db/src/client-options.test.ts # existing db suites pass
npx vitest run server/src/__tests__/cloud-tenant-transient-db-retry.test.ts # 6 passed
cd packages/db && npx tsc --noEmit # clean
```
New tests cover the replay, the `.values()` form Drizzle uses, the
attempt budget, an error that must not replay, and one execution per
pending query. A Drizzle round trip runs through a fake wire-protocol
server, which shows the wrapper is transparent to ordinary queries.
## Risks
Low risk.
- The retry only fires for a failure during the socket write, where the
server never received the query. A replay therefore cannot duplicate an
effect.
- A permanently unreachable database costs two extra attempts and 150 ms
before the same error surfaces.
- Transactions keep exactly their current behavior.
- Existing `retryOnTransientDbConnectionError` in the auth middleware
stays. It wraps a broader set of codes for one path, and it is
unaffected.
- No migration. No configuration change. No API change.
## Model Used
- Claude Fable 5 (`claude-fable-5`), 1M context, extended thinking, run
through Claude Code with tool use and code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] I have updated relevant documentation to reflect my changes — no
documented behavior or configuration changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green — pending first run
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups —
pending first review
- [x] I will address all Greptile and reviewer comments before
requesting merge
|
