mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-01 23:43:16 +01:00
b7f1c3769c00a107ffe4128b48b0ef8cc452cf3f
442
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 DottaandPaperclip
|
3ee4c5b613 |
Make saved execution receipt recovery react to durable work (#18409)
Reviewed and verified at
|
||
|
DottaandPaperclip
|
47f81e850b |
fix(testops): verify unexpanded result job runs (#18462)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Chat feature evaluations send their results to TestOps. > - The final reader checks the result against its current GitHub run. > - A real result omitted the nested job object and had no presentation links. > - The reader rejected that passed result after a successful upload. > - This change joins the embedded job-run ID to the separate job-run response. > - The reader keeps the source, case, status, and attachment checks. ## Linked Issues or Issue Description Refs #18394 and #18322. **What happened?** The native chat test and cleanup passed. TestOps stored a passed result and three attachments. The final read-only step failed with `chat_testops_upload_mismatch`. The result response had a job-run ID, launch ID, and current GitHub run URL. It did not include the nested job object. Its presentation links were empty. **Expected behavior** The reader must accept that response when the separate job-run response proves the exact job, launch, and current GitHub run. It must reject conflicting identity or source data. **Steps to reproduce** 1. Return a valid result with an unexpanded job run and empty links. 2. Return the matching full job run from the existing job-run endpoint. 3. Call `verifyChatUpload`. The old reader rejects the result. **Paperclip version or commit** Observed on |
||
|
DottaandPaperclip
|
ab51cabbbd |
test: await async outcomes and drive saved deadlines deterministically (#18416)
Reviewed and verified at
|
||
|
DottaandPaperclip
|
f2db909393 |
fix(ui): retain startup autosave through session recovery (#18412)
Reviewed and verified at
|
||
|
DottaandPaperclip
|
d5766e4074 |
perf(chat): scope task questions and batch inbox queue checks (#18422)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Standing chats show recent messages and current task questions. > - Recent-message pagination bounds history work, but it does not bound question work. > - The question read built the full company attention feed before it selected the chat's tasks. > - Each feed read also checked seeded queue membership with one insert per item. > - This pull request scopes question collection in the attention owner and batches queue membership checks. > - Chats keep their current question cards and initial layout while doing less work. ## Linked Issues or Issue Description Follow-up to #18358 and #18323. Related work: #18312. **What happened?** A hosted standing chat remained slow after recent-message pagination shipped. On commit `4eaa435bf4`, an embedded-browser reload made the conversation usable at about 8.6 seconds. Recent comments finished at about 1.9 seconds. The pending-question request held the initial reveal until about 8.4 seconds. That request called the complete company attention feed and then filtered its results. The separate badge summary also did broad work. **Expected behavior** Chat question cost should depend on the relevant tasks. Reading an existing queue must not require a separate insert for every member. Current access and decision policy must remain in force. **Steps to reproduce** 1. Create a standing chat with questions on tasks assigned to its agent. 2. Add many unrelated tasks and pending questions to the company. 3. Open the chat from a fresh application cache. 4. Compare recent-comment arrival with the pending-question response and content reveal. **Paperclip version or commit** Baseline: `909e29dd071b1aeeb24c2933af049da90a6066d3`. Its affected services match the observed hosted revision `4eaa435bf4`. ## What Changed - Add an internal task-question projection to the attention service. Filter pending interactions by company, task IDs and question kind before card enrichment. - Keep the same access, audience, dismissal, queue, snooze, retention, archive and training-example logic. Ranks are local to the returned question list. - Use the projection from the conversation receipt service. Keep current chat audience and session checks there. - Batch seeded membership inserts in groups of 500. Keep the queue lock, conflict protection, audit events and activity logging for each new membership. Existing members produce no new audit records. - Add real PostgreSQL regression cases and a reproducible paired benchmark. No schema migration or UI reveal change is needed. ## Verification - `pnpm -r typecheck` passed. The final server typecheck also passed after the queue change. - `pnpm build` passed. - Real PostgreSQL service coverage: 57 attention and chat-handoff tests passed. Seven queue-route tests passed, including concurrent reads, mixed existing/new items, duplicates, disabled seeds and audit preservation. - Red/green proof: the unrelated-attention regression fails with the baseline conversation implementation and passes with the scoped read. - `node scripts/check-module-boundaries.mjs` and `git diff --check` passed. - Full `pnpm test:run` was attempted and stopped with exit 130 after embedded PostgreSQL startup failures and 17 failures in the untouched native-session executor suite. It is not a local pass. The focused Docker/PostgreSQL results above completed before Docker Desktop became unavailable. A direct native PostgreSQL probe identified the host semaphore limit (`could not create semaphores: No space left on device`). No shared kernel settings or other test clusters were changed. [Full CI](https://github.com/paperclipai/paperclip/actions/runs/41162314995) passed on `2809f096b3fb2c01b44a96a1b82a27809c9359df`, including all eight browser shards. Final status: 55 successful checks/statuses, three skipped and one neutral. Greptile is 5/5 with no findings or unresolved threads. The fresh local browser retest is blocked because Docker Desktop cannot start and native PostgreSQL cannot allocate semaphores. The original hosted reproduction above is complete; no candidate browser pass is claimed. Local service benchmark: PostgreSQL 17.11, two relevant tasks, five alternating pairs per scenario, after warmup. These are local service timings, not a claim about deployed latency. | Fixture | Read | Baseline median (range) | Candidate median (range) | SQL calls | | --- | --- | --- | --- | --- | | 1,000 unrelated tasks/questions | Chat questions | 670 ms (632–1,905) | 12 ms (11–15) | 1,076 → 22 | | 1,000 unrelated tasks/questions | Company badge summary | 641 ms (622–717) | 167 ms (161–171) | 1,068 → 69 | | Small company | Chat questions | 38 ms (37–52) | 13 ms (12–16) | 52 → 22 | | Small company | Company badge summary | 34 ms (33–39) | 34 ms (32–40) | 44 → 43 | All question fields and ordering matched, except rank is relative to the scoped list. Summary fields matched except the generation timestamp. The interaction plan changed from 1,002 selected rows to two rows through the existing company/issue/status index. The benchmark loads baseline attention, conversation and queue modules together so the queue change cannot contaminate the baseline. ```sh pnpm --filter @paperclipai/server exec tsx ../scripts/benchmark-chat-questions.ts \ --baseline-ref=909e29dd07 --unrelated=1000 ``` ## Risks - Scoped and complete feeds must share policy. This change reuses the same attention collection and finalization path and tests dismissals, named audiences, snoozing, archives and private access. - Batched inserts must audit only new members. The concurrent PostgreSQL test checks unique membership and one event/activity record per new item across batch boundaries. - Full inbox summaries still collect the company feed. The small-company summary timing is unchanged. Other chat requests and transcript hydration can still contribute to hosted latency. - The first reveal still waits for question cards to preserve layout. This follow-up has not been deployed, so its hosted effect remains unverified. ## Model Used OpenAI Codex, based on GPT-6, with reasoning, repository tools, code execution and browser profiling. 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 #` 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
|
492d537340 |
fix(native): recover controllers across webserver replacement (#18343)
## Thinking Path > - Paperclip manages AI agents and their work. > - Native execution runs in a remote Runner while the webserver owns its controller. > - Encrypted PRP checkpoints and session snapshots already survive a webserver restart in Postgres. > - Ordinary warm and per-turn admission still needs local markers and supervisor metadata to find that state. > - Replacing the webserver with an empty native controller directory can therefore lose a usable retained session. > - This change adds bounded recovery metadata to the existing snapshot transaction and reconstructs local projections from the selected durable source. > - The benefit is native execution continuity across webserver replacement, with the existing ownership and agent lifecycle policies. ## Linked Issues or Issue Description **What existing behavior does this improve?** Native remote controller recovery and retained-session continuation during a webserver upgrade. **Subsystem affected** server/ — native runtime, checkpoint persistence, recovery scheduling and internal upgrade reporting. Product E2E qualification also changes. **Current behavior** Without both the local PRP marker and supervisor file, normal continuation can return before discovering the saved Postgres checkpoint. The snapshot already exists in Postgres, but its local reuse compatibility envelope cannot be reconstructed. **Proposed behavior** Select the existing durable source, validate its identity and ownership, restore its PRP state and supervisor projection, then use normal authenticated admission. Missing or inconsistent evidence holds recovery. It does not authorize fresh provider work. **Reason and benefit** A replacement webserver must recover active execution and later continue its retained session without the old server's native files. This is one component of deployment continuity. HTTP traffic switching, workspace storage migration and Railway rollout remain separate. **Breaking changes** No new public endpoint, schema migration or database pool. The private run-profile envelope is additive. Disk-independent rollback also requires a reader for the staged controller format and compact checkpoint part; the first production release containing these readers is the floor. Older file-backed sessions need an owner checkpoint or the bounded internal importer. **Additional context** Related foundations: #17401, #17916, #17254, #14621, #14622. The implementation contract and open acceptance gates are in [the dated plan](doc/plans/2026-10-30-native-controller-checkpoints.md). ## What Changed - Persist a versioned recovery envelope, at most 4 KiB, beside the redacted session snapshot in its existing transaction. Keep it out of public run projections. - Compare frozen admission separately from prepared attachment instructions during staged recovery, preserving exact template/identity/authority checks. - Bound the existing lazy checkpoint pool at three connections after measured two-slot burst backlog; keep request/renewal capacities and SQL operation counts unchanged. The tuning is a separate commit. - Discover and restore the exact selected retained source without local native markers. Preserve the 32-candidate window, progress barriers and current authority checks. - Fence snapshot and result writes. Acquire coordinator, issue and run locks in order and check expiry after lock waits. Combine snapshot ownership/read work into one statement. - Reuse existing recovery deadlines and drain/resume scheduling. Retained-owner waits preserve the original unstarted run and actual lease deadline. Stop, ownership and provider-progress fences precede deferral. Add bounded native evidence reporting, retained-source import and retention protection. - Admit a settled warm source with an authenticated idle goal-read suffix while retaining active-work and mixed-authority holds. - Permit lifecycle `turn.stop` checkpoints after accepted results under existing ownership fences. Claim the existing successor coordinator before retiring the old warm session, and stage shared attachments only after verified retirement. Preserve omitted isolated-home defaults and existing fresh skill-selection order during recovery without hiding real configuration changes. - Carry the exact durable six-field PRP binding during active recovery after a healthy warm handoff. Keep the Runner mismatch guard and preserve opaque attachment IDs. - Preserve an already verified active or idle warm process during remote preparation; revalidate its proof/liveness and retain later authenticated PRP attachment. Ordinary resumed leases retain suspension checks. - Drain retained source events under their original authority, reconcile immutable source accounting and block unknown source tool dispatch before target work. - Persist compact staged handoff intent through authenticated adoption and recover direct/listener command-loss windows without another Runner or duplicate attachment. - Rebuild suspended lifecycle receipts for resumed prepared tasks, and retry missing lifecycle reads after partial PRP materialization without overwriting contradictory evidence. - Add normal warm/per-turn diskless continuation coverage, real Postgres races and SQL-cost checks. Add a reproducible paired load workload. - Give only the explicit outage fixture a 15-minute provider idle window, with a red/green catalog regression preserving all production timing and lifecycle policies. - Publish the goal-less task-session pointer in the existing post-acceptance terminal checkpoint transaction, with current ownership, accepted-result, reset and newer-progress fences. Keep a private bound receipt so late completion cannot duplicate or resurrect publication. - Use a retained warm sandbox in the dedicated continuity fixture, and accept an exact output token followed by an explanation. Keep the strict identity oracle. - Extend the explicit Daytona cell to a five-minute active outage and another replacement after completion. Require retained provider identity, a new tool receipt and a useful follow-up result. Add an owned loopback Postgres E2E fixture for hosts that cannot start embedded Postgres. ## Verification **Merge recommendation — `fc68584ce6231a02e47ecd743f55b036ccea034f`: ready for the qualified native-controller component.** Full CI, fresh review, local application paths, controlled performance qualification and the strict live Linux/Daytona case pass. Independent resource cleanup is complete. Broader rollout, browser reconnection and the unexecuted transport/platform combinations remain integration gates. Integrated 484 passed full local recursive typecheck/build and ordinary active→completion→cache-loss→C component qualification. Fresh review then found a valid attached-wake failure: preparation appends attachment instructions before the verifier compares execution to the frozen admission. 9f4 fixes that by separately checking the pre-preparation admission and exact prepared execution. Both hashes are captured privately per admission; the full template, identity, configuration and authority checks remain. Its regression fails before the fix; 36 focused cases, 42 staged source/target cases and server typecheck pass afterward. The review thread is resolved. Exact-head full CI passes ([run 41168228822](https://github.com/paperclipai/paperclip/actions/runs/41168228822): 48 successful jobs, one intended skip); the complete `pnpm test:run` partitions are forced, with the five native suites in the Runner lane. This is not a claim that an unsharded local command passed. The fresh exact-head Greptile check succeeds with 5/5 and all six threads resolved. The original two-connection database comparison closes **steady PASS and burst FAIL**. Every relative CPU/WAL/request-p99 gate passes and all operations/ownership remain correct, but candidate arm 3 drains in 1,015.99 ms and baseline arm 4 in 1,059.67 ms against the unchanged 1,000 ms absolute gate. Those results remain failed. A separate 2→3→4→2 diagnostic reproduces drain times 1,051.57 / 956.79 / 882.02 / 1,065.72 ms: three is the smallest tested passing cap, and the repeated two-cap failure survives run-order/cache effects. Separately reviewable `fc68584ce6` changes the existing lazy checkpoint pool default to three connections. Request capacity stays ten and liveness two; idle release, shutdown fencing, transaction counts and pool count are unchanged. Thirteen DB lifecycle/cap tests and DB typecheck pass. Full recursive typecheck/build pass at exact |
||
|
DottaandPaperclip
|
322fa6b65c |
fix(ui): restore task document authoring and add persistence journey (#18356)
## Thinking Path > - Paperclip helps people manage AI agents and their work. > - Task documents keep versioned work beside each conversation. > - The default chat interface could read general documents but hid their editor. > - People also need unfinished edits to survive a tab switch or reload. > - This pull request restores the shared editor in Artifacts. > - A portable browser journey checks exact saved content and revision history. > - The benefit is a usable document flow with independent deployment checks. ## Linked Issues or Issue Description **What happened?** A person using the default task interface could read a general task document but could not create or edit one. The existing editor was available in Classic mode. A pending create could also leave newer text as a new form after its editor unmounted, causing a conflict against the person's own successful creation. **Expected behavior** A person can create and edit general documents through Artifacts. Titles, bodies and revision history persist after reopening and reloading the task. Unfinished text survives tab switches. A successful create makes retained newer text an edit of that confirmed revision. Later writes from another actor still cause a conflict. **Steps to reproduce** 1. Open an ordinary task in the default chat interface, then open Artifacts. 2. Try to create a general document. The previous interface has no authoring control. 3. With authoring restored, submit a new document, keep typing, and switch to Properties before the request finishes. 4. Reopen Artifacts. The previous draft path restores a new form despite the successful create. Related public work: #17598 adds document copy/download and #18157 adds image galleries. The selected actual merged parent is `f1d780838218f975064f38333946805e0ec103a5` from #18427. It includes the stable profile-entry ordering repair and the exact recovery-fixture prerequisite. The resulting PR diff contains 32 document and reporting files. ## What Changed - Reuse `IssueDocumentsSection` for general documents in default-chat Artifacts. Keep Classic authoring in its existing task body. Ordinary tasks, conversations and Coordinator mode share the same editor options and signed-in permission wrapper. - Preserve Plans, file-review documents, attachments, uploads and work products. Hide duplicate general-document cards. Use the same general-document key rules for visible rows, validation and refused writes. - Forward image uploads, mentions, references, author maps and feedback to both editor mounts. Keep empty mention/project values stable while queries load. - Back up unfinished forms within their signed-in account, task and editor scope. Preserve text entered before its key. Record a successful create on its matching retained newer buffer, including after unmount or an early reopen. Preserve newer title/body edits, real remote conflicts and matching cancellation cleanup. - Add the model-free `task-document-persistence` journey. Setup creates only an owned empty backlog task. The browser creates and edits the document; normal APIs independently check company, task, actor, exact content and both revision records. - Reopen the owned task through the explicit Backlog view, then reload it. Require exact UI readback and zero native runs. Add only the Doc registration, definition-hash inputs, report mapping and four closed checks to the shared harness. - Count explicit zero-run browser receipts after private journey metadata is removed for publication. Preserve missing billing evidence, positive usage and native/protocol accounting. - Update browser assertions for the restored inline editor while retaining exact download bytes, revision collisions, paused-state and no-run checks. Update the feature map, usage docs and shared editor stories. ## Verification - Latest focused checks use official Node 24.22.2: 568 journey, catalog, report, billing and attempt-accounting tests across 25 suites pass. All four admitted journeys remain registered. The seven relevant UI suites pass 529 tests, including the task-page container after the parent merge. Maps, token gates, the PR-scope diff check and the owned credential-pattern scan pass. The normal QA import preflight also passes with an empty environment and network access blocked. - Billing regression: before the repair, the explicit empty public receipt counted one run, and two positive-usage receipts were discarded by browser metadata. Those three cases fail before and pass after the repair. Two additional red regressions cover aggregate usage beside an empty run list. That ambiguous shape now keeps an unavailable receipt and a minimum count of one. Missing receipts retain unavailable billing, and native/protocol one-or-more run totals retain their measurements. - The previous fresh CI run failed its unchanged snapshot-only recovery fixture: one global sweep did not call its callback. The original failure remains recorded. An isolated PGlite probe with all canonical migrations exercises the real reconciler: it processes two earlier sources, then reaches the owned third source on the next sweep. It preserves source evidence, files and run IDs. This proves bounded pagination, not full native snapshot repair. - Local native recovery tests could not start PostgreSQL because the host semaphore limit was reached. That run executed zero cases. Canonical Linux CI later passed all 143 fixture cases, including the original snapshot-only case and the new competing-candidate control. The exact fixture repair is now inherited from #18427 and is absent from this PR diff. - The portable oracle rejects 19 stale, missing, foreign or incorrect outcomes. Result validation also rejects model/provider metadata and native runs for this case. - Five real Artifacts-to-Properties unmount regressions fail before the repair and pass after it. They cover both editor modes, reopening before and after completion, later typing, and a foreign revision that must still conflict. The edit-storage failure case also checks that Cancel clears the promoted fallback copy. - UI coverage includes default chat, Classic, conversations, Coordinator, Plan rendering, document-card deduplication, reserved keys, pending-query stability, permissions and rich editor options. - Earlier real browser checks passed five upload cases and two artifact-arrival cases. They retain exact CSV bytes, original/revised content, collision, paused-state and no-run assertions. Current-head normal CI runs the updated browser suites again. - The branch merges only the selected actual parent `f1d780838218f975064f38333946805e0ec103a5`. The merge has no conflicts or manual resolution. The SQL service and recovery fixture match that parent exactly. The own document editor, persistence flow, saved-state oracle and billing repair retain their previously qualified source blobs. Follow-up, clarification, question/resumption, screenshot helpers and the Steering map are byte-identical to the merged parent. Existing project/search/question/follow-up registrations remain present. - Required document source blobs remain exact: Doc UI `22e80663b053363e5f75e1432dd9043dc0e70243`, its unmount tests `16f3f4ed0dc9b11aaff129e06e6cf2bbf4522713`, readiness helper `800a747284db19b8484650b6759e8bd1f356839d`, and budget service `3a15d489f0aa55c22e8f7fe6a6bc6a0e14e5914e`. - Local typing is limited by borrowed declarations from another checkout. They omit current `skill_test` and sandbox creation-operation fields. No dependency declarations or repository config were changed. The authoritative clean CI Typecheck + Release Registry gate passes on exact head `2ffce29827ae170420d740940c1c069c7225a74c`. - Full monorepo typecheck, tests and build were not run locally. Borrowed build declarations from another checkout limit local typing. Authoritative [clean CI](https://github.com/paperclipai/paperclip/actions/runs/41171717008) passes on exact head `82d15946c5844d3ae41c6ade62fa224053f20a9a`: 55 successful checks, 3 skipped and 1 neutral, with zero pending or failed checks. Typecheck, build, Runner, server and browser lanes pass. Greptile gives that same head a fresh 5/5. All 15 review threads are resolved, with no further thread page. The PR is mergeable and clean. - Failure history remains recorded. The earlier recovery failure was repaired with a bounded, owned-source fixture. A later CI worker shut down after 2,375 Runner unit tests passed. Its one authorized narrow retry passed those tests and all five native integration suites. The separate AgentConfig browser failure was reproduced and fixed by the merged #18427 parent. These earlier passes do not replace current-head CI. - Final-source live qualification passed on exact harness and served application `82d15946c5844d3ae41c6ade62fa224053f20a9a`. QA run `41177696858` completed all four closed checks and all five phases with zero native runs. Cleanup passed; the durable manifest is complete with no pending evidence. The coordinator verified the original grade and manifest. - The [public qualification report](https://pages.paperclip.ing/evals-qa-smoke-41177696858-1/) records zero agent runs, empty run IDs and empty run usage. Pricing remains unavailable and billing remains incomplete; the report does not claim measured zero cost. - The earlier harness `36c8d9ddd76d0b7ddb28a9a6c674d9e538f6df88` pass against application `2ffce29827ae170420d740940c1c069c7225a74c` remains historical evidence. It is separate from the final-source measurement above. - The selected parent brings profile-entry ordering, proactive chat, a database migration, Runner updates, and task-page chat/read behavior. Its new task-page container is included in the final application measurement. The coordinator retains the original evidence and owns the primary merge and daily-plan admission. ## Risks Artifacts now mounts the existing shared editor. Its autosave, conflict, revision, annotation and server access contracts remain in place. Tests cover both interface modes, shared scopes, later remote writes and unavailable edit storage. Plans and generated file-review documents keep their existing flows. The PR adds no database schema or API change. The shared billing repair requires explicit captured empty run IDs and an unambiguous usage object containing only an empty run list for public browser-only receipts. Other fields beside an empty list keep conservative unavailable billing. Missing measurements remain unavailable. Contradicting positive usage remains counted. The inherited recovery fixture keeps the production sweep limit and its source, readback and no-agent-work assertions. The live journey must fail when authoring controls or saved outcomes are absent. It must not create the expected document through setup APIs or switch to another interface to pass. Final deployment verification must retain the actual app and harness revisions separately. ## Model Used OpenAI Codex, based on GPT-6, with tool use and code execution. The exact runtime model ID and context window were 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
|
8c9dd1e444 |
test(runner): qualify Kimi Code lifecycle and recovery (#18029)
Add the Kimi lifecycle and recovery qualification harness. Preserve original grades and cleanup evidence, strict native ownership audits, bounded startup and cleanup, and early rejection of unsupported local fault execution before credentials or resources. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
b866fa4bdc |
feat(runner): add pinned Kimi Code native runtime (#18028)
Add the pinned Kimi Code runtime, native instruction and model admission, scoped credentials, and pending-candidate recovery contracts. Retain current master release attestations and provider-pack compatibility. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
d06c5e0eff |
fix(evals): preserve opaque TestOps parameters (#18394)
## Thinking Path > - Paperclip verifies real agent work through Product E2E tests. > - The fixed Chat feature workflow uses TestOps to select and report one case. > - TestOps exposes job parameters as ID/value pairs. > - The schema does not establish that those IDs identify global environment variables. > - The reader used that unsupported relationship to infer parameter names. > - This pull request keeps IDs opaque, preserves the strict allowed values, and records closed read diagnostics. ## Linked Issues or Issue Description Refs: #18322, #18375, #18385. **What happened?** A zero-model bootstrap returned HTTP 404 during context readback. The exact failed GET was not retained and remains unknown. Schema review also found no established relationship between job-parameter IDs and global environment-variable IDs, although the reader queried that namespace to label parameters. **Expected behavior** Read only the existing job-run, launch, and job records. Preserve the job's exact permitted ID/value vector without inventing parameter names. Record which fixed GET class failed, its positive resource ID, state, and available HTTP status without raw responses or URLs. **Steps to reproduce** Use the actual API helpers with a job-parameter ID that has no global environment-variable definition. The new pure controls reject an unexpected definition lookup and verify the opaque vector and diagnostics without a backend or model call. **Paperclip version or commit** This change starts from `8595efc20bda3ce20a9fbbf34a5d156a6fcf6d57` after #18385. **Deployment mode** GitHub Actions metadata setup in the protected TestOps reporting environment. ## What Changed - Remove the unsupported global environment-variable lookup from job-parameter handling. - Preserve an exact positive ID and `refs/heads/master` value. Use the fixed workflow's existing `evaluate` default. - Treat an operation parameter as such only when its ID matches the protected configured ID and its value is allowed. Keep rejecting empty, duplicate, conflicting, or unknown vectors before writes. - Project only an opaque ID with a fixed-default-ref kind and value, or an unresolved kind with the value withheld. - Add closed diagnostics for the existing job-run, launch, and job GETs. A successful call keeps HTTP status null because the API abstraction does not expose the status. - Sync one reporter-test assertion and document opaque IDs. Reporter runtime and workflow source are unchanged. ## Verification - Four new actual-helper controls fail on the parent. - `node --test tests/runner-e2e/chat-feature-testops-api.test.mjs tests/runner-e2e/testops-chat-reporter.test.mjs`: 58 affected pure/mock/FS controls pass on Node 26.4.0. - Three MJS syntax checks and `git diff --check` pass. TypeScript noEmit was not run for this MJS/documentation change. - Two intermediate exact-projection assertion failures were retained before the assertions were synced to the new diagnostic contract. - [Normal Node 24 CI](https://github.com/paperclipai/paperclip/actions/runs/41137009579) passes on exact head `3c256fb8f7559093c5fb063170e26c01b6e8a0a4`: 48 jobs passed and one skipped. Both required `ci / e2e` and `ci / verify` pass. - Greptile reviewed this exact head at 5/5 with no actionable findings or unresolved threads. At this verification cut, the optional Contributor trust check was still in progress; it is not an effective required status context. - No actual CLI, API, provider, model, app, database, build, install, or additional bootstrap was run to validate this change locally. ## Risks The original HTTP 404 route remains unknown. Removing an unsupported schema assumption does not prove that it caused that failure. The live job mapping and dispatch remain unqualified. The strict source, actor, selection, run-number, nonce, once-only send, private custody, case, model, grader, and clock checks are unchanged. Unknown metadata still stops. This PR adds no endpoint, workflow mode, permission, credential path, or passing result. ## Model Used OpenAI GPT-6.1 Sol (`gpt-6.1-sol`) with reasoning, code editing, tool use, and pure test review. The runtime does not expose the 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 - [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
|
8ff812cdde |
feat(chat): deliver proactive task messages to agent chats (#18374)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Agent Chat gives each person a persistent conversation with each
agent.
> - Tasks and routines produce results outside that conversation.
> - Those runs need a safe way to bring a requested result back to the
person.
> - The server must select the agent's own chat and check the source
task's access rules.
> - This pull request adds proactive chat messages and unread
indicators.
> - People can see the result, open its source task, and reply in the
same chat.
## Linked Issues or Issue Description
**Subsystem affected**
Server, native runner tools, shared contracts, database, and UI.
**Problem or motivation**
An agent can report progress on a task or comment on an authorized task.
It cannot send a routine result to its own standing chat with the
responsible person. A morning review routine therefore cannot leave a
visible digest in that person's chat.
**Proposed solution**
Add `send_chat_message({body, idempotencyKey})` and `POST
/api/agents/me/chat/messages`. The server selects the company, agent,
and responsible person from the active run. It saves an attributed
comment and audit record in one transaction. The chat shows a
source-task link and per-person unread state.
**Alternatives considered**
A destination parameter on a generic comment tool would expose
unnecessary routing authority. A new delivery worker would add another
durable delivery path. This change uses a dedicated tool and the
existing comment and live-event paths.
**Roadmap alignment**
This extends the experimental Agent Chat capability in `ROADMAP.md`. The
existing `enableAgentChat` setting controls it. Related public work:
Refs #13282, Refs #13283, and Refs #14408. Searches for proactive chat
and unread chat found no duplicate PR.
## What Changed
- Add a focused messaging service shared by the native tool and HTTP
route. It checks current run ownership, execution mode, recipient
membership, task access, and task privacy.
- Reuse the canonical company/agent/person conversation. Save the
comment and audit record together. Keep task completion and chat
execution separate from delivery.
- Add source-task provenance and source-scoped retry keys. Identical
retries return the saved receipt. Changed content returns a conflict.
- Enforce the conversation's fixed agent at the shared comment-write
boundary. Keep normal reply cancellation and session checks.
- Add chat unread state, accessible agent dots, aggregate Chat
indicators, and visible-message read cutoffs. Refresh through existing
live events.
- Add the request-key constraint, native tool contracts, documentation,
capability maps, stories, and regression coverage.
## Verification
- Local delivery/security proof: all 38 proactive-message cases pass.
Eight credential-scope regressions and four quota regressions fail
against the prior source. The repaired delivery and canonical
cross-issue cap suites pass all 49 cases; all 31 OpenAPI tests, server
typecheck, and capability-map checks pass. The shared per-run cap covers
native and HTTP sends, mixed ordinary mutations, concurrent last-slot
contention, and retry/recovery without double charging.
- Credential tests preserve the authenticated person and authorizer;
reject bridge/test credentials and persisted skill-test runs after
task-mode changes; and recheck API-key revocation, scope, and
responsible person after destination locking. The earlier
delivery/native-authority suites passed all 74 cases before the four
quota additions.
- Four PostgreSQL contention cases cover native and HTTP sends plus
retries during a native chat mutation. All four deadlock against the
earlier source and pass with destination-before-source/run locking.
Recovery and human-reply delivery enqueue only the human comment; a
proactive digest never wakes the chat agent.
- Passed 108 companion conversation, native-authority, and OpenAPI
tests; 112 runner catalog/workflow-report tests; 47 completion-delivery
tests; and 25 project-handoff tests. Focused runner tests cover native
tool availability without generic API tools.
- All 327 affected UI tests, UI typecheck, token gates, and catalog
checks pass. The read hook retries at most three times, respects server
retry delay, coalesces only seen cutoffs, recovers on reconnect, and
cancels on scope changes.
- Real browser journeys with a deterministic process agent verify
routine-to-chat delivery, first-chat creation, unread state across
reload, the source link, secondary chat navigation, and a follow-up
reply referring to the digest. This does not qualify live GitHub
retrieval or model output quality.
- The Storybook showcase uses production components with deterministic
network fixtures. Rendered interactions pass for the landing, ordinary
read success, transient read recovery, and persistent failure after
bounded retries. Both Chat navigation variants update from two unread
chats to one while only Kodi's dot clears. Mobile and light variants
were reviewed. The existing Coordinator Recents label contrast finding
is unchanged; no Linux visual baseline archive exists.
- Current master's independent sandbox migration retains 0418.
Regenerated the feature migration as 0419 from the combined schema and
retained exactly the newest five snapshots; numbering/safety, replay,
and schema-drift checks pass.
- Workspace typecheck and build passed before PR preparation; canonical
server/UI typechecks passed again on the combined tree. Migration
`0419_slippery_impossible_man.sql` passes safety and schema drift checks
on disposable PostgreSQL, including interrupted index repair, concurrent
migrators, one-time owner read-state backfill, preserved receipts, and
two SQL replays. A default five-second local snapshot comparison timed
out; a diagnostic 30-second run passed without drift.
- Earlier failures are retained in the record: full local `pnpm
test:run` and broad shards encountered embedded-PostgreSQL startup,
timing, and provider-recovery failures. Thirteen selected
external-channel cases pass in isolation on the original and changed
sources; an unordered comment-row assertion now explicitly orders both
reads. The earlier route registry, native tool count, runner catalog,
and accessible Chat badge-name failures were fixed and passed full
remote CI on `920b3bb7db`.
- Repeated managed-skill server8 failures came from CI's newer synthetic
merge: its Cloud guard correctly rejected the old local-disk fixture.
Master independently repaired the fixture in #18385; this branch keeps
that repair unchanged. All 57 managed-skill and Cloud prerequisite cases
pass on the combined tree, and that fixture has no remaining PR diff.
- Current reviewed head `dc16f314739a6d977bed21cb1446d0e80a541578`:
Greptile 5/5 with no findings, Superagent security assessment passed,
advisory Paperclip Review 5/5 with no findings, and zero open review
threads. All 55 current-head check runs pass, with five expected skips
and the advisory review neutral at 5/5; Snyk also passes. No checks are
pending or failing. Protected squash auto-merge is enabled and is
waiting for required code-owner approval. A normal exact-head merge was
refused by the base-branch policy; no admin bypass was used. The
previous head's Superagent quota finding is fixed and covered above.
## Risks
- Chat remains company-visible. Delivery rejects a source task with
private-task, project, or ancestor restrictions.
- Read state uses a comment timestamp cutoff. The server retains
database precision and never moves a read receipt backward.
- The new unique constraint changes comment query planning. An existing
regression now uses an explicit row order. The migration builds the
unique index concurrently, then takes a brief table lock to attach it.
It uses the existing top-level migration and interrupted-index recovery
path.
- The upgrade backfills missing chat-owner read-state rows at the latest
existing agent comment. Existing read receipts remain intact. New chats
can still be unread before the human participates.
- Quota observation commits atomically with delivery. A rejected send
rolls back its attempted denial audit row and any first-chat creation;
the canonical warning remains in server logs, and previously spent quota
remains enforced. Identical/conflicting retries and receipt recovery
spend no extra slot.
- This version delivers in-app only. It adds no external-channel, push,
or email delivery.
- Local full-suite acceptance remains limited by the failures recorded
above. The feature stays behind `enableAgentChat`.
## Model Used
OpenAI Codex. PR preparation and CI work used `gpt-6.1-sol` with tool
use and code execution. The earlier implementation also used Codex, but
its exact model ID was not recorded in the visible session. The
context-window size was not provided.
## 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 focused tests locally and they pass; the broad local
failures are recorded 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>
|
||
|
github-actions[bot]andlockfile-bot
|
4eaa435bf4 |
chore(lockfile): refresh pnpm-lock.yaml (#18414)
## Thinking Path > - Paperclip uses a dependency lock for reproducible builds. > - Dependency changes on master need matching lock entries. > - Paid eval reports also pin a reviewed offline correction. > - Reports must install before paid model work starts. > - This automated PR keeps the lock and report proof together. ## Linked Issues or Issue Description **What happened?** Dependency manifests on master changed after the last committed lock refresh. **Expected behavior** Frozen installs must match those manifests. Eval reports must retain matching offline lock proof. **Steps to reproduce** Run a frozen install before and after this lock refresh. ## What Changed - Refresh the dependency lock from the manifests on master. - Include the matching offline report correction and hashes. ## Verification - Pinned pnpm resolution completed with lifecycle scripts disabled. - The report proof refresh accepted the reviewed committed baseline. - The workflow rejected changes outside the lock and report proof files. ## Risks The lockfile bot merges this PR without CI or review when every added package version is at least 3 days old on npm. Otherwise the PR waits for a maintainer, and the daily refresh tries again. The bot merges only the exact commit this workflow generated from manifests already reviewed on master. Unknown report lock baselines remain blocked. This update changes no permissions or paid enable gates. ## Model Used None — generated by pnpm and the deterministic report-lock updater. ## Checklist - [x] Thinking path and problem description included - [x] Generated lock and matching report proof included - [x] Release-age gate decides between bot merge and maintainer review Co-authored-by: lockfile-bot <lockfile-bot@users.noreply.github.com> |
||
|
DottaandPaperclip
|
1353a232f2 |
test(e2e): add a portable active follow-up journey (#18345)
## Thinking Path > - Paperclip lets people direct agents while their work is active. > - Product E2E checks must prove that saved user input reaches the intended work. > - The existing active chat story depends on a file in a local agent workspace. > - That setup cannot qualify an attached deployment with remote agent execution. > - This change adds one shared active follow-up journey through the normal task UI and API. > - Independent native operation, saved input, final reply, and cleanup checks make the result inspectable across targets. ## Linked Issues or Issue Description **What existing behavior does this improve?** Product E2E coverage for a user follow-up sent while a native agent is working. **Current behavior** The classic chat story holds work open with a local file. The shared deployment-targeted harness has no active follow-up journey. A running heartbeat alone cannot prove that work was active when input arrived. **Proposed behavior** Run one bounded command through the real native provider. Observe its active execution before and after one follow-up is saved. Require one final reply with the late requirement, correct task and agent identity, an empty input queue, and visible readback after reload. Use the same story on managed and attached targets. ## What Changed - Add a shared `followup-while-running` journey with normal task creation and reply submission. - Require a bound native turn and pending shell execution at the input boundary. Retain a closed receipt with source identity and a row hash. - Accept one steered work run or one normal successor (1–2 native runs). Preserve every other journey's declared run contract. - Reuse the saved native final reply observer, run binding checks, usage collection, target lifecycle and private evidence path. - Accept canonical builtin/process execution events independent of provider tool display names; calibrate against real OpenCode, Codex and ACP mappers. Reject Paperclip semantic tool names across native wrappers/prefixes and unnamed or ambiguous builtin calls. - Recognize only the observed legacy PRP v1 builtin/bash/unknown/readOnly=false command shape and fail promptly when owned work finishes before a live boundary can be observed. - Ready the task page/composer before the observed command boundary, use a bounded 90-second command, and preserve active-after-persistence proof. - Add adversarial oracle and retained-result calibration. Register the journey, include its sources in the definition hash, and document its coverage. - Keep six real browser document-readiness checks and verify shared-deadline behavior with a deterministic clock/state test that rejects timeout-reset mutations. ## Verification Exact current head `e72e20ad111bd94728d38606015d1fdac3c92929` passed [full CI run 41142552195](https://github.com/paperclipai/paperclip/actions/runs/41142552195): all 49 workflow jobs reached accepted terminal results, including typecheck, build, every test lane and aggregate verify. All 59 reported checks are terminal: 55 success, three expected skips and one neutral 5/5 Paperclip review. Snyk passed. [Fresh Greptile review](https://github.com/paperclipai/paperclip/pull/18345#issuecomment-6528913235) explicitly names this exact head and gives 5/5. No review threads remain open. The PR is mergeable. These source gates do not establish a live QA pass at this head. Fresh protected QA at this exact source and the separately recorded deployed application/runtime remains required before primary merge and daily admission. Historical live QA staging **PASS** on qualified commit `af9e091b15b5e24b6b2cc1b424523430cc2ab024`: 6/6 checks, all five phases including cleanup, two native work runs, and the configured model/environment check passed. The [sanitized qualification report](https://pages.paperclip.ing/evals-qa-smoke-41122620146-1/) records that exact measurement. Final integration head `e72e20ad111bd94728d38606015d1fdac3c92929` merges exact actual squash `18b1637b6371a6a6adade857e2a73654ee626fef` from [#18254](https://github.com/paperclipai/paperclip/pull/18254) into previous head `e54d64743964b1a19d165537bebbe62c50389a23`. The merge had no conflicts. All 18 files in this PR's change set, the complete 45-file shared journey directory, and six shared native/evidence helpers retain their exact previous-head Git blobs: 59 scoped files in total. The production clarification helper remains blob `800a747284db19b8484650b6759e8bd1f356839d`. Follow-up 1–2 runs, question 1–3 runs, older exact counts, owned native event checks, active-before/after persistence proof, cleanup/accounting and feature mappings remain unchanged. Source integration head `e55d74491f5bcd99c1455e9ace02bd7ad7319fae` integrates exact actual squash `8595efc20bda3ce20a9fbbf34a5d156a6fcf6d57` from [#18385](https://github.com/paperclipai/paperclip/pull/18385). The merge had no conflicts. All 17 files in this PR's change set and six shared native/evidence helpers retain their exact Git blobs from previous head `162e341cda5b627639b486b4019a1675bb018ea2`. This preserves the follow-up and question readiness checks, feature mappings, six closed follow-up check IDs, immutable before/after active-operation proof, follow-up 1–2 runs, question 1–3 runs and older exact counts. Previously reviewed head `e54d64743964b1a19d165537bebbe62c50389a23` adds a separate test-only reliability repair on top of that source integration. It replaces one scheduler-sensitive browser case with a deterministic Vitest clock and locator-state model that calls the actual production helper. The six other real Chromium checks remain byte-for-byte unchanged and pass **6/6**. Focused unit suites pass **51/51**, including the deterministic deadline case. This is six browser checks plus focused unit coverage, not seven browser checks. The model makes the panel ready at 100 ms, the supplied answer ready at 600 ms, and full valid content ready at 1,300 ms. It requires rejection at the shared 1,000 ms deadline and records immutable rejection state before simulating a 10-second scheduler stall. Both restarted-timeout mutations fail: resetting all waits wrongly resolves at 1,300 ms; resetting matcher timeouts while keeping the final real deadline guard rejects only at 1,300 ms with valid content present. The genuine helper rejects at 1,000 ms with the answer still incomplete. The production helper is unchanged, Git blob `800a747284db19b8484650b6759e8bd1f356839d`. That two-file test-only patch changed no journey, application, run, accounting, model or runtime behavior. Its six-browser/51-unit results are prior-head evidence; the complete patch is unchanged in the new integration. The final source also retains the actual merged question-primary squash `d9354bbfc35ef74333e90e596ce21dc8e59149b9` from [#18346](https://github.com/paperclipai/paperclip/pull/18346). Later commits `409117451da260a03fc28a3107c7808ecbe68f8f` and `ea3e7b1d1bed562351fcf4722c7175001b987b59` bound question document readiness to one campaign deadline and test that waits cannot restart it. Commit `162e341cda5b627639b486b4019a1675bb018ea2` adds question/follow-up report feature mappings and coverage. These files are unchanged by the final merge. The imported upstream source has real application changes: [#18334](https://github.com/paperclipai/paperclip/pull/18334) makes managed skill content durable and extends native skill-test contracts; [#18350](https://github.com/paperclipai/paperclip/pull/18350) requires managed Cloud shared storage and bootstrap prerequisites. #18385 updates TestOps job metadata and repairs their combined maintenance fixture. The new merge additionally imports [#18383](https://github.com/paperclipai/paperclip/pull/18383) at `81ecaf0654fe79a2a2dbda6588f033e39ee610e4`, which bounds failed lifecycle verification; [#18357](https://github.com/paperclipai/paperclip/pull/18357) at `45a3a92e6dc52d20923b9aa34a22fee2ec5443f9`, which persists sandbox-creation ownership and updates cleanup/adapter/plugin contracts; and #18254 at `18b1637b6371a6a6adade857e2a73654ee626fef`, which accepts a valid late Antigravity authorization URL while native authentication drains. This PR's diff against exact upstream `18b1637b6371a6a6adade857e2a73654ee626fef` contains only journey tests, report coverage and documentation. Harness source and deployed application/runtime identity are separate inputs. This source integration does not deploy the imported application changes or claim the QA target runs this application revision. This final SHA has not been measured by the historical run above: **fresh QA at this exact final head is required before primary merge**, followed by normal TestOps verification/admission. Fresh exact-head CI and review have passed as recorded above. The live measurement must record the actual separately deployed QA application/runtime identity. Seven whole-file blobs remain identical to the qualified SHA. All paths are relative to `tests/runner-e2e/`: | Journey/helper | Unchanged Git blob | | --- | --- | | `product-journeys/active-followup-evidence.ts` | `83a5d9fab52abb9849a40e83eca6b6a9c29bf587` | | `product-journeys/active-followup.test.ts` | `728160c610c18292d2cf06d4a87d25a8719f1734` | | `native-task-reply.ts` | `59d11f1a178dad9db3da8fa2342ff93643cb9922` | | `native-event-envelope.ts` | `305b0b456c59c511ee328e529ad574c0583075a0` | | `run-observations.ts` | `08c8ad184a821a7cd536926b3d7e3d1156a1e3da` | | `user-actions.ts` | `e9611348dc41e17899367dab2e3b36d3d1be9480` | | `stop-redirect-flow.ts` | `4580bd605a425bc72511e02d988492acbec8d60e` | The follow-up preservation comparison against historical `af9` has three explicit exceptions; no whole-file identity is claimed for them. The separately documented question readiness and report additions above also require final-source verification: 1. `product-journeys/agent-flows.ts` adds `question-answer-resume` to the `AgentJourney` type exclusion, alongside follow-up. Its remaining contents and runtime helper bodies are byte-for-byte unchanged. Qualified blob `a0afaa968c1d854476159fef21de7b6dd50a4a98`; final blob `0cc797c0ae594e7769338e63a80c4f67e29e4a54`. 2. The common run-count helper is centralized in `success-contract.ts` and re-exported by the catalog. It accepts follow-up **1–2**, clarification **1–3**, and each older paid journey's exact count. The follow-up catalog entry, maximum of two runs, six required checks, native dispatch, collection and retained-run contract remain byte-for-byte unchanged. The helper's text differs; its preserved count semantics are verified explicitly. 3. `product-journeys/active-followup-flow.ts` changes exactly one composer readiness assertion from Playwright's implicit five-second default to an explicit 30-second timeout, matching `submitTaskReply` for attached targets. Prompts, actions, native evidence, before/after persistence proof and grading otherwise compare byte-for-byte unchanged. Qualified blob `64392329198a1a4d9828e32834d4507e175216e6`; final blob `2746b5ba89e2af26b4097ba7897913186786ac5d`. This correction is why a fresh final-head QA run is required. - **310 focused tests passed at current final head `e72e20ad111bd94728d38606015d1fdac3c92929`**: 286 journey/evidence/catalog/report/shared-deadline tests in 14 files and 24 native authentication/real-child stream-order tests in two files. Both invocations used Node24 and one Vitest worker. The earlier `e55` integration also passed 365 focused tests, including 40 Cloud configuration and 40 TestOps metadata cases; those results remain prior-source evidence. - **157 explicit count assertions passed again at the current final head**, including both interactive ranges, older exact counts, invalid/fractional/non-number values and catalog/contract export identity. The maximum bounds and six closed follow-up check IDs also matched. - Maps and diff checks passed. The official QA runtime-import preflight passed on Node24 at the current final head without network access. It uses a dummy adapter and no credentials. It does not prove target connectivity, sign-in, model readiness or deployed application/runtime identity. Those require the normal protected QA run. No model/provider work or optional Runner release was dispatched. - Current local runtime dependencies use ignored isolated workspace-source views. No shared dependencies or compilers were changed. A previous local typecheck attempt hit stale reused Runner declarations. No full repository typecheck, test run or build was repeated locally; the successful exact-head full CI run above now supplies these gates. - CI on previous head `162e341cda5b627639b486b4019a1675bb018ea2` failed the Cloud/local maintenance fixture in [server shard 9](https://github.com/paperclipai/paperclip/actions/runs/41133716802/job/124331643100), plus dependent verify. The actual tested merge combined that head with master `198218bf554877fe3b7a83742eefb3b8f12392c6`. A Cloud child selected local-disk storage and hit the new shared-storage policy. This was a deterministic upstream fixture mismatch; no unchanged failed-job retry was used. - The actual #18385 repair configures a valid Cloud child, verifies local bytes independently, checks the expected provider mismatch and retains exact stored-config readback. Its own final head `b76d5dd40d7711f1b5f6b6316b1114c0aae02d57` passed 59 checks, including [server shard 9](https://github.com/paperclipai/paperclip/actions/runs/41133876636/job/124331450635). That prior result supports the repair; it does not replace this final head's fresh CI. Separate repair candidate #18392 and its source/evidence are retained as superseded history. - A local attempt to run the database fixture could not start PostgreSQL because the host exhausted macOS semaphores. A first setup attempt also failed package resolution before test execution. Skipped database cases and these setup failures are not passes. No kernel settings, shared processes or databases were changed. - The complete previous-head CI run [41138618869](https://github.com/paperclipai/paperclip/actions/runs/41138618869) finished with one underlying failure: [Runner Vitest shard 1](https://github.com/paperclipai/paperclip/actions/runs/41138618869/job/124347859445) expected native Antigravity login success but observed failure at `antigravity-auth.test.ts:101`. The lane had one failed test, 1,994 passing tests and eight skipped tests. Aggregate verify failed only because of this lane. All other 57 checks were accepted, including build, typecheck, all server/browser suites, Runner shard 2, Snyk and exact-head 5/5 reviews. No unchanged retry was requested. - The independently reviewed #18254 repair removes an invalid cross-pipe ordering assumption: a valid URL on stderr can arrive after the authenticate response on stdout during drain. Its real-child regression forces both orders. The current final source imports that actual squash; the 24 focused authentication cases pass here. Fresh full CI and exact-head review have now passed for this source, as recorded above. The earlier failed result is not relabeled. - The original live compatibility failure is retained: an older PRP v1 pack labeled builtin `bash` as operation `unknown`, causing the observer to miss the command and never send the follow-up. Its exact shape was reproduced red before repair. Generic unknown operations, semantic tools, wrappers, read-only work and foreign provenance remain rejected. The historical qualified pass used the repaired oracle; no failed result was relabeled. - Earlier CI on `6fad965d69` hit an unchanged heartbeat test's three-second adapter-call wait; its original failure and one narrow rerun remain recorded. The qualified SHA subsequently passed CI. An earlier report-catalog subprocess failure is retained. The final focused Node24 run now includes that suite and passes after the documented report feature mappings. ## Risks The fixture asks the native agent to run one bounded `sleep 90` command. The observer must see actual active native execution; a fast or noncompliant turn fails rather than passing from elapsed time. A completed result requires one or two bounded native work runs, source attribution, configured model/environment, and collection. Unknown provider prices stay unknown. This change does not alter production execution or grant new access. ## Model Used OpenAI Codex, GPT-6 family, with reasoning and code execution. The host does not expose an exact underlying model version 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 available 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; existing follow-up execution fixes do not provide this deployment-targeted test - [x] I have described the issue in-PR following the enhancement template - [x] I have not referenced internal/instance-local Paperclip issues or links - [x] My branch name describes the change and contains no internal ticket or instance-derived details - [x] I have run the new focused tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation - [x] I have considered and documented risks - [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 merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
8595efc20b |
fix(evals): support official TestOps job metadata (#18385)
## Thinking Path > - Paperclip manages AI agents and verifies real work through Product E2E tests. > - The fixed Chat feature workflow uses TestOps to select and report one backlog case. > - Official registration exposes a workflow-query URL and a Branch parameter. > - The existing reader accepts only workflow-file or numeric URLs, and the admin launcher requires an operation parameter. > - These assumptions reject valid official metadata before the feature can run. > - This pull request supports the exact observed form while preserving source, selection, and once-only execution checks. ## Linked Issues or Issue Description Refs: #18322, #18375. **What happened?** The second zero-model bootstrap stopped with a generic job-mismatch code. Ordinary TestOps UI inspection then showed an official workflow-query URL outside the accepted literals and a Branch-only job. The exact first failed predicate in that bootstrap remains unknown. **Expected behavior** Accept that exact fixed-workflow URL. Preserve the existing master Branch value when the job has no operation parameter. Keep rejecting foreign URLs, unknown parameters, wrong branches, and conflicting operation IDs before mutation. **Steps to reproduce** Pass the official query URL and Branch-only metadata through the existing bootstrap, selection, and admin-launch owners. The new pure controls exercise those actual helpers without a CLI, backend, or model call. **Paperclip version or commit** This change starts from `0e15c92073a09e50df99190a37f69a86286b989b` after #18375. **Deployment mode** GitHub Actions and the protected TestOps reporting environment. ## What Changed - Share one exact URL predicate across bootstrap, evaluation selection, and admin preparation. Add only the observed official workflow-query literal. - Support Branch-only, operation-only, and Branch-plus-operation jobs with strict definition IDs, names, values, and duplicate checks. - Preserve the exact master Branch value. Set an existing operation parameter to `evaluate`; otherwise use the fixed workflow's existing default. - Use the same checked parameter vector for job and plan writes and their readbacks. - Report distinct fixed bootstrap predicate codes without response values. - Document when the protected operation-parameter ID is required. - Align one existing Cloud-maintenance test with the inherited managed-Cloud S3 policy. Preserve the same-version local-storage proof in the parent, use valid synthetic Cloud settings in the child, and check the expected provider mismatch without S3 requests. Retain the no-seed and unchanged persisted-row assertions. ## Verification - URL controls: three failures on the parent; all pass with the correction. - Branch controls: two failures and one correct rejection on the predecessor; all pass with the correction. - `node --test tests/runner-e2e/chat-feature-testops-api.test.mjs`: 40 owning pure/mock controls pass on Node 26.4.0. - Two MJS syntax checks and `git diff --check` pass. TypeScript noEmit was not run for this MJS/documentation change. - The first [normal CI run](https://github.com/paperclipai/paperclip/actions/runs/41131094974) failed only in the managed-skill Cloud fixture and its dependent verification gate: 46 jobs passed, two failed, and one skipped. The tested merge included a new upstream policy that rejects the test's local-disk Cloud configuration. The exact caught runtime exception remains unknown. - Three pure controls using the actual policy validator pass: the old invalid Cloud fixture rejects, the new synthetic S3 fixture admits, and an invalid master key rejects. An initial tsx wrapper IPC permission error executed zero controls and remains a separate setup failure. - The PostgreSQL/maintenance-child test was not run locally. On new-head AWS CI, `managed-skill-content.test.ts` passed all 17 tests in server shard 9. - [Normal CI](https://github.com/paperclipai/paperclip/actions/runs/41133876636) is green on exact head `b76d5dd40d7711f1b5f6b6316b1114c0aae02d57`: 48 jobs passed and one skipped. Both required `ci / e2e` and `ci / verify` pass. - Greptile reviewed this exact head at 5/5 with no actionable findings or unresolved review threads. - No actual CLI, backend, model, provider, app, database, build, or bootstrap was run to validate this correction locally. ## Risks The exact original bootstrap predicate is still unknown. This fixes a demonstrated metadata false negative and preserves the prior failed run. Actual backend parameter types and integration dispatch still need verification. Unknown metadata stops before the once marker and writes. This PR changes no workflow, credential, case, model, grader, clock, or private-evidence policy. It does not claim a passing feature result. ## Model Used OpenAI GPT-6.1 Sol (`gpt-6.1-sol`) with reasoning, code editing, tool use, and pure test review. The runtime does not expose the 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 - [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
|
0e15c92073 |
feat(skills): make managed content independent of web-server disk (#18334)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents use installed skills and exact version pins during execution. > - Managed skills still depended on mutable directories on the web server. > - A replacement server could lose acknowledged edits or remove an unassigned skill whose directory was absent. > - This pull request makes immutable versions the authority for managed content and installed repository reads. > - The server can rebuild delivery files from durable content after replacement. > - This establishes skills readiness for ephemeral web servers. General server draining and workspace recovery remain separate work. ## Linked Issues or Issue Description **Subsystem affected** Skills services, shared contracts, database receipts, Skill Studio, and runtime delivery. **Problem or motivation** A fresh web server must discover, read, edit, and deliver protected installed skills using the same database, object storage, secrets, and supported application artifacts. It must not need its predecessor's files or process memory. **Proposed solution** Use `managed` ownership with a null source locator. Read installed content by immutable version. Prepare packages before a short publication transaction. Persist execution selections before materialization. Recover exact pinned runtime assets and reject deletion while active or recovering pins exist. **Alternatives considered** Reconstructing a mutable installed directory preserves the old source-of-truth ambiguity. Permanent dual writes create two competing authorities. This change uses one managed model after a bounded legacy conversion. **Roadmap alignment** This improves the existing Skills Manager, Studio, version history, and Cloud execution capabilities. The maintainer supplied the implementation contract. This does not add a separate skills product or deploy fleet changes. Related work: #18082 supplies immutable package storage. #17273 addresses edits outside the API for linked local skills; this PR keeps that source category separate. #2648 is a closed earlier ephemeral-filesystem proposal. ## What Changed - Add managed ownership and focused content publication and execution materialization modules. Keep the archive format and package storage tables. - Publish managed create, save, delete, fork, rename, restore, and catalog content with version checks. Add durable create receipts. Preserve file-save receipt responses. - Return file version identities. Keep Studio drafts after a conflict and provide an explicit reload action. A delayed reload cannot overwrite newer typing or navigation. - Persist exact runtime selections, including empty selections. Rebuild pinned assets and verify their admitted digests on a replacement server. - Preserve source access on forks and disconnected copies. Keep the existing workspace policy category for managed skills. - Add bounded conversion, eligibility, and independent verification phases. Load persisted Cloud configuration in the standalone CLI without seeding or changing it. Reject partial current versions against installed inventories; recover only complete saved content or the exact source. Document storage migration blockers and the compatible rollback floor. - Add separate-process PostgreSQL/S3 qualification and update domain, retention, policy, and UI regressions. - Pass the plugin worker manager into Studio execution, admit restricted `skill_test` mode through native Runner contracts, and preserve admitted scope across native and REST tools. Native tests cannot create ordinary tasks, browse company data, or use external tools. - Select the exact durable Studio revision even when unassigned or Beta skills is off. Persist Studio output and audit atomically with native task completion, preserving the result on replay. ## Verification - Final revision `13e8af480b88ba6fea5737bac4bd81104d604792` passed full CI: all 49 workflow jobs completed successfully ([run 41127396098](https://github.com/paperclipai/paperclip/actions/runs/41127396098)). This includes typecheck, test suites, Runner checks, browser checks and build. Greptile review 15 is 5/5 on that same revision, with zero unresolved threads. The PR is ready for review. - Local workspace typecheck and build passed after integrating master. Focused checks passed: 33 execution/completion cases, 52 Runner HTTP cases, 16 managed-content cases, 17 Studio cases, the new real-process CLI regression, map checks and UI token gates. The CLI regression proves inspection does not seed missing Cloud configuration. Server typecheck passed after that fix. - The independent-process PostgreSQL/S3 matrix passed through normal HTTP skill routes. It covers A → B → C replacement, lost commit responses, upload crashes, concurrent saves, binary contents and executable delivery. Metadata listing and warm selected delivery retain zero object downloads. Local timing measurements are recorded; they do not establish hosted latency. - Hosted A → B → C passed on the owned disposable test tenant with external Postgres, S3 and a real remote Paperclip Computer. Predecessor temporary homes were absent. C read and executed B's acknowledged managed save. New runs used repository v1 while v2 was installed. Hashes matched an independently checked exact upstream commit. - Compatible image rollback to `6518b323` passed. Studio retained the unassigned fork's C save. A fresh remote run executed the newer managed content and historical repository pin. The actual shell event confirmed expected hashes, stdout, exit 0 and mode 0555. - Runtime candidate `9200e3fcb4` deployed successfully ([Cloud run 41124147876](https://github.com/paperclipai/paperclip-cloud/actions/runs/41124147876)). The final Studio test delivered only the unassigned fork's exact v3, executed its script and persisted a successful output snapshot. Independent command-event inspection confirmed all three hashes, stdout, exit 0 and mode 0555. All 14 installations were version-backed. Independent verification checked all 25 versions: 23 remote objects, two bundled database snapshots, zero blockers. - The deployed image is `9200e3fcb4`. Later master integration, the client-only delayed-reload draft guard and the CLI bootstrap have separate automated proof. The CLI bootstrap was also exercised against the tenant's stored configuration and storage through the same read-only services. This PR does not claim that the deployed image is the final PR revision. - Live Studio checks exposed worker wiring, admission, access scope, exact selection and completion-publication gaps. They are fixed. Failed attempts remain in the evidence. The full local suite exceeded 80 minutes under host contention and was stopped. Later combined/isolated local attempts also timed out; they are not counted as passes. Full CI covers the integrated suites. Earlier CI found an obsolete restart-scope fixture, which was corrected and passed all 52 HTTP cases. - Reproduction and migration: [`doc/skill-ephemeral-web-servers.md`](doc/skill-ephemeral-web-servers.md). Source revisions, deployments, versions, run IDs, hashes and failure history: [`doc/plans/2026-10-30-ephemeral-web-skills.md`](doc/plans/2026-10-30-ephemeral-web-skills.md). ## Risks - Legacy conversion requires stopped admission and retired filesystem-authoritative writers. Pre-conversion binaries are not supported rollback targets afterward. - Cloud readiness requires remote object storage. Local provider receipts and provider mismatches remain blockers until objects are copied and verified. - Linked local sources retain filesystem semantics and are outside the cloud guarantee. - Cold reads still fetch the existing archive format. Warm runtime preparation retains the verified-cache path. - Bundled reconciliation preserves an installed head. Existing release snapshots remain selectable. An older server must not replace a newer head with its shipped files. - No production volumes, fleet defaults, or general deployment coordination change in this PR. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, repository tools, code execution, and browser use. The exact backend model ID and context-window size are not exposed in this session. No subagents were used. ## 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
|
d9354bbfc3 |
Add a portable agent question and answer resumption journey (#18346)
## Thinking Path
> - Paperclip helps people manage agent work and keep human decisions
attached to tasks.
> - Product E2E journeys verify the browser, native agent, saved state,
and cleanup.
> - The classic suite already checks a choice question followed by a
text question.
> - Deployment-targeted journeys need the same behavior on an attached
instance.
> - Production questions can use provider input, the dedicated semantic
tool, or the authenticated native API fallback.
> - The basic journey must verify the original card and native
continuation for each supported path.
> - This pull request shares the scenario and browser actions, adds
strict API receipt evidence, and verifies the saved document after
reload.
> - Separate tool-documentation qualification and the original failed
measurements keep their existing scope and outcomes.
## Linked Issues or Issue Description
**What existing behavior does this improve?**
It adds portable coverage for answering agent questions and resuming the
same task on an attached deployment.
**Current behavior**
The classic suite has the two-question scenario. The portable catalog
does not expose it. Its shared native-path grader recognizes provider
and dedicated semantic receipts, but rejects a valid completed native
API creation receipt.
**Proposed behavior**
Expose `question-answer-resume` as a deployment-targeted journey. Answer
one choice question and one text question in the browser. Require
server-owned creation evidence, unchanged forms, exact board answers,
bound continuation, an agent-authored task document, and readback after
reload.
**Reason and benefit**
The same story can qualify local and hosted targets through their normal
provider setup and cleanup. It catches lost answers, duplicate or
foreign work, premature output, and stale documents.
## What Changed
- Register the journey with a three-run maximum and positive model
allowance. Accept one to three attributed work runs at both collection
and retained-result boundaries.
- Share the existing two-question scenario, browser answer actions, and
behavioral grader. Keep separate tool-documentation grading unchanged.
- Recognize a completed native API creation receipt only when its HTTP
201 result retains the original pending card, exact
task/company/agent/source run, idempotency key, and full normalized
form. Require the receipt to remain unchanged after the answer and
through the final checkpoint. Missing, pending, foreign, or altered
proof still fails.
- Check signed-in board ownership, each native continuation, no
premature or extra output, document authorship, and the saved revision
after reload.
- Wait for the authenticated product region before screenshots. Recheck
origin and path after readiness. Bound readiness and capture by the
campaign deadline.
- Add closed stage-specific errors, catalog coverage, sanitized receipt
calibration, and documentation. Include the provider-form and
browser-question presentation helpers in the definition fingerprint.
## Verification
- Fresh QA staging qualification passed at
`7f7de8be286a436010a48af89aba1727b4451493`: all 11 required checks
passed, cleanup passed, and all three native runs used the configured QA
target and model. The [sanitized run
report](https://pages.paperclip.ing/evals-qa-smoke-41121966520-1/)
records the measured pass. This is one successful qualification, not a
reliability claim.
- Final integration head `ae0d9d289831c770a413658d3520fb1dc8ad60e3`
contains the one main merge through
`fd008979f5794fda3d5f54151d36acd31d0de140` and an exact cherry-pick of
the four-file qualified search fix
`4edd10804e2b42941aa2038e90c362c13a0b98dc` from #18370. Patch identity
matches. It preserves the project and search registrations and atomic
search readback. The question live qualification ran at the earlier SHA
above; the final integration head has separate local and CI
verification.
- Eighteen question journey, helper, and test file blobs are identical
between the qualified and integration SHAs. This includes clarification
flow and evidence, question grader and browser helper, the sanitized
receipt fixture, screenshot helper, provider-form helper, browser
question presentation, continuation scenario/readiness/flow, shared
agent validation, and user actions. The question run-count contract and
dispatch branch are also unchanged. Reviewers can compare those paths
directly between the two commits.
- All 622 selected Product E2E, catalog, question-resumption,
necessary-question, report, and readiness tests pass after the main
merge. After the exact search delta, all 145 narrow
question/search/project/contract tests and all nine real Chromium search
regressions pass. Map checks, combined credential-free preview, source
fingerprints, and the earlier workspace-link preflight pass. The local
TypeScript attempt reports stale Runner declaration errors for
`onAccountingEvent` in unchanged native-session code; fresh official CI
provides the clean-build type gate. The three search source/test blobs
match the qualified search merge, and the three project source blobs are
unchanged.
- An in-memory probe executes the actual launcher fingerprint
expression. Each project, search, and question implementation input
changes the digest. Separate red probes first reproduced the omitted
provider-form and browser presentation helpers; both pass after their
source inputs are added.
- The sanitized retained native API response-shape regression fails
before the grading correction and passes afterward. Negative controls
reject missing, pending, foreign, changed, spoofed, or wrongly bound
receipts, answers, continuation, model, environment, and document
evidence. All nine provider/semantic/API path combinations pass
calibration.
- The real local Chromium screenshot probe fails before the readiness
repair and passes afterward. Login redirects and sign-in pages at task
URLs remain excluded with no screenshot files.
- The two original attached failures remain retained: the first stopped
at early screenshot capture; the second stopped before answering at the
former API provenance gate. An earlier sandbox run also retained one
report subprocess failure; the full selected tests then passed with
local subprocess IPC access.
- The prior integration CI had an unchanged canary mock `EADDRINUSE`
failure. That history is retained; its documented retry is owned by the
coordinator. No application API, schema, production UI, model, runtime
policy, credential, or deployment configuration change is included.
Final integration head
|
||
|
DottaandPaperclip
|
a1845b3b64 |
fix(evals): capture TestOps registration after watch (#18375)
## Thinking Path > - Paperclip manages AI agents and verifies their work through Product E2E tests. > - The fixed Chat feature workflow prepares a TestOps case and plan before any model run. > - Its first zero-result bootstrap completed, but the watched child did not receive the generated job-run ID. > - The official watch log contained a job-run ID, so the missing child variable did not prove that registration was absent. > - This pull request reads official context after watch and validates the linked backend records. > - The workflow retains closed registration metadata and stops on unknown or foreign context. ## Linked Issues or Issue Description Refs: #18322. **What happened?** The successful zero-result bootstrap retained `official_job_run_id_missing` from its watched child. It could not prove the backend job and workflow binding needed for the next setup step. **Expected behavior** After successful watch, read the generated context, validate its current GitHub run and fixed workflow, and retain only closed IDs and parameter metadata. **Steps to reproduce** Run the fixed workflow's `bootstrap` operation on master. Inspect its setup artifact after watch completes. No evaluation or synthetic passing result is needed. **Paperclip version or commit** The original bootstrap used merged #18322. This follow-up starts from `76fde891ccc29283b3f5c8dcdce089c5b3d937d1`. **Deployment mode** GitHub Actions metadata-only setup in the protected reporting environment. ## What Changed - Run an internal context reader after successful bootstrap watch in the same protected step. - Parse only numeric official job-run and launch IDs in memory. Do not log or save raw command output. - Validate current-run, project, launch, job, workflow, and bounded parameter metadata before writing an exclusive registration receipt. - Retain the two known registration receipt paths with the existing setup artifact. - Preserve the original inside-watch receipt, case reconciliation, provider isolation, and evaluation limits. ## Verification - `node --test tests/runner-e2e/chat-feature-testops-api.test.mjs tests/runner-e2e/testops-chat-reporter.test.mjs`: 48 pure/mock/FS controls pass (36 prior and 12 new). The actual local Node version was 26.4.0. Normal Node 24 CI remains required. - Four MJS syntax checks and `git diff --check` pass. No TypeScript noEmit claim is made for this MJS-only helper change. - Actionlint passes for the changed workflow. - Normal [GitHub Actions CI](https://github.com/paperclipai/paperclip/actions/runs/41123894542) passes on exact head `58396de10f0c49294b0bbc0f4709c29990d57d53`: 48 jobs succeeded and one skipped. Both required `ci / e2e` and `ci / verify` pass. - Greptile reviewed that exact head at 5/5 with no actionable findings or unresolved review threads. - No local CLI, provider, model, app, database, or build qualification was run. No replacement bootstrap or paid evaluation was dispatched. ## Risks The official context or backend DTO shape can differ from the supported format. The reader then stops with a closed failure receipt. Pure tests do not qualify the remote CLI or backend. The original zero-result bootstrap and its missing child-context observation remain part of the history. The case's actual outcome remains unmeasured. ## Model Used OpenAI GPT-6.1 Sol (`gpt-6.1-sol`) with reasoning, code editing, tool use, and pure test review. The runtime does not expose the 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 - [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
|
4edd10804e |
fix(evals): batch task search result readback across renders (#18370)
## Thinking Path
> - Paperclip lets people manage AI agents and saved work.
> - Product journeys check user actions against saved state.
> - The task search journey compares the exact visible task identities.
> - A search can replace rows while the harness reads an old indexed
locator.
> - The attribute read can then time out after the correct results
appear.
> - This pull request reads the visible links together and preserves
strict result checks.
## Linked Issues or Issue Description
Refs #18355.
**What happened?**
An attached Cloud attempt passed the task search journey. A later
attempt on the same harness and app revision failed at search-result
readback. Its owned screenshot showed the correct target and completed
control. Authentication, collection, and cleanup passed. The failure
remains recorded and the journey has not been admitted to the daily
plan.
The reader called `all()`, then awaited `isVisible()` and
`getAttribute()` for each indexed locator. When a search replaced the
old collection between those reads, the next attribute read waited for
an index that no longer existed.
**Expected behavior**
Read the matching visible links together. Compare the complete identity
set against the owned target and controls. A render transition must not
consume the whole poll deadline on a stale row index.
**Steps to reproduce**
1. Render three task links in the task collection.
2. Let the reader observe the third link as visible.
3. Replace the collection with the two correct search matches before its
attribute read.
4. The old reader times out waiting for `nth(2)`, even though the two
correct rows are present.
## What Changed
- Read visible link attributes in one batch. Discard a detached snapshot
and retry within the existing poll deadline.
- Keep the exact identity comparison, duplicate detection, control
exclusion, saved-content checks, and zero-run checks.
- Add a real Chromium render-transition regression and invalid-result
cases. Extend the stateful flow test for a detached snapshot.
- Document the readback boundary. No app UI, target configuration, or
cleanup behavior changes.
## Verification
- Review follow-up: the existing journey wait helper is now exercised
with the real Playwright poll. Its first read evaluates the actual
reader callback on detached browser DOM nodes and returns no evidence.
The second read contains an extra owned control. Only the third, exact
pair succeeds. The test observes all three reads. The 30-second deadline
and grading behavior are unchanged.
- Rebased onto current master, including the merged project journey.
Credential-free discovery confirms exactly one `project-task-context`
and one `task-search-filter` entry.
- Latest rebased checks passed: all nine Chromium search regressions,
all 381 Product journey unit tests, runner typecheck, discovery, maps,
and diff checks. Exact head
|
||
|
DottaandPaperclip
|
6dcf4160dd |
perf(chat): load recent messages and page history on scroll (#18358)
## Thinking Path > - Paperclip lets people manage work done by agents. > - A standing agent chat can hold months of conversation. > - The chat route forced full-history mode, bypassing compact task loading. > - A cold visit fetched lifetime run summaries and hydrated historical transcripts. > - This PR starts with 20 recent messages and loads earlier pages on upward scroll. > - It preserves the reader's position, current questions, and draft while history grows. ## Linked Issues or Issue Description **What happened?** Standing agent chats forced full-history loading. A long conversation fetched full activity and run history before it became readable. Earlier comments also used an explicit load control. **Expected behavior** Open a small recent conversation quickly. Load older messages when the reader scrolls upward. Show an inline loading state. Keep the visible message and composer mounted through paging, live updates, and Back navigation. **Steps to reproduce** 1. Open a standing agent chat with hundreds of messages and completed runs. 2. Use a fresh browser query cache. 3. Observe full activity and run reads plus transcript requests for historical runs. 4. Compare the recent window and upward paging after this change. Related work: #18312 and #18323 restored compact task and collection loading. #14848 added reopening of visited agent chats. A GitHub search found no open PR for this chat-history optimization. ## What Changed - Return 20 contiguous recent comments for standing chats, with a stable oldest cursor. Keep the existing compact task exchange contract. Use the existing company/issue/time index for cursor reads. - Keep chats compact even if their task view previously expanded history. Retain observed messages through live catch-up and cache-backed return visits. - Keep fetched history pages, acknowledged sends, and the observed recent cursor separate. Reconcile live replies without losing history after a send or deleted cursor. - Hydrate metadata only for newly revealed message/run IDs; preserve prior batches while paging. Slow-send recovery refreshes the recent window. Receipts failures have an independent retry. - Hydrate run summaries only for loaded messages and the latest turn. Validate batches of at most 100 run IDs and keep the existing issue/company access rules. - Reuse hydrated terminal native streams when only history membership changes. Preserve retry, resume, mode changes, reconnect recovery, and interrupted recovery reads. - Load older messages, decision receipts, and task handoffs through upward scrolling. Support wheel, touch, and keyboard input. Keep pending questions visible. - Reserve a status slot while loading. Preserve a logical message anchor and the mounted composer. Pause failed page reads until an inline Retry. - Add database, route, controller, and scroll regressions, plus an opt-in production-build browser rig for long chats on desktop and mobile. Update design and feature/access documentation. ## Verification - The real PostgreSQL cold-open suite passes all 12 tests, including contiguous pages with equal timestamps, complete history, company denial, and scoped run IDs. - Activity route and OpenAPI suites pass 53 tests. The route run allowed local cold-import overhead; CI retains the repository's normal timeouts. - Three new paging regressions fail against the unchanged source. The candidate's focused paging/window suite passes 26 tests. Broader UI verification is recorded below when complete. - Desktop and mobile browser acceptance covers recent-only startup, per-frame reveal geometry, delayed and failed older pages, stable reading anchors and drafts, Back, older comment links, and one persisted live reply. - Full workspace typecheck and build passed locally. The final UI source also passes its typecheck and production build. - Affected UI suites pass: 212 controller/history/paging tests, 352 thread/scroller tests, 26 native-stream tests, plus metadata batching and live-cache hydration coverage. The issue mutation/access route suite passes 179 tests. - All 22 final desktop/mobile browser journeys pass, including delayed image growth. The isolated second-page request proof also passes six repeated desktop/mobile runs. A separate 6,000-message, 3,000-run fixture passes cold-open and delayed-paging checks on both screen sizes, still opening 20 messages and loading one 50-message page. - Final matched production timings use baseline `aa58fa1073` (`index-BaRiMbxd.js`) and candidate UI source `f9a26b9783` (`index-DPWTbsFF.js`, identical application source at final head `bf71ffd5af`). Chromium 151.0.7922.34, 1440×900, 4× CPU throttle, no added network latency, browser HTTP cache disabled, fresh application cache for each cold visit. Both variants block service workers and use the same warm local API server and immutable 600-message/300-run fixture. Only document HTML is intercepted for both variants; API requests are real. The common API revision is `109981f8c1`; subsequent commits change only UI and tests. | Usable conversation | Baseline median (range) | Candidate median (range) | | --- | --- | --- | | Cold direct URL | 6.57 s (5.69–7.45) | 3.07 s (2.50–3.28) | | Warm in-app return | 3.76 s (3.54–4.76) | 0.67 s (0.58–1.46) | | Back return | 2.53 s (2.12–4.20) | 0.51 s (0.47–1.18) | Five pairs alternate order. Every paired difference favors the candidate for all three journeys. Readiness requires a revealed, fully visible latest reply above the composer and an editable composer. Return timing starts after the Tasks route renders. Cold API requests fall from 953–954 to 71; API wire bytes fall from 517–559 KB to about 49 KB. Separate CPU profiles and request waterfalls confirm the removal of lifetime event hydration and are excluded from the five timing pairs. Both variants have the same expected missing optional plan document (404) and no browser exceptions. Control recent-window API reads have medians of 37 ms / 32 ms, with shared-host outliers up to 376 ms. The 6,000-message recent-window SQL uses `issue_comments_company_issue_created_at_idx` (22 index rows read for a 21-row probe). Geometry assertions retain a 2-pixel tolerance for the reading anchor and first-reveal layout. These are local synthetic measurements, not deployed staging latency. Earlier timing probes are excluded because they did not isolate service workers or composer clearance. Final head `bf71ffd5afb5684ba9105a3127f617e7adf5139c` has 55 successful checks, three optional skips and one neutral result, with no failures or pending checks. Greptile is 5/5 with no open review findings. ### Failure record Early local attempts exposed an unconditional scroller wrapper that changed the old task layout, and automatic retries on chat page failures. Both were corrected. The original task scroller structure stays in place. Metadata request assertions wait for initial connection recovery to settle, then verify only the new page’s reads; retry/resume/reconnect revalidation remains intentional. The initial review caught a send-before-scroll cache defect; its controller regression failed before the fix. Later rapid-paging checks caught scroll readiness being gated by metadata after the first reveal; readiness now stays enabled once revealed. Live replies use the separate acknowledged-message cache, then catch up from the server cursor. Visual review exposed a pagination wrapper bypassing existing composer clearance. The body is again a direct scroller child; cold-open browser checks require the full latest reply above the composer. A resize-observer regression failed before the fix and now passes. Controlled delayed inline images prove both pinned following and a stable reading anchor on desktop/mobile. The first performance fixture lacked valid native event envelopes and caused missing-log fallbacks. It is excluded from latency claims. The final fixture uses saved native assistant events and conversation session generations. The live test uses a temporary fixture-agent key, attributes the reply to an existing synthetic run, and revokes the key. It starts no model or agent process. Local database attempts failed during a Docker shutdown. The retained disposable database was restored and its suite passed. Early fixture errors included a missing conversation identity field and session generation. Mobile test setup also selected an offscreen sidebar link and reused queued human input. Optional announcements could intercept input; the fixture now dismisses them before the journeys. The scroll observer now waits for the delivered wheel event and settled animation frames, including when the viewport is already at the top. A concurrently running live test briefly changed another probe’s row count; the final acceptance and timing runs were sequential on separate fixtures. The final rig uses the visible mobile navigation and a saved agent reply. Anchor observation waits for the user's scroll to settle before measuring the effect of live updates. The full local `pnpm test:run` attempt hit repeated embedded PostgreSQL initialization failures and failures in untouched native-session executor suites; it was stopped after the repeated environment failures. The targeted real-PostgreSQL suite passed, and final-head CI is required before handoff. The first CI run also caught an exact mock expectation missing the added company scope; the expectation is updated and its 179-test suite passes. The final-head first CI attempt also stalled an assigned child task in the existing plan-approval scenario, then hit a heartbeat-run row lock timeout during fixture cleanup. The failed shard and aggregate check passed on an unchanged rerun of the same commit. Scenario assertions and timeouts were unchanged. The failure and rerun are retained in [CI run 41116120430](https://github.com/paperclipai/paperclip/actions/runs/41116120430). ## Risks - Historical transcript diagnostics no longer load on a cold chat open. Saved messages and conversation events remain available; diagnostic run views keep their full records. - Cursor paging must not omit comments with equal timestamps. Database tests cover timestamp/UUID ordering and the complete set. - Prepended or late content can move the reader. Browser checks cover retained row identity, viewport offsets, mounted drafts, and Back. Previously loaded pages stay in the query cache for its existing lifetime. - Local synthetic evidence does not establish deployed staging latency. This PR does not merge or deploy the change. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, tool use, and code execution. The exact backend deployment ID and context window were 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 (affected checks; limits are recorded 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> |
||
|
DottaandPaperclip
|
9e1735ec17 |
test(evals): add a fixed TestOps AWS chat feature route (#18322)
## Thinking Path > - Paperclip lets people manage AI agents and inspect their saved work. > - Chat feature evaluations test a real request against saved tasks, audit records, links and reload behavior. > - The current suite has Codex and OpenCode cells but no dedicated runnable DeepSeek 1017 backlog plan in TestOps. > - TestOps must select the real case before the workflow can access a model key. > - This PR adds one fixed, default-branch AWS route for native OpenCode and DeepSeek 1017. > - The route keeps original evidence private and sends only a verified result and approved screenshots to TestOps. ## Linked Issues or Issue Description **Subsystem affected** Product E2E evaluation definitions and GitHub Actions delivery. **Problem or motivation** A dedicated TestOps plan cannot run the existing backlog chat journey with native OpenCode. A result must prove saved work and cleanup rather than report that a model call completed. **Proposed solution** Add one OpenCode backlog cell and one fixed public workflow. TestOps selection validation runs in a token-isolated job before builds or inference. The workflow retains private originals and reports the bound outcome to the selected TestOps case. **Alternatives considered** A private-repository wrapper would need a new cross-repository GitHub write credential. The new route uses the existing TestOps GitHub integration instead. Existing manual candidate approval paths remain separate. **Roadmap alignment** This extends the existing feature evaluation infrastructure. It changes no production feature, global runtime default or model default. Current main #18319 supplies the v2 grader, delegation queue accounting and final screenshot capture; these improvements are preserved. Related product work: #16814 and #17126 concern backlog handling. This PR changes neither product implementation and does not claim those changes are required. ## What Changed - Add exactly one native OpenCode backlog cell with DeepSeek 1017 under the separate `runner-opencode-deepseek-1017` profile. Preserve all four existing Codex and production OpenCode cells and their models. - Add a default-branch-only workflow with fixed evaluation, zero-result registration and one plan-launch modes. Require GitHub actor checks and authenticated singleton selection before inference. - Add a separate shipped-main evidence authority. Reuse the existing strict private storage, scanners, inventories and scoped AWS sessions. - Bind sanitized results, generated JUnit and producer-approved screenshots to the exact source, run, attempt and case. Keep failed grades and cleanup failures failed. - Add TestOps API, CLI and result adapters with pure controls and normal unit-test wiring. Keep the reporting token out of the evaluated process. - Verify the current server case result and attachment metadata after upload, without masking an earlier failed upload or evaluation. - Apply only the existing reviewed lock correction, verify unchanged tracked dependency inputs and bind the lock receipt across the compiled-output handoff before frozen installation. - Bind the initial administrator launch to one approved workflow run number and live GitHub source/actor metadata before TestOps authentication. Use bounded REST case reconciliation and only the reviewed existing GitHub integration. - Make two existing CI tests observe their intended state deterministically: a fake clock for copy-feedback lifetime and a completed-pass barrier for retry-cancellation batches. Production behavior and the original assertions remain unchanged. The two test-only fixes are intentionally included because their observation boundaries blocked this delivery’s required normal CI. They preserve every original assertion and change no production behavior. The original failed runs remain recorded, and the exact causes of those observations remain unknown. ## Verification - 157 focused profile, catalog, oracle, filesystem admission, export, private-path and workflow controls passed on the current-main composition. The existing stock OpenCode models and the new exact DeepSeek 1017 selection are checked separately. - 44 API, CLI, filesystem and post-upload readback controls passed, including foreign selection, changed hashes, unsafe files, failed outcomes, expired readback and no-resend behavior. - The new exact private session and trace layout controls produced four genuine parent failures and pass with the selected-path extension. The frozen 244-case candidate approval registry remains unchanged. - The lock guard has seven filesystem controls plus seven existing canonical-owner controls. The unrestricted old workflow failed its actual policy calibration; the corrected frozen-install route passed. - Eval-owning TypeScript checks, workflow lint and Git whitespace checks passed. The separate UI test correction was checked with the focused jsdom unit suite. The server test barrier has pure wrapper and real scheduler controls; its DB-backed route was not run locally. No additional local typecheck/build was run for these test-only changes. - Full native builds, provider execution, app/database behavior and TestOps registration are reserved for GitHub Actions/AWS. They have not run locally. - The initial normal CI run failed the new-workflow policy allowlist and a heartbeat scheduling assertion. The policy fix has an actual parent red/current green control and rejects five unsafe route variants. Heartbeat source is unchanged and its failure cause remains unknown. The original run is retained. - The launch-fence actual caller fails against the published parent and passes the five own-run contradiction controls before authentication. - The prior `a5aa359d` normal CI completed successfully. The complete `f9e1f289` run passed all leaf jobs except an existing copy-feedback UI assertion, plus its dependent verification gate. The four owning UI files were unchanged from the earlier passing head and actual tested merge; the exact scheduling cause remains unknown. - The correction changes only that test to actual React `act` and a scoped fake clock. All eight component tests pass; early-hide and never-clear source variants each fail the intended lifetime assertion. Production clipboard/card behavior is unchanged. - The `2f27f3fc` run passed the corrected UI shard and every leaf job except the existing retry-cancellation intermediate-count assertion, plus its dependent verification gate. The seven relevant owners match the earlier passing run and historical source comparison; the exact CI scheduling cause remains unknown. - The test-only barrier observes a persisted completed worker pass before asserting the unchanged 20-item cap and two-company fairness, then releases it for the original 29-item drain and 180-second idle check. Two extracted-wrapper controls and three actual scheduler controls pass. The real DB-backed test awaits AWS qualification. - On `836e80ac`, both repaired test shards passed in AWS. The original attempt retained a separate E2E cleanup HTTP 500 from a PostgreSQL lock timeout and one runner shutdown. No production repair or precise contention/shutdown cause is claimed. - Exactly one targeted retry of E2E shard 3 and its aggregate gate passed in attempt 2; exactly one recovery of the interrupted server shard and its verification gate passed in attempt 3. Other passed jobs were carried forward. The final normal run is successful. No whole-workflow retry or evaluation/model retry was launched. - Both literal required checks (`ci / verify` and `ci / e2e`) are successful on the exact final source `836e80ac`. Greptile is 5/5 and no review threads remain open. Actual TestOps registration and the OpenCode/DeepSeek feature result remain unqualified until their separate executions. ## Risks - Actual zero-result registration, tenant API shapes and public-repository dispatch access must be checked before the first evaluation. Missing or ambiguous metadata stops the route. - The real OpenCode/DeepSeek backlog result and Allure outcome remain unqualified until that run. No passing result is seeded. - One human query, one accounted work run, 120-second turn, 420-second attempt, 300-second settled-state bound and zero retries remain unchanged. Setup usage is outside work accounting and stays unknown. - The reporting token was configured through the separately reviewed protected credential bootstrap. It is unavailable to AWS evaluation jobs. - Metadata reconciliation is capped at ten pages, 1,000 cases and four parallel detail reads. Missing or inconsistent metadata stops before inference. The initial administrator launch counter must be created once after registration; it does not restrict normal TestOps Run actions. - Existing manual workflows, historical evaluation grades and candidate approval records remain unchanged. The original CI failures are preserved; the test corrections do not establish retrospective causes. ## Model Used OpenAI GPT-6.1 Sol (`gpt-6.1-sol`) with reasoning and tool use. The runtime did not expose its 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> |
||
|
Dotta
|
a4308b9791 |
test(e2e): add browser project and task context journey (#18353)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - People group tasks into projects and use those projects to keep work in context. > - Deployment journeys already test task creation and assignment. > - They do not yet prove that a new project and its task keep the same saved relationship after reload. > - This pull request adds an explicit browser journey for that gap. > - Independent API reads and visible task properties must agree on the exact records and selected settings. > - The journey uses no model runs. Each deployment must pass before a daily plan can include it. ## Linked Issues or Issue Description **What existing behavior does this improve?** The deployment-targeted Product E2E suite can now check project creation and task context through the production dialogs. **Current behavior** The suite covers standalone task creation and assignment. It does not prove that a newly created project contains the task selected in the composer, or that the relation survives reload. **Proposed behavior** Select `project-task-context`. Create a unique project and a self-assigned task with that project selected. Reopen the project, open its task, and reload. Require exact saved company, project, title, prompt, user assignment and standard work mode. Require no agent runs before or after reload. **Reason and benefit** The same journey can qualify different deployments. It catches lost project links and incorrect saved context without a model call or a separate QA-only test implementation. ## What Changed - Add the browser flow and an independent saved-state oracle with positive and negative cases. - Scope project and task navigation to unique links in the main content so sidebar links cannot bypass list coverage. - Bind reopened routes to saved project and task references. Verify the main project heading and visible task content instead of a hidden duplicate sidebar label. Keep browser failures as closed stage codes. - Register the journey, its closed success checks, and its source files in the definition hash. Default model-matrix selections stay unchanged. - Associate qualification-report coverage with project lifecycle, project task context and task creation. Reject retained browser-only passes containing run IDs, provider/model metadata or native/legacy execution. - Add `project` to the existing durable ownership resource contract. A target provider remains responsible for cleanup of its own project fixtures. - Update discovery expectations, the deployment journey guide and the projects feature map. ## Verification - Focused runner tests: 110 passed across the new oracle, discovery, journey contracts and existing report subprocess checks. The first sandboxed attempt blocked the existing report subprocess IPC socket; the same selection passed with local IPC enabled. - Late review regression checks: both report association and unexpected browser-only model work reproduced as failing tests before the fix. The final contract, saved-state oracle and report catalog selection passed all 72 tests. Runner typecheck passed again after these changes. - Browser readback: the old project-name selector failed in Chromium when a hidden sidebar label preceded the production project heading. All six browser regression cases now pass, including missing rows, wrong project/task destinations and a wrong project property. The fixture uses the production heading and task-row overlay conventions. Credential-free Node 24 imports pass. - Current-master integration: 208 focused catalog, project/search, campaign and reporting tests passed. All nine project/search browser regressions passed. Runner typecheck, map checks and Node 24 imports for both journeys passed. The combined catalog, dispatch, source hashes, success checks, browser-only guard and feature associations retain both selections. - The first deployment qualification created both owned records but failed during later readback with a generic browser code. Authentication, setup, collection and cleanup passed. That failed grade is retained. The corrected source subsequently passed a fresh attached QA campaign; the original failure remains preserved. - New oracle alone: 7 tests passed after the final catalog mapping change. - `node scripts/check-maps.mjs` — passed; structural checks only. - Credential-free `--suite product-journeys --list` — passed and lists this browser journey. - `node node_modules/typescript/bin/tsc -p tests/runner-e2e/tsconfig.json` — passed after building required plugin, runner and eval-kernel declarations. The initial fresh-checkout check failed because those declarations were not built. - All current-head CI gates passed, including repository typecheck, tests, build and browser checks. Greptile reviewed the exact head at 5/5 with no unresolved findings. Fresh attached QA passed at `f19557c9eebde14c2def37195d6b154c23eb9690`: [sanitized report](https://pages.paperclip.ing/evals-qa-smoke-41119722363-1/). All five independent checks and authentication, setup, journey, collection and cleanup passed; zero agent runs started. ## Risks - This adds an opt-in test and changes no production behavior or schema. UI control changes can fail the journey. - The target provider must support durable `project` ownership before it can run this case. Project cleanup must settle owned tasks, reject foreign linked tasks and archive the project while preserving history. - Source registration alone does not include the candidate in any saved daily plan. A fresh pass against that deployment is required first. ## Model Used - OpenAI Codex, GPT-6 family, with code inspection and tool execution. The host does not expose the exact underlying model revision 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 (no duplicate journey found) - [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 |
||
|
DottaandPaperclip
|
a0e3ac62a7 |
test(evals): add reusable task search and filter journey (#18355)
## Thinking Path > - Paperclip lets people manage AI agents and their work. > - The task list lets people find saved work and open its context. > - Product journeys can exercise the same task UI on local and Cloud targets. > - The shared suite lacks a search and status-filter journey with control tasks. > - This pull request adds that journey with exact task identity checks. > - The benefit is reusable evidence that search, filters, and task navigation work together. ## Linked Issues or Issue Description **What existing behavior does this improve?** Product E2E coverage for task search, status filters, and saved task navigation. **Current behavior** The shared journey suite covers task creation and assignment. It does not prove that a search excludes an unrelated task, that a status filter excludes a matching completed task, or that the selected task retains its saved content. **Proposed behavior** An explicit `task-search-filter` selection creates three owned tasks. It searches through the production task list, applies Backlog, and opens the exact target. It verifies reload and Back navigation, saved task fields, and zero agent runs for all three tasks. **Reason and benefit** The same journey can run against a managed local stack or an attached Cloud target. Independent assertions reject missing evidence and plausible wrong results. Deployment qualification and daily-plan admission remain separate gates. **Breaking changes** None. Existing journey selections keep their behavior. Related work: #16979 fixes status filtering before pagination. #15403 covers saved task views. This journey tests the existing search/filter interaction and does not add a new UI or saved-view feature. ## What Changed - Add a zero-run browser journey with one target and two owned controls. - Compare exact visible task identities after search, filtering, reload, and Back navigation. - Verify saved company, owner, title, description, status, and zero agent runs through the public API. - Register dispatch, source hashing, closed success checks, and browser reporting validation. - Add adversarial oracle tests and a campaign dispatch test. Document the selection and its feature-map scope. - Exercise the real browser flow with a stateful UI/API port across Coordinator, combined Tasks, and legacy shells. Inject missing ownership, wrong destinations, lost filters, changed saved context, and unexpected native work. - Use the combined Tasks shell's named Backlog view for reload persistence, as required by its existing view-preset contract. Preserve the successful polled result snapshot and link the case to feature coverage. ## Verification - First attached Cloud attempt failed during Backlog selection. Its cleanup passed and it started no model work. The original failed result is preserved; the corrected candidate passed a fresh attached campaign below. - Local Chromium reproduces the cause: the legacy StatusIcon and text label produce the accessible checkbox name `BacklogBacklog`, so the prior exact `Backlog` lookup finds no status control. - The selector now stays inside the Status group and accepts the explicit and duplicated names. Three real Chromium regressions pass, including a separate assignee called Backlog. - Latest focused journey checks: 130 passed. Runner typecheck and map checks passed. Unexpected browser/API failures retain only closed source-owned stage codes. Screenshot privacy guards and all exact identity, content, filter, and zero-run checks are unchanged. - Fresh attached Cloud qualification passed at `e3dfc0ae37e8ac74e9a07ca24421ea6c96f9bcfc`: [sanitized report](https://pages.paperclip.ing/evals-qa-smoke-41114233380-1/). All five independent search/filter/reload/reopen/no-agent checks and all five phases, including cleanup, passed. The browser journey started zero agent runs. - Runner typecheck passed: `node node_modules/typescript/bin/tsc -p tests/runner-e2e/tsconfig.json`. - Focused oracle, contract, and campaign tests passed: 101 unique tests. The campaign suite was rerun after its new dispatch test: 30 passed. - Direct browser-flow tests passed: 23 tests. These call the real `taskSearchFilter` function through a stateful UI/API port. They do not establish live browser or Cloud qualification. - Latest focused search, oracle, catalog, and report checks passed: 127 tests across five files. - Full runner Vitest completed with 5,204 passes, two skips, and three failures outside this change. Two local embedded PostgreSQL fixtures failed at bootstrap. The Pi controls test expects profile version 35 while the unchanged base declares 37. These failures remain recorded; the PR does not change those fixtures or pins. - The remaining Node support suite passed separately: 137 tests. The standard unit launcher stops before these tests after a Vitest failure. - Credential-free discovery passed: `node --import ./cli/node_modules/tsx/dist/loader.mjs tests/runner-e2e/launch.ts --suite product-journeys --list`. - `node scripts/check-maps.mjs` passed. `git diff --check` passed. - All 58 current-head check runs are green, the additional security status passed, Greptile reviewed this exact head at 5/5, and no review threads remain unresolved. Typecheck, test, build, browser CI, and fresh attached Cloud acceptance passed. Daily-plan admission follows merge. ## Risks - Task-list control labels and link structure can change. Such drift fails the journey and needs an update. - The journey creates three board-assigned tasks. It records ownership before creation and records each returned ID immediately. The existing target provider owns cleanup. - Local assertion tests do not prove that a deployed Cloud target passes. The target owner must record a fresh live pass before admission to a daily plan. ## Model Used OpenAI Codex, GPT-6 family. The host does not expose an exact model version or context limit. Used repository search, TypeScript tools, and tests. ## 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> |
||
|
github-actions[bot]andlockfile-bot
|
ade402c75f |
chore(lockfile): refresh pnpm-lock.yaml (#18360)
## Thinking Path > - Paperclip uses a dependency lock for reproducible builds. > - Dependency changes on master need matching lock entries. > - Paid eval reports also pin a reviewed offline correction. > - Reports must install before paid model work starts. > - This automated PR keeps the lock and report proof together. ## Linked Issues or Issue Description **What happened?** Dependency manifests on master changed after the last committed lock refresh. **Expected behavior** Frozen installs must match those manifests. Eval reports must retain matching offline lock proof. **Steps to reproduce** Run a frozen install before and after this lock refresh. ## What Changed - Refresh the dependency lock from the manifests on master. - Include the matching offline report correction and hashes. ## Verification - Pinned pnpm resolution completed with lifecycle scripts disabled. - The report proof refresh accepted the reviewed committed baseline. - The workflow rejected changes outside the lock and report proof files. ## Risks The lockfile bot merges this PR without CI or review when every added package version is at least 3 days old on npm. Otherwise the PR waits for a maintainer, and the daily refresh tries again. The bot merges only the exact commit this workflow generated from manifests already reviewed on master. Unknown report lock baselines remain blocked. This update changes no permissions or paid enable gates. ## Model Used None — generated by pnpm and the deterministic report-lock updater. ## Checklist - [x] Thinking path and problem description included - [x] Generated lock and matching report proof included - [x] Release-age gate decides between bot merge and maintainer review Co-authored-by: lockfile-bot <lockfile-bot@users.noreply.github.com> |
||
|
DottaandPaperclip
|
aa58fa1073 |
perf(ui): restore fast Coordinator task and inbox loading (#18323)
## Thinking Path > - Paperclip lets people manage work done by agents. > - Tasks, Inbox, and Needs You must stay usable as work history grows. > - The Coordinator views bypassed some existing collection optimizations. > - Navigation fetched complete decision cards for a count, and Tasks mounted every loaded row. > - This PR uses the canonical attention rules for small count and task-ID responses. It also restores bounded rendering and the optimized Inbox selection. > - The benefit is less data and less initial rendering work without a second definition of pending work. ## Linked Issues or Issue Description **What happened?** Coordinator navigation fetched the full attention feed on every cold collection visit. Tasks rendered the entire loaded collection. Coordinator could also select the old Inbox loader when Streamlined UI was off. **Expected behavior** Fetch only the projection needed by each surface. Start with a bounded set of rows. Keep complete task membership, accurate counts, current permissions, and stable loaded rows. **Steps to reproduce** 1. Enable Coordinator and open Tasks or Inbox with an empty browser query cache. 2. Observe a complete attention-card download for the navigation badge. 3. Open a large task list and observe that every loaded row mounts at once. 4. Disable Streamlined UI while keeping Coordinator enabled, then open Mine and observe three overlapping compact task reads. Related work: #17756, #3744, #17983, and #18312. ## What Changed - Add board-only attention count and complete task-ID projections. Reuse canonical company, privacy, addressee, deduplication, dismissal, snooze, and archive rules. Skip card bodies, plan/image reads, and presentation roster lookup. - Share the small count read across desktop and mobile navigation. Reuse the page-owned response on Needs You and Needs attention. Keep the newest known count through navigation, including zero. - Use the task-ID index for Needs attention and keep task row hydration in batches of at most 100. - Render 100 task rows initially, across the whole expanded tree. Reveal 150 more at a time through scrolling or a button. Keep existing rows through refresh and append. Defer offscreen leaf layout with shared tokens. - Keep Coordinator on the optimized Inbox when Streamlined UI is off. Use index reads for expanded project recents. - Add real-database, component, and production-build browser regression checks. Update the access and feature maps. ## Verification - All required CI checks pass on `1c68ac8957490a6c8585ba0939355f536c8120bf`: typecheck, build, server/workspace tests, serialized database suites, and eight E2E shards. Greptile is **5/5**, with its row-retention finding fixed and the review thread resolved. Storybook checks were intentionally skipped; Paperclip Review completed neutral. - Workspace typecheck and production build passed locally. The final rebased source also passed UI typecheck and production build. Token gates and map checks passed. - The attention service suite passed all 34 tests on disposable PostgreSQL. Final targeted checks cover compact projections, complete-set counts, quality approvals, snoozes, private tasks, and board/company access. Database response capture confirms compact projections do not read plans or images. All 31 OpenAPI tests passed after registering both endpoint contracts. - Six production-build browser journeys passed: desktop, mobile, Coordinator Inbox with Streamlined UI off, Needs You, Needs attention, and an older server page followed by task open/Back. They verify the initial 100-row budget, scrolling, retained DOM identity/height, one owner per attention read, complete task membership, 220 retained rows on Back, and less than 2 px of return-position difference. - The row-budget and Inbox regressions fail against original source and pass with this change. The final three affected UI suites passed 40 tests, and the broader Inbox/IssuesList suites passed 101 tests serially. ### Controlled production-build comparison Baseline `b31ec0a2dc079e32272a8f618425fe92b68bbaec` versus PR head `1c68ac8957490a6c8585ba0939355f536c8120bf`. Isolated baseline asset `index-CS9uPxMx.js`; candidate asset `index-BxJeTQ30.js`. Both UI builds used the same warm disposable API server and database, with 220 tasks and 110 pending board approvals. Chromium **151.0.7922.34**, 1440 × 900 viewport, **4× CPU throttle**, loopback with no added network latency, alternating baseline/candidate order. Cold cases use a new browser context with network cache disabled; warm cases navigate within the same context. Other host workloads are uncontrolled. Readiness requires the destination collection's first visible task row, then two animation frames. Inbox readiness waits for its canonical Mine route so previous-page rows cannot satisfy the measurement. Five paired samples per case; cold Inbox has a second five-pair confirmation wave because the first result was slower. Times below are **milliseconds**; parentheses show min–max. Paired savings are the median of individual before-minus-after pairs, not the difference between medians. | Journey | Pairs | Baseline median (range) | Candidate median (range) | Median paired saving | Faster pairs | | --- | ---: | ---: | ---: | ---: | ---: | | Cold Tasks | 5 | 2752 (2542–3025) | 2300 (2275–2524) | 328 | 5/5 | | Cold Inbox | 10 | 2536 (2283–2927) | 2582 (2117–2856) | 37 | 5/10 | | Warm Tasks navigation | 5 | 2054 (1826–2479) | 1448 (1419–1548) | 635 | 5/5 | | Warm Inbox navigation | 5 | 2020 (1485–2254) | 1568 (1289–1848) | 350 | 5/5 | | Task → Back to Tasks | 5 | 1747 (1609–1948) | 1327 (1238–1355) | 472 | 5/5 | Cold Tasks, warm Tasks, warm Inbox, and Back improved in every measured pair. **Cold Inbox latency is inconclusive:** the two waves pointed in opposite directions, with the candidate faster in 5/10 pairs. This PR claims less badge work there, not a proven cold Inbox latency improvement. The canonical badge response shrank from **200,039 to 359 decoded bytes** (encoded HTTP transfer: **8,125 to 598 bytes**). The compact response also skips card payloads, plans, images, and presentation-only roster work. These are measured mechanism reductions, separate from end-to-end latency. Needs You still owns its full card feed; Needs attention owns its complete task-ID projection. Reproduction setup and assertions are in `tests/perf/collections/README.md`. <details> <summary>Individual paired readiness measurements (ms)</summary> | Journey | Pair | Baseline | Candidate | | --- | ---: | ---: | ---: | | Cold Tasks | 1 | 3025 | 2283 | | Cold Tasks | 2 | 2542 | 2275 | | Cold Tasks | 3 | 2752 | 2424 | | Cold Tasks | 4 | 2783 | 2300 | | Cold Tasks | 5 | 2641 | 2524 | | Cold Inbox | 1 | 2927 | 2702 | | Cold Inbox | 2 | 2422 | 2629 | | Cold Inbox | 3 | 2409 | 2606 | | Cold Inbox | 4 | 2621 | 2769 | | Cold Inbox | 5 | 2539 | 2856 | | Cold Inbox | 6 | 2283 | 2117 | | Cold Inbox | 7 | 2466 | 2558 | | Cold Inbox | 8 | 2674 | 2286 | | Cold Inbox | 9 | 2696 | 2389 | | Cold Inbox | 10 | 2533 | 2183 | | Warm Tasks navigation | 1 | 2479 | 1423 | | Warm Tasks navigation | 2 | 2228 | 1548 | | Warm Tasks navigation | 3 | 2054 | 1419 | | Warm Tasks navigation | 4 | 1826 | 1448 | | Warm Tasks navigation | 5 | 1954 | 1518 | | Warm Inbox navigation | 1 | 2020 | 1661 | | Warm Inbox navigation | 2 | 2198 | 1848 | | Warm Inbox navigation | 3 | 1733 | 1497 | | Warm Inbox navigation | 4 | 2254 | 1568 | | Warm Inbox navigation | 5 | 1485 | 1289 | | Task → Back to Tasks | 1 | 1721 | 1330 | | Task → Back to Tasks | 2 | 1747 | 1275 | | Task → Back to Tasks | 3 | 1948 | 1355 | | Task → Back to Tasks | 4 | 1609 | 1327 | | Task → Back to Tasks | 5 | 1790 | 1238 | </details> ### Failed attempts and limits - The local full repository command hit embedded PostgreSQL startup failures and 16 failures in the untouched native session executor suite; it was stopped. The full local suite is not claimed green. CI on the final revision provides the broad test result. - A concurrent local UI run hit timing failures under host contention. Inbox/IssuesList subsequently passed all 101 tests serially. No tests were weakened to address these timeouts. - CI caught two integration gaps introduced during this work: missing OpenAPI entries for the new endpoints, and an attention fixture incompatible with the routine filters added on master. Both were fixed; the affected 31 OpenAPI and 40 UI tests pass. - The final revision's first CI attempt failed two untouched tests: scheduled-retry batch cancellation and the tool-access audience/archive race. Both failed jobs passed when rerun without changing the revision. Their original failed attempts remain visible in [the workflow history](https://github.com/paperclipai/paperclip/actions/runs/41083379875). - Browser evidence uses synthetic disposable data. This PR is not deployed staging or production verification. ## Risks - New attention sources must preserve canonical eligibility. Both projections call the existing collector, so they share its permissions and complete-set behavior. - The count and task-ID caches must not stand in for full decision cards. Each projection has its own key under the existing attention invalidation prefix. - A row budget can omit descendants if applied per parent. The tests cover one global budget across a large expanded tree, refresh, append, reset, and incremental loading. - Local browser evidence uses synthetic data and is not staging deployment proof. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, tool use, and code execution. The exact backend model deployment ID and context window were 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 (affected suites; broad-run limitations are listed 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> |
||
|
Nicky LeachandClaude Opus 5.5
|
abc270ee4c |
fix(runner): attest upstream releases and bind ACPX uniformly for every harness (#18325)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Native ACPX runs admit a provider only when its saved profile and `commandDigest` match the current profile exactly. > - For Copilot, Grok and Pi, `commandDigest` hashed dozens of Paperclip source files. Claude, Codex, Cursor and Antigravity hash only their upstream release. > - Any change to a bound source therefore advanced these profiles. #18212 alone moved Copilot from v88 to v91 and Grok from v15 to v19. Each bump reset in-flight sessions and added immutable identity fixtures (about 160 on master). > - #18237 added the provider pack `contractDigest`, which now guards Runner source compatibility at the pack boundary. > - This pull request changes the Copilot, Grok and Pi digests to upstream release attestations, as the other providers already use. > - It also binds the shared ACPX layer the same way for every ACPX harness: the profile's `acpxVersion` and the pack `contractDigest`, which now covers the Runner's ACPX patch. > - The benefit is that Runner source changes no longer reset these sessions or need new profile versions and fixtures. ## Linked Issues or Issue Description Refs #18237, #18212, #17972. **What happened?** A Runner change that touched a bound source file changed the Copilot, Grok or Pi `commandDigest`. The author had to advance the profile version, regenerate an identity fixture and keep every old fixture. On deploy, saved sessions for that provider failed the exact admission check and started fresh. **Expected behavior** A provider's `commandDigest` changes only when its upstream release changes: an upstream version, a platform closure, the ACPX patch or the Grok launcher. Runner source compatibility is the job of the pack `contractDigest`. **Steps to reproduce** 1. On master, change a comment in `src/drivers/acpx/runtime-host.ts`. 2. Run `node packages/paperclip-runner/scripts/grok-profile-identity.mjs --check`. 3. The check fails. The Grok profile must advance, and the next deploy resets Grok sessions. ## What Changed - Add `packages/paperclip-runner/scripts/acpx-release-attestations.mjs`. It computes the release attestation for each provider: - Copilot and Pi: a canonical-JSON hash of the profile version, upstream package versions and the platform closure pins (Pi also binds its Node version). - Grok: the SHA-256 of `launcher.cjs`. Admission already hashes these bytes. - The attestation binds no Paperclip source, shared ACPX layer, absolute path, pack manifest digest or install layout. - Bind the shared ACPX layer once for every ACPX harness (Claude, Codex, Copilot, Cursor, Grok, Pi, Antigravity): - Add the Runner's ACPX patch (`patches/acpx@<runner acpx version>.patch`) to the pack `contractDigest`. A patch change now retires older image packs for every harness. Other packages' ACPX patches are not read. A build context without the Runner's patch fails. - Remove the ACPX patch hash from Cursor's release declaration (`cursor-contract.json`) and advance Cursor once to profile 17. Copilot, Pi and Cursor no longer bind ACPX in their own digests, like Claude, Codex, Grok and Antigravity. - Saved inputs for every harness already must match the current `acpxVersion`. - Claude and Codex digests do not change. - `generate-acpx-profiles.mjs` writes these digests to `acpx-profiles.json`. `check:acpx-profiles` refuses a stale digest. - Move the Copilot closure pins to `copilot-closure-pins.ts`, so that the generator can import them. `copilot-installation.ts` re-exports them. - Advance each profile once and then freeze it: Copilot 92, Grok 20, Pi 37. A version change at the switch keeps the rollout on the same path as every earlier profile bump. - Replace the historical version lists in `profile-compatibility.ts` with a range. Every revision from 1 through the current one decodes. Claude and Codex still decode revisions 1 through 5. Launch and recovery admission still require the exact current profile and digest. - Delete: - `scripts/copilot-profile-identity.mjs`, `scripts/grok-profile-identity.mjs` and its test - the Grok launcher source stamp - 159 Copilot, Grok and Pi identity fixtures, including three unreferenced `copilot-mainline-v*` fixtures. The Cursor fixtures stay. - Rewrite the affected tests as behaviour checks: - Each provider admits only the exact current profile. Every earlier version decodes and is refused at admission. - The last source-bound digests, and a changed digest at the current version, are refused by recovery and by the backend factory. - The Rust Pi descriptor test uses the generated profile instead of fixtures. - New `acpx-release-attestations.test.mjs`: a relocated copy of the inputs gives the same digests, each upstream input changes only its own providers, and the generator refreshes and checks the manifest. - Remove the fixture exceptions from `.dockerignore`. Add the attestation script and the generator to the Daytona image content inputs. - Update the Pi install probe and the Rust integration constant for the new Pi digest. - Update the docs to describe the current attestation model: `doc/grok-native-runner.md`, `doc/RUNNER-INTEGRATION-PLAYBOOK.md`, `doc/observability.md`, `packages/paperclip-runner/docs/durable-recovery.md`, and the Copilot, Pi and rich ACP architecture reports. Each report now opens with the current profile identity; earlier source-bound profiles remain as dated chronology. No doc references a deleted file. ## Verification - In `packages/paperclip-runner`: - `node --test scripts/acpx-release-attestations.test.mjs scripts/contracts/contract-fingerprints.test.mjs`: 19 tests pass, including a Runner ACPX patch edit that changes the contract digest, other patch edits (including `acpx@0.12.0` for `adapter-utils`) that do not, and a build context without the patch that fails. - `node scripts/generate-acpx-profiles.mjs --check` passes. - `vitest run src/drivers/acpx src/backends src/contracts`: 1647 tests pass, 9 skipped. - `tsc -p tsconfig.json --noEmit` passes. - From the repository root, `vitest run --config tests/runner-e2e/vitest.config.ts tests/runner-e2e/daytona-image.test.ts`: 11 tests pass. - In `server`, the 8 test files that use ACPX profiles, pack digests or `contractDigest`: 1617 tests pass. - Not run locally: the Rust tests and `cargo fmt`, because this machine has no Rust toolchain. CI runs them. - Before this change, a synthetic upgrade check built two Runner revisions that differ only in shared source. A Claude checkpoint saved under the first revision rebinds under the second, and a changed Claude `commandDigest` starts fresh. This confirms that the provider JSON has no other build-dependent field. ## Risks - **One reset at rollout.** The Copilot, Grok, Pi and Cursor digests and versions change once. Saved sessions for these providers start fresh on their next run, as after any earlier profile bump. After this change, Runner source changes no longer cause this reset. - **Lost "source changed, so requalify" signal.** The old digest forced a new profile version when bound source changed. Nothing enforced requalification: Copilot was last qualified at v47. Pack-boundary contract changes are now caught by the pack `contractDigest`. - **Image packs are retired once for every provider.** Adding the ACPX patch changes the pack `contractDigest`, so the controller uploads its bundled pack (about 16 s per remote start) until images are refreshed. Claude and Codex sessions are not reset: their digests and versions do not change. - **Grok launcher and closure pins.** The Grok digest covers only the launcher bytes. Grok's runtime version is still checked separately at admission, as before. - **Grok digest equals early Grok profiles.** The launcher bytes have not changed since profile 1, so the new Grok digest equals the published profile 1 digest. The profile version still separates them: launch, recovery and the backend factory require profile 20. An image pack from that period has no `contractDigest`, so the pack gate refuses it. `remote-provider-pack-verification.test.ts` now covers this case. - **Docker context checks** run `generate-acpx-profiles.mjs --check`, which now imports `.ts` pin files through Node type stripping. `materialize-pi-distribution.mjs` already relies on this on Node 24. ## Model Used - Claude Opus 5.5 (`claude-opus-5-5`) in Claude Code, with tool use (shell, file edits, local test runs). ## 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 5.5 <noreply@anthropic.com> |
||
|
DottaandPaperclip
|
1511c92965 |
fix(evals): account for superseded delegation wakes (#18319)
## Thinking Path > - Paperclip lets people ask agents to delegate work and return the result. > - The chat feature eval checks task creation, worker execution, saved output, and delivery to the source chat. > - A task can first belong to the CEO and then move to Morgan. > - Paperclip correctly cancels the old assignment wake before it starts. > - The eval treats that queue record as a failed provider run and stops before Morgan works. > - This change accounts for that record separately and preserves the complete outcome checks. > - The report also keeps partial check results when an attempt fails. ## Linked Issues or Issue Description **What happened?** The delegation journey stopped on an unstarted queue record with `issue_assignee_changed`. The CEO had created a task, assigned Morgan, saved the brief, and returned a task link. The test stopped before worker execution. Its report showed no individual checks even though the saved observation contained partial results. [Original public measurement](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/chat-progress-39e-112-latest-121-measurements-with-screenshots-40649033672/?view=case&case=chat-features.delegate-announcement). **Expected behavior** Keep the obsolete queue record and its events. Continue observing the intended worker. Pass only after three successful native work runs, a current Morgan-authored announcement, a completed task, automatic source-chat delivery, and browser readback. **Steps to reproduce** 1. Ask the CEO to have Morgan write and save a launch announcement, then return it to chat. 2. Create the task under the CEO, then assign Morgan before the queued CEO run starts. 3. Observe the cancelled old assignment. The previous eval aborts at that point. **Paperclip version or commit** Based on `6da112ab40d7b3cc0bcb79e08ddbb9b915a268a4` from `origin/master`. **Deployment mode** Disposable Product E2E fixtures. The live check uses the trusted default-branch workflow and AWS fleet. Related contribution: #17794 introduced these two journeys. The open-PR search found no separate fix for this queue-accounting defect. ## What Changed - Recognize only explicit dispatch-gate reassignment cancellations with a null start time and no native issue. Verify the company, prior CEO owner, and target task now assigned to Morgan before excluding that record from work-run counts. - Keep every queue record in the full observation and final ledger. Retain its events with an explicit reason for the absent provider log. - Let the feature flow own its complete saved-state and browser oracle. Return only actual work runs to the common provider checks. - Record saved-state checks on failures as well as successes. Capture the required `final-state.png` after the feature-specific browser readback. - Add two explicit local OpenCode cells alongside the two existing Codex subscription cells. Use the current production-qualified model. The earlier linked campaign used a different model, so this is a new measurement. - Advance the grader and suite versions. Keep the original request, output requirements, 120-second run cap, five-minute observation bound, budgets, and single-attempt policy. ## Verification - Red/green regressions cover queue accounting, missing-log collection, partial checks on evidence failure, and the required final screenshot. Flow tests invoke the actual polling callbacks: restoring the original callback reproduces both early-stop errors. Started cancellations still fail. - All 87 focused flow/grading/queue tests pass. Catalog/report checks and the 137 additional Node checks pass. Eval typechecking passes on final commit `0a6e11408c4470683417e6bf7965c73cec2a9a60`. - All current CI tests/builds pass. Greptile is 5/5 on the final commit, with no unresolved review threads. - The final commit adds only flow test coverage. Its registered definition digest equals live target `e4e0be875ee12c134fe8dd2ce15a88bfcbb100ef`: `ac6bba21d4b9cb86088be5d00df6442af90268186d6e94c9bea7d611d0488563`. - Replaying the original saved observation still fails for missing Morgan execution and delivery. Its stale queue record no longer counts as provider work. | Operation | Selected coverage | Observed result | | --- | --- | --- | | [41071710105](https://github.com/paperclipai/paperclip/actions/runs/41071710105), `b8dff816` | Backlog and delegation | Refused at the exact private-retention gate before fixtures or providers. Preserved without feature coverage. | | [41073708150](https://github.com/paperclipai/paperclip/actions/runs/41073708150), `39be9e45` | Backlog and delegation | Both passed their full product/browser oracles. Retained grades correctly fail missing `final-state.png`; the feature flow had saved the image under another name. Fixed in `e4e0be87`. Each original archive remains private and intact; later workflow concurrency superseded only the report aggregation. | | [41076854978](https://github.com/paperclipai/paperclip/actions/runs/41076854978), `e4e0be87` | Backlog and delegation | Backlog passed in 92.576 seconds, with a 47.388-second native run, complete required artifacts and passed cleanup. Delegation was interrupted by a runner-host shutdown at 15:19:23 UTC; all later evidence-upload steps were skipped. Its private archive is absent, so behavior, usage and cost remain unverified. | | [41080065077](https://github.com/paperclipai/paperclip/actions/runs/41080065077), `0a6e1140` | Delegation only | Passed on the final commit in 240.786 seconds. All 14 product checks plus browser readback passed; exact private identity, current worker output, Done state, correlated source-chat reply, required screenshot and cleanup verified. The three native runs took 42–60 seconds each. Passing control was not repeated. One-cell workflow and report enforcement passed. | The completed delegation in `41073708150` passed all 14 product checks plus browser readback in 222.593 seconds. Morgan saved the current two-sentence announcement, the task reached Done, and the CEO automatically delivered that saved text to the original chat. Its three successful native runs took 49.117, 40.060 and 39.107 seconds. The failed required screenshot remains a failed result in that measurement. All live operations use explicit IDs, concurrency one, no automatic retries, the unchanged 120-second per-run cap, and scoped private evidence retention bound to the exact workflow, target, original run, selector and launcher digest. These narrow measurements do not qualify the full catalog or establish repeated-run reliability. The broad local support suite passed 5,139 tests, with two skipped. Two database tests hit exhausted host semaphore capacity after the installed Postgres package's missing library symlinks were repaired. Their recheck reported `could not create semaphores: No space left on device`. No host IPC resources or kernel settings were changed. The corresponding full CI tests passed on a fresh host. ## Risks - A broad cancellation exception could hide failed work. This exception requires the dispatch gate's exact receipt, a null start time, the expected company and prior owner, and the actual reassigned task. Started and unproved cancellations still fail. Backlog creation permits no task wake. - New manual OpenCode cells can spend provider money only when explicitly selected. Default scheduled selection is unchanged. - This changes eval interpretation and evidence collection. It does not change product scheduling or task ownership. ## Model Used OpenAI Codex, GPT-6 family, with code execution and tool use. The exact model ID and context-window size are not exposed to 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
|
5a7a6b32ba |
fix(ui): restore compact Coordinator task loading and responder continuity (#18312)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - People open tasks from their personal workspace and send replies to agents. > - Cold task opens should load recent conversation before older history. > - Classic settings could expand history before the Coordinator preference arrived. > - Coordinator lists also read full tasks, and response startup used the previous assignee. > - This pull request restores compact reads and keeps the selected responder visible through runtime discovery. > - People can read and reply without waiting for unrelated history or losing response feedback. ## Linked Issues or Issue Description Related: #17593, #17751, #17756, #17885, #17983, and #18092. **What happened?** A cold task could select Classic full history while its personal workspace preference was pending. The full-history choice stayed cached after Coordinator became active. Coordinator Tasks used full task reads. Needs attention fetched one enriched task per pending task. Sending to a newly selected agent could show no responder, show the previous assignee, or clear feedback when assignment changed. Legacy runtime discovery then replaced the startup identity with a generic activity line. **Expected behavior** Wait for the workspace preference before selecting Classic history. Show the task instruction and recent conversation with **More messages** for older content. Read collection indexes and fetch diagnostics for visible rows. Keep the selected responder mounted from Send through saved input and runtime startup. **Steps to reproduce** 1. Enable Classic task UI and Coordinator. Delay the personal workspace preference on a cold task open. 2. Open Coordinator Tasks or its Needs attention view with many pending tasks. 3. Open a task assigned to a person. Select another agent in the composer and send a message. 4. Let the message save before runtime reads return. Then allow a legacy run to appear. ## What Changed - Wait for task interface settings before persisting Classic history expansion. - Use lightweight indexes for Coordinator Tasks. Resolve pending task IDs in batches of at most 100. Keep search and participant filters on the server. Enrich at most 50 visible rows through the existing hook without replacing list membership or order. - Preserve an error when a pending task cannot load, instead of showing a false all-clear. - Register board cards for row enrichment. Observe newly mounted cards after column expansion or pagination, and track visibility by element so hidden list rows cannot clear visible board cards with the same task ID. - Track the recipient of each submitted message through reassignment. A later draft's recipient does not change the response already starting. - Share a stable response header across startup, native runtime, and legacy runtime rendering. Keep the existing transcript renderers. - Add regression tests, production-browser checks, design guidance, and feature-map notes. ## Verification - Red/green: delayed Coordinator preference started a full comments read before the fix. It now retains compact history and still expands on request. - Red/green: sends from human or other-agent assignment showed no identity or the wrong identity. Both now retain the selected responder through assignment, acknowledgement, later draft selection, and native/legacy startup. - In a 105-task Needs attention fixture, the previous code made 105 enriched detail requests. The fixed code makes two bounded index requests. - All 659 final affected tests pass across task detail, thread, runner turn, history window, Inbox, board, row enrichment, and Coordinator collection suites. - Review follow-up: the production board lost its next-step badge before card registration. The regression now passes, along with column pagination. The focused 82-test enrichment run and 371-test header/thread run also pass. - All 12 compact-startup browser checks pass against the production UI and an isolated PostgreSQL container. They cover delayed preferences, list navigation, saved-card geometry, document/comment anchors, history expansion, draft retention, retry, and actual legacy process startup. The final deterministic process case recorded 118 frames with zero missing or replaced responder identities. It does not call a model provider. - Full workspace build and typecheck pass. Final UI build and typecheck pass after the shared-header change. Token gates, map checks, module boundaries, and diff checks pass. - The full UI rerun before the board review follow-up passed all 10,034 tests across 844 files. The 659 affected tests pass after that follow-up. The first full UI run passed 10,030 tests and timed out in one unchanged IssuesList large-tree test. Its two variants pass in an isolated rerun and in the full rerun. Those later passes establish recovery; the earlier timeout's cause remains unproven. - The full `pnpm test:run` attempt was stopped after repeated embedded PostgreSQL initialization failures in server fixtures. No complete local server-suite pass is claimed. Browser verification recovered with a disposable container database. - A later avatar browser rerun reused the previous same-title task and its cancelled run through the existing 48-hour title deduplication. That fixture never reached fresh runtime startup. The test now uses a unique title and checks its initial assignment/status; a full 12-case rerun and a second avatar run on the same instance pass. The failed trace is retained locally. - Final commit `7a18372da3` has 55 successful checks, two intentionally skipped Storybook checks, and one neutral Paperclip Review. All required verification gates are green. Greptile is 5/5 with no actionable findings or open review threads. The advisory Paperclip review paused without a complete assessment; no passing Paperclip assessment is claimed. - These checks establish bounded reads and stable rendering. No reliable staging latency benchmark or staging deployment is claimed. ## Risks - Collection projections omit task bodies. Detail navigation must retain its normal core/full task reads. Search remains server-side. - Startup feedback must end on rejection, unavailable agents, terminal outcomes, or a visible run. Existing cases cover those transitions. - The shared header changes the legacy live-turn layout to retain identity beside its existing status line. - No migration, mutation permission, provider protocol, or execution authority change. ## Model Used OpenAI Codex, GPT-6 family. The exact serving deployment ID and context-window size are not exposed in this session. The model used reasoning, code editing, shell tools, automated Chromium checks, and manual browser inspection. ## 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 suites and browser checks; full-suite limitations are recorded 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> |
||
|
DottaandPaperclip
|
ac4f13d205 |
fix(e2e): report committed child blockers promptly (#18306)
## Thinking Path > - Paperclip manages AI agents and their work. > - Product E2E checks the ordinary onboarding conversation and its delivered work. > - A native run can finish successfully while its task remains blocked for a human. > - The onboarding journey kept waiting for such a child until the campaign deadline. > - This pull request reports the committed child blocker when no owned work remains pending. > - The journey still requires the same independent checks before it can pass. ## Linked Issues or Issue Description **What happened?** An approved child can stop with a committed current-track blocker owned by the board. Its native run succeeds, but the child task stays blocked. The onboarding journey reports a deadline failure instead of the observed child blocker. **Expected behavior** Report `onboarding_child_task_blocked` when the latest child run commits that blocker at the child's current status version. Keep waiting for parent dependencies and pending native or workspace recovery. **Steps to reproduce** Use the existing onboarding fixture with an approved blocked child. Give its latest native run a succeeded status, committed finalization, blocked work disposition, and the matching task status version. The old journey waits until its deadline. The new journey fails at that settled blocker without downloading or accepting any delivery. Related public context: #18298. Public issue and PR searches found no duplicate blocked-child harness fix. This change reports the failure boundary; it does not repair artifact delivery. ## What Changed - Check the owned child's board blocker after the existing run and recovery checks. - Require a unique latest producer, native committed results, and the current task status version. - Preserve parent dependency waits and all artifact, document, download, link, scope, and budget checks. - Add a behavioral regression and controls for active work, native recovery, workspace replacement, stale versions, newer producers, and agent-owned blockers. - Document the distinct failure code and the existing control-plane writers for its metadata. ## Verification - The new behavioral regression failed on the old code with `onboarding_first_task_deadline`; 147 controls passed. - The current head passed all 148 onboarding tests, including the valid agent-owner object control. - The initial implementation passed 305 tests across all 10 Product E2E files. The later update changes only writer links and the owner control; its 148 focused tests and harness typecheck also pass. - `pnpm test:e2e:runner:typecheck` passed after building the isolated checkout's workspace declarations. - `git diff --check` passed. - All required CI checks passed on `1b3bd04d193fc98d0d9f2eb3b826a4ef4bdb6dea`, including full tests and build. They were not repeated locally. Greptile reviewed that exact head at 5/5. - The full Cloud journey remains unverified by these source and CI checks. This change does not claim that the underlying artifact delivery issue is fixed. ## Risks The early failure depends on the public task and run projections. Missing, stale, or ambiguous producer metadata keeps the existing bounded wait. A pending recovery or newer producer prevents the old blocker from ending the journey. No success requirement changes. ## Model Used OpenAI Codex, exact model ID `gpt-6.1-sol`, reasoning effort `ultra`. The model used source inspection, code editing, and local test tools. The runtime did not expose its context window size. Root supplied the model and effort metadata for this agent 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
|
0c70cedce9 |
Wait for the saved Studio file before editing a skill (#18305)
## Thinking Path > - Paperclip lets people manage AI agents and use their saved work. > - Product E2E journeys check these user paths in classic and Cloud environments. > - The shared Skill Studio helper edits a skill created by an agent. > - Studio can mount the editor before its saved file finishes loading. > - Typing into that empty draft can replace or lose content when the file arrives. > - This change waits for the existing saved body before it starts the edit. > - Browser regressions check delayed loading and refusal to edit an unloaded file. ## Linked Issues or Issue Description Refs #18222 for the shared Product E2E journeys. No duplicate hydration fix was found. **What happened?** The shared Studio helper started typing as soon as the editor control was visible. A delayed file query could replace that draft. The helper could then lose the original content or fail to save its complete marker. **Expected behavior** The helper must load the original file before editing it. It must preserve that content, save the new marker, reload, and read the result from the task card. **Steps to reproduce** Run the controlled Studio browser fixture with a 1.5-second file load. The old helper saves the marker without the original `Verify checks.` body. With no file load, it still sends input events to the empty editor. ## What Changed - Wait for the original saved body before typing in the shared Studio helper. - Add browser checks for immediate load, delayed load, and an unavailable file. - Include these checks in the credential-free browser support suite. - Document the saved-file readiness requirement. ## Verification - Against the old helper, both new failure cases fail: the delayed case loses the original body, and the unloaded case sends 13 input events. - With the change, all three browser checks pass. They verify the saved original body and marker, reload, task-card readback, and zero input or save calls for an unloaded file. - `node node_modules/typescript/bin/tsc -p tests/runner-e2e/tsconfig.json --noEmit` passes. - `git diff --check` passes. Full repository CI passes in [41058312896](https://github.com/paperclipai/paperclip/actions/runs/41058312896), including typecheck, tests, build and all eight E2E shards. Greptile reviewed this exact head at 5/5 with no open threads. No live acceptance result is claimed by these controlled fixtures. ## Risks This adds a readiness assertion to an existing test helper. Both callers already use the same skill fixture and verify `Verify checks.` in the task card. An unavailable file now fails before an edit can occur. The change does not alter application behavior, model instructions, grading, or deadlines. ## Model Used OpenAI Codex, GPT-6 family, with reasoning and code execution. The host does not expose the exact underlying model version 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 described the issue in-PR following the bug template - [x] I have not referenced internal/instance-local Paperclip issues or links - [x] My branch name describes the change and contains no internal ticket id - [x] I have run tests locally and they pass - [x] I have added tests where applicable - [x] I have updated relevant documentation - [x] I have considered and documented risks - [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 reviewer comments before merge Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
cd5f930558 |
Add shareable image galleries to task documents (#18157)
<!-- entire-trail-link-start --> https://entire.com/gh/paperclipai/-/trails/247/changes/paperclip/247 <!-- entire-trail-link-end --> ## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Task documents hold plans, reports, and image evidence. > - Task attachments already have an image gallery. > - Inline document images had no gallery navigation or image links. > - This change gives each document a gallery that uses the existing modal. > - People can review the image sequence and share a link to one image. ## Linked Issues or Issue Description **What existing behavior does this improve?** Image review in task documents, including the side panel and Classic interface. **Current behavior** Inline document images do not open a gallery. People cannot move through them with arrow keys or share a link to one image. **Proposed behavior** Select an image to open the gallery. Use arrow keys or Previous and Next. The URL records its one-based position, for example `#document-report&image=2`. Reloading that URL opens the selected image. Revision links keep their exact saved content. **Reason and benefit** People can inspect a document's image evidence in reading order and share the exact view. No matching issue was found. The latest 300 open PR titles were checked for related document gallery work. ## What Changed - Add a document renderer that collects rendered images in reading order, including reference-style and repeated images. - Reuse the gallery modal with optional index-change and focus-return controls. - Keep the document owner in the saved tab identity when people switch between local and other-task documents. - Match explicit current-task document owners with local document tabs. Keep review and viewer parameters during gallery navigation. - Treat another Classic review receipt as a new request while image navigation retains the selected revision. - Make inline image controls work with a keyboard. Keep focus in the modal and restore focus on close. - Keep document, revision, review, and maximized-view link parameters during image navigation. - Fix a Classic revision-cache race found by the browser test. Recheck a pinned revision when the saved document head changes. - Update the feature and agent access maps. No new server API or permission is needed. ## Verification - Passed 119 focused tests for document galleries, the shared gallery, document rendering, the revision panel, and hash routing. - Passed all 3 Chromium journeys: desktop, mobile, and Classic. These use real uploads and saved document revisions. - The browser journeys also check explicit current-task owners, review parameters after reopening, other-task links after tab switches, and a second Classic review link after Return to latest. - Browser checks cover keyboard and button navigation, wraparound, shared-link reload, exact historical content, invalid positions, focus return and containment, scroll locking, and Escape, Close, and backdrop dismissal. - UI typecheck, token gates, map checks, and diff whitespace checks passed. - The UI production build passed in the browser harness. - Earlier browser runs found and then verified the Classic revision-cache fix. A later build exited 137 while the full test suite ran. The browser run passed after that suite stopped. - Full repository typecheck and build could not complete because Cargo is absent. The full repository test run did not complete within this sandbox. These checks remain subject to CI. ## Risks - Image positions follow document reading order. Editing an unpinned document can change those positions. A revision link retains the saved sequence. - The shared modal now uses Radix Escape handling. The existing gallery tests and nested mobile dialog journey pass. - No schema change or migration. ## Model Used OpenAI GPT-6 through Codex. The runtime did not expose a more specific model ID or context window size. Used reasoning, code editing, shell commands, and browser automation. ## 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
|
8f50704135 |
fix(e2e): wait for pending native recovery (#18298)
## Thinking Path
> - Paperclip manages agents and their work.
> - Native runs can retain a bounded automatic recovery on the same run.
> - Product E2E observes those runs before it checks the delivered work.
> - The onboarding journey treated every failed run as a final failure,
even when the server had saved its next recovery attempt.
> - This pull request waits only for the explicit agent-owned native
recovery projection.
> - The journey still requires successful execution and the same
independent delivery checks.
## Linked Issues or Issue Description
**What happened?**
The onboarding journey can stop after a native turn completes without
its semantic result. The server marks that attempt failed and saves a
same-run retry. The journey rejects the failed row before the saved
retry is due. This also affects a native child that follows a proven
workspace deferral.
**Expected behavior**
Observe the saved automatic recovery within the original campaign
deadline and run cap. A pending recovery cannot satisfy completion.
Ordinary failures and exhausted, terminal, or board-owned recovery must
still fail.
**Steps to reproduce**
Use the existing onboarding fixture with a native failed row whose
nativePhase is retryable_failure. Give it the server-owned
native_session_interrupted cause, an agent-owned retry_scheduled
projection, a valid retry time, and attempt 1 of 3. Transition that same
row through reconnecting and running to success. The old journey rejects
the first row.
Related public context: #18292. A search of related native-recovery PRs
found no duplicate harness fix.
## What Changed
- Observe the authenticated native recovery state in the existing
workspace-deferral module. Require matching nonempty cause strings and
every owner, phase, attempt, schedule, deadline, scope, and successor
guard.
- Apply it to the proposal wait, failed native rows, workspace successor
tails, and final settlement.
- Keep defensive rejection of known permanent, permission-declined, and
exhausted causes. Diagnostic cause strings are not a positive retry
policy.
- Add behavioral recovery, deadline, and rejection controls to the
existing workflow tests.
- Document that this wait preserves the original campaign and final
delivery requirements.
## Verification
- Baseline workflow regression: 4 expected premature failures with 90
controls passing. The final structural calibration rejected 20 valid
pending-recovery cases before the correction, with 120 controls passing.
- Final regression: 140/140 pass, including existing and new diagnostic
causes on direct and workspace-successor paths. Known terminal,
ambiguous, malformed, mismatched, and exhausted states still fail.
- Focused Product E2E compatibility: 297/297 pass across 10 files.
- pnpm test:e2e:runner:typecheck: pass after building the normal SDK and
Runner TypeScript dependencies.
- git diff --check: pass.
- Final source passed these checks before the reviewed commit.
- Full repository tests/build and live Cloud acceptance were not run
locally for this harness-only change. Required CI run 41049692194 passed
on
|
||
|
DottaandPaperclip
|
8035f8b245 |
test: require saved onboarding approval cards (#18292)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Product E2E tests exercise real user workflows through the browser. > - The onboarding journey requires a human to accept one proposed child task. > - A live agent tried to create the child before it saved a proposal card. The product correctly rejected that action. > - The scripted request named the required result but did not name the proposal tool or require its saved receipt. > - This pull request makes that handoff explicit. It keeps the existing human approval and independent result checks. > - The benefit is a clearer scripted workflow without a change to product authority. ## Linked Issues or Issue Description Refs #18288 and #18266. **What happened?** The agent treated the opening task request as a step toward acceptance without first saving a proposal. It then stopped after child creation was denied. Its visible response referred to a pending proposal, but no proposal card existed. **Expected behavior** The agent saves a native confirmation card, checks its receipt, and waits for the human to accept it through the browser. It must not record its own approval. Only an accepted card authorizes the child task. **Steps to reproduce** Run the scripted Cloud onboarding journey. Give the canonical onboarding task its first task request. The observed agent skipped the saved proposal and tried to create the child. Model behavior varies, so this is a fixture guidance change rather than a deterministic product defect claim. **Paperclip version or commit** Fixture base: `7a7ade4b9360aaf8e1a2808bb2b1a752c080f564`. The serving product was Core `965184d2968f40b1183803529b88f288264c832f`. **Deployment mode** Testing Cloud workspace with native Codex. This pull request selects no product release or deployment. ## What Changed - Name the existing `request_human_input` confirmation or revision-bound checkbox contract and `wake_assignee` policy in the parent request. - Require a saved pending-card receipt. Require corrections when tool arguments are rejected. - Forbid self-approval. Require the actual accepted human card decision before child creation. - Document this scripted handoff. Keep the child brief, final download links, browser actions, and all grading assertions. ## Verification - Focused onboarding and Cloud journey tests: 79 passed. - Shared evidence, campaign, and contract tests: 66 passed. - `pnpm test:e2e:runner:typecheck`: passed after building the existing workspace type dependencies. - Credential-free preview under a network-deny guard: passed with zero external calls. Its preparation plan matches the base fixture. - `git diff --check`: passed. - No prompt-string test was added. Existing tests continue to enforce approval order, child scope, independent file checks, and real result links. - Full local repository build and database suites were not repeated for this fixture text change. Required PR CI must pass before merge. - The original live result remains unchanged. This local work does not prove that a later model run will comply. ## Risks - The model can still ignore the instructions. A new real campaign must prove the workflow. - More explicit guidance makes this a scripted product journey. It is not a model discovery benchmark. - There are no product runtime, schema, credential, model, budget, or approval-policy changes. ## Model Used OpenAI Codex, exact model `gpt-6.1-sol`, ultra reasoning effort. Parent runtime metadata confirmed this model. Context window size was not reported. Used source inspection, local code execution, and an independent agent review. ## 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
|
7a7ade4b93 |
test: require real onboarding result links (#18288)
## Thinking Path > - Paperclip lets people assign work to agents and use their saved results. > - The Cloud onboarding journey tests this handoff through a real parent and child task. > - The child can publish a correct file while the parent names that file without linking it. > - The existing journey correctly rejects that reply because the user cannot download the named result from it. > - This pull request makes the scripted parent request explicit about real Markdown links and returned attachment paths. > - The independent artifact and parent-link checks remain unchanged. ## Linked Issues or Issue Description Refs #18266 and #18258. **What happened?** A Cloud onboarding attempt produced a correct child artifact and saved document. The parent then named the child, document key, and filename in inline code. It omitted the attachment URL. The attempt failed with `onboarding_parent_result_links_missing`. **Expected behavior** The parent's final reply must link the saved child results. The file link must use the real published attachment reference. **Steps to reproduce** Run the existing explicit Cloud onboarding journey. A reply that only names the file must remain a failure, even when the child artifact is correct. **Paperclip version or commit** The branch starts from `965184d2968f40b1183803529b88f288264c832f`. ## What Changed - Ask the parent for real Markdown links in its final reply. Use the published attachment's returned content or download path. Do not invent IDs or URLs. - Add a filename-only reply to the existing negative oracle cases. - Document this scripted fixture requirement. ## Verification - Focused onboarding tests: 79 passed. - Shared task, campaign, and contract tests: 66 passed. - Existing artifact and native-completion Node checks: 137 passed. - `git diff --check`: passed. - Harness typecheck, `pnpm -r typecheck`, and `pnpm build`: passed after building the existing workspace dependencies. The initial missing-declaration and local-IPC setup failures remain recorded. - The full local database test suite was not rerun. Required repository CI runs the broader suites on Linux. - Required repository CI passed on the exact PR head. No paid model calls or Cloud operations were run for this patch. - [Initial CI](https://github.com/paperclipai/paperclip/actions/runs/41027635545/attempts/1) failed one unchanged retention timing test (3310 ms versus the strict 3000 ms bound). The unchanged case passed once on an isolated local Postgres17 fixture. One failed-jobs-only rerun passed in [attempt 2](https://github.com/paperclipai/paperclip/actions/runs/41027635545/attempts/2). The original failure remains visible. No timing threshold or retention source changed. - The previous live attempt stays failed. Local tests do not prove that a model follows the revised request. A new bounded live attempt must establish that result. ## Risks - The extra instruction can still be ignored by a model. The unchanged oracle will reject missing links. - This changes the fixture definition hash. Historical attempts retain their original source and result. - No product runtime, API, approval policy, model profile, budget, or cleanup behavior changes. ## Model Used OpenAI `gpt-6.1-sol`, ultra reasoning, with code execution and repository tools. Model metadata was confirmed in the parent task. The context window size is 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 `#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
|
81ebe479e8 |
fix(evals): bind rendered Markdown replies to saved run evidence (#18264)
## Thinking Path > - Paperclip lets people assign work to AI agents and read saved results. > - Product journeys must prove that a result is visible and belongs to the completed run. > - The shared observer compares stored Markdown with browser text. > - A document link can render as an issue chip, so those strings differ. > - This PR keeps run and comment attribution, then compares full text and exact link destinations. > - The benefit is a valid browser result without a false timeout or a weaker provenance check. ## Linked Issues or Issue Description Refs: #18248, #18255. **What happened?** The protected-action story completed both decisions and all four native turns. Its final explanation contained a document link. The browser rendered that link as an issue chip. The observer waited for raw Markdown to equal rendered text, so it could not accept the valid result. **Expected behavior** Accept complete matching content from the final run after reload. Keep the exact link destination, task anchor, author, run and durable final-message checks. Reject missing or changed content. **Steps to reproduce** 1. Save an attributed reply with bold text and a Markdown document link. 2. Render the link as an issue chip in the task bubble. 3. Run `readBoundTaskFinalReply`. The original observer times out. The new browser regression preserves that failure. ## What Changed - Add a UI-owned content projection using its existing Markdown renderer and GFM/issue-reference plugins. Keep DOM observation, attribution and destination checks in the task-reply observer. Keep source Markdown and observed text separately. - Read visible DOM text, link labels and destinations. Require the structured projection to match the actual visible text. Allow only recognized product chips to replace a link label. Ordinary links keep their source label. Reuse the production issue-reference and company-route helpers, plus the current task URL. Ordinary root-relative anchors keep their original path. Preserve exact marker checks and all run/comment/native-event attribution checks. - Check visible decline explanations and unread data in preserved native source text. - Add browser regressions and source/content tests. Reuse the UI’s existing `react-markdown` and `remark-gfm` dependencies. The final diff changes no package manifest or lockfile. - Match generated GFM checkbox states and reject changed states. Document supported content and fail-closed behavior. ## Verification - Original observer: the decorated-link browser regression fails at the final-reply predicate. - Fixed observer: all 21 browser regressions pass, without credentials or model calls. - Focused observer, decline, content and journey contract suites: 64 tests pass. - E2E and UI TypeScript checks pass. All UI token gates pass. The unchanged frozen install passes. - Full CI passed on the preceding |
||
|
Nicky LeachandClaude Opus 5.5
|
fa760dbce4 |
fix(runner): reuse image packs only with matching contract digest (#18237)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Remote native runs use a provider pack. The controller reuses a protected pack from the sandbox image when the pack is compatible. > - #17749 gates this reuse on a hand-maintained interface revision in `provider-pack-contract.json`. > - Later changes altered the runner/pack boundary (#17922, #18188), but nobody bumped the revision. A stale Claude or Codex image pack can still be reused across a sidecar contract change. This is the failure class of #17972. > - A manual revision has a "forgot to bump" failure mode. Source-byte hashes change on every refactor, so they are too brittle. > - This pull request declares the sidecar `session.open` contract in its JSON schema and computes a contract digest from the JSON contracts and Rust shapes. > - Image-pack reuse now requires an equal digest. > - The benefit is that an incidental change keeps image packs reusable, and a real contract change retires them automatically. ## Linked Issues or Issue Description Refs #17749, #17972, #17922, #18188. **What happened?** The controller reused an image provider pack when `compatibility.acpx` and the selected agent's profile digest matched. For Claude and Codex, that profile digest binds only the upstream executable. The ACPX interface revision stayed at 1 after later sidecar contract changes. A pack built before those changes passed the reuse check. **Expected behavior** The controller reuses an image pack only when the pack uses the same runner/pack contracts as the controller's bundled pack. Comments, formatting and refactors do not change that result. **Steps to reproduce** 1. Build an image pack from a commit before a sidecar input change, such as a new `session.open` field. 2. Run a controller from a later commit that sends the new field, with the same interface revision and Claude version. 3. Before this fix, `verifyRemoteProviderPack` accepts the old image pack. ## What Changed - Declare the sidecar `session.open` parameter set as a closed `sessionOpenParams` definition in `acpx-sidecar.schema.json`. - `generate-acpx-sidecar-contract.mjs` now generates `GENERATED_ACPX_SIDECAR_OPEN_PARAMS` from it. - `parseOpenParams` uses the generated list instead of an inline `Set`. The accepted set is unchanged. - This is the surface of #17972. The schema is now its source of truth. - Declare the nested `session.open` contracts that the sidecar parses strictly: - `protocol/provider-schemas/runtime-context.schema.json` closes every runtime context object. `parseNativeRuntimeContext` uses the field lists that `scripts/generate-runtime-context-contract.mjs` generates from it. - The sidecar schema declares the closed `providerPolicy` object, and `parseProviderPolicy` uses its generated field list. - The accepted fields do not change. - Add `packages/paperclip-runner/scripts/contracts/contract-fingerprints.mjs`. It fingerprints these contracts: - the PRP and sidecar JSON schemas, with annotation text removed - the generated catalogs in `generated/` - the Rust serde shapes, `json!` payload keys, match literals and env keys in `runner-core/src`, without test-only modules - the reserved env key lists in `configured_environment.rs` - Read only the trees that every pack build context contains. Fail when a registered source is missing, so two build contexts cannot silently produce different digests. - Use only Node built-ins. The registry adds no dependency. - Write `contractDigest` as a top-level payload field in `build-provider-pack.mjs`. The manifest digest algorithm and the existing field layout do not change. - Require an equal `contractDigest` for image-pack reuse in `remote-provider-pack-verification.ts`. A pack without a digest is not reused. The controller uploads its bundled pack, as it does today for any incompatible pack. - Re-attest the source-hashed profiles, because this PR changes their bound sidecar sources: - Grok becomes profile 15 and now also binds the generated allowlist, so a Grok field change in the schema still invalidates the Grok identity. - Copilot becomes profile 88, and also binds the generated runtime context fields. - Earlier profiles stay decodable. - Add the registry script and `generated/` to the Daytona image content inputs. - Document the digest and the schema-first rule in `docker/daytona-runner/README.md` and `doc/DEVELOPING.md`. ## Verification - `node --test scripts/contracts/contract-fingerprints.test.mjs` (in `packages/paperclip-runner`): 4 tests pass. - These changes keep the digest: - schema description text - schema reformatting - renaming `parseOpenParams` - Rust comments, `#[serde(skip)]` fields and `#[cfg(test)]` modules - edits to the `#[path]` test module `tests/provider_api_validation.rs` - These changes alter the digest: - a new `session.open` schema field - a changed constraint on the `remainingWork[].description` property - a new reserved env key - a new sidecar env key - a new checkpoint field - A build context without a registered source fails. - `node --test test/acpx-sidecar-contract.test.mjs`: 13 tests pass. A new case rejects an unknown `session.open` parameter and accepts each declared parameter. - `node --test scripts/grok-profile-identity.test.mjs`: 4 tests pass. A change to the Grok field in the generated allowlist invalidates the Grok identity. - `generate-acpx-profiles.mjs --check`, `generate-acpx-sidecar-contract.mjs --check` and `generate-protocol-manifest.mjs --check` pass. - Runner vitest suites pass: 470 passed, 1 skipped. They cover the sidecar, sidecar input and protocol, Copilot and Pi installation, profile compatibility, native execution, recovery identity, the backend factory and the environment. - Server tests: pack verification, executor and artifact preparation pass (1,004 tests). - `tsc --noEmit` passes for `packages/paperclip-runner` and `server`. - `tests/runner-e2e/daytona-image.test.ts`: 10 of 11 tests pass. The failing test also fails on `master`. It expects a literal `profile-v19-identity.json` line in `.dockerignore`, which uses a `profile-v*` glob. - Not run locally: - A complete `build-provider-pack.mjs` run. The builder rejects the local Homebrew Node because it is not portable after relocation. - Rust tests and `acpx-controller-recovery.test.ts`, because `cargo` is not installed. - The full `pnpm -r typecheck`, `pnpm test:run` and `pnpm build`. ## Risks - Only JSON contracts and Rust shapes are part of the digest. The remaining TypeScript-only definitions are the `paperclip_finish` and `paperclip_block` tool schemas in `completion-result.ts` and the TypeScript env key lists. They are not covered until they move into a schema. The nested `expectedIdentity` and `tools` values ignore unknown fields, so a change there cannot cause a startup rejection. - A change there can let a stale image pack be reused. The sidecar then rejects the input with its strict parser error. - Today, Claude and Codex image packs have no contract check at all. - The digest covers all registered contracts for the whole pack. A contract change for one provider kind also retires image packs for the other. Each replacement costs about 16 s per remote start until the image is refreshed. - Existing image packs have no digest. After deploy, the controller uploads its bundled pack until the image is rebuilt. - The Grok and Copilot re-attestations start fresh sessions for those agents once after deploy, as each profile bump does today. - The Daytona image content ID formula changes, so the next image build does not reuse a cached image. > 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 Anthropic Claude Opus 5.5 (`claude-opus-5-5`) through Claude Code in the desktop app, with repository tools, shell execution and research subagents. The session does 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 - [ ] 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 5.5 <noreply@anthropic.com> |
||
|
DottaandPaperclip
|
145214201d |
test: preserve the onboarding child brief at delegation (#18266)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Product E2E journeys check onboarding through the real app. > - The onboarding story asks a parent agent to delegate a file and document task. > - A parent can preserve the goal but omit the exact tool instructions from its child description. > - The child can then publish a note instead of the required JSON file and task document. > - This pull request asks the parent to copy the full child brief without changing it. > - The existing independent checks still decide whether the result passes. ## Linked Issues or Issue Description Refs #18258. Includes merged prerequisite #18270 for the existing retention scheduling failure found by required CI. **What happened?** The scripted onboarding request gave the parent an ordered tool recipe. It did not require the parent to keep that recipe in the child description. A delegated child could compute the result but publish a Markdown note and omit the Paperclip task document. **Expected behavior** The child receives the full approved brief. It must publish `connection-proof.json`, save its Paperclip task document with the returned file link, and report completion through the current contract. **Steps to reproduce** Run the explicit Cloud onboarding journey. Inspect the approved child description and its published result when the parent summarizes the request. **Paperclip version or commit** This change is based on public master `b3c2172d417ea6b783e0c796aaed4a91e1f9d381`. **Deployment mode** Scripted Product E2E against a live Cloud testing workspace. ## What Changed - Put the existing ordered instructions in one marked child brief. - Require the parent to copy that brief verbatim into `create_task.description`. - State the exact JSON filename and workspace-relative content reference for `register_deliverable`. - State that a Markdown workspace note does not replace the `write_document` step. - Document this scripted scope in the Product journey guide. ## Verification - Existing focused Product journey tests: 227 passed in 10 files. - Credential-free Product journey discovery passed through the direct tsx loader. - Source comparison confirms that all code after the opening reply is unchanged. The download, calculation, hash, provenance, document, parent-link, approval and run-limit checks retain their original behavior. - No prompt word-matching tests were added. This is an instruction-only change. - `pnpm build`, `pnpm -r typecheck`, and `pnpm test:e2e:runner:typecheck` passed. - Full support Vitest check: 5,042 passed, 2 skipped, 2 failed. Both failures occur during embedded PostgreSQL startup in the controller database lifecycle fixtures. The log includes a missing bundled `libzstd.1.dylib`. The separate Node support checks passed all 137 tests. - `pnpm test:run` was attempted. It reached the general server suite and repeated embedded PostgreSQL startup failures. I stopped only this owned local test process group. The full local test command is incomplete. Required Linux CI remains the merge gate. - The first harness typecheck ran before workspace build outputs existed and reported missing workspace modules. It passed after the full build. The earlier failure remains recorded. - The initial CLI discovery wrapper hit a local tsx IPC permission error. The direct loader passed with the same source and arguments. - The original required CI run found an existing skill-retention page delay. Its unchanged test passed once against a dedicated local PostgreSQL fixture. Separate PR #18270 added a deterministic regression and fixed the scheduler. After that PR merged, this branch received one normal merge from master. Its diff remains two fixture files. The original failed CI run remains preserved. - After the branch refresh, all 70 onboarding first-task tests and 7 retention/deadline controls passed. Harness typecheck passed. Fresh required CI and review passed on head `608e266c6567c19981fd3f8a07305c785346753b`. - The refreshed CI run passed the retention shard but failed one unchanged Antigravity authentication fixture. The same Runner lane passed on the prior fixture head and the merged prerequisite. The exact test passed once locally. Source review identified a possible stdout/stderr ordering race, but the failed run did not retain its exact error subtype. The one approved retry of the failed CI job and its verification aggregate passed on unchanged source. The original failure remains preserved. This proves recovery, not its cause. - The live Cloud journey has not been rerun with this source. Historical failed attempts retain their original results. ## Risks - An agent can still fail to follow the brief. Existing independent assertions will reject the wrong file or missing task document. - The fixture definition changes. A later live attempt must record this new harness revision. - No application, model, effort, budget, cleanup or acceptance rule changes. ## Model Used - OpenAI `gpt-6.1-sol`, ultra reasoning effort, with repository tools and code execution. The parent verified this model ID from current runtime metadata. The context window size is 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 `#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
|
b3c2172d41 |
fix(evals): verify native protected-action outcomes (#18255)
## Thinking Path > - Paperclip manages work assigned to agents. > - Product E2E stories ask for useful user-facing results. > - Classic runner and deployed QA reuse the same service approve/decline task prompts. > - The service returns a receipt marker that belongs in a saved briefing document. > - State that destination explicitly and ask for the document link when complete. > - The deployed decline path also needs the same durable final-reply proof as Plan and Stop. ## Linked Issues or Issue Description Refs #18222. Builds on the merged reply observer in #18248. **What happened?** The shared service tasks ask for titles and a “verification code” without specifying that the code belongs in the document or asking for its final link. A native QA attempt created its briefing but then repeatedly failed the completion tool’s array validation. Those invalid arguments remain a separate observed failure; the wording change does not establish their cause or relax validation. Source inspection also found that the declined-action oracle requires a projected comment even when the settled native thread renders a durable final message. **Expected behavior** Ask for page titles and the service receipt code in the saved briefing, then a document link. After a declined action, require the existing explanation-only fallback without another service call. **Steps to reproduce** Run the shared service-approve/service-decline stories through the classic runner or QA Product E2E adapter. ## What Changed - Clarify the natural-language task for both shared service stories: put page titles and service receipt code in the document and provide its link. - Use the reloaded run-bound final reply for QA service declines. Pass an explicit native receipt to the existing decline oracle; do not synthesize issue comments. Require the final native run to start after the saved decline and keep exact decision, truthful explanation, no unread marker, zero calls and unchanged connection checks. - Keep classic callers on their existing attributed-comment path. No internal tool instructions are added to the user prompt; ownership, service data and model/runtime limits stay strict. ## Verification - Decline regression observed RED: two cases reject a valid saved native explanation without projected comments. After the repair, 79 focused tests passed across native reply, decline guidance, agent evidence and journey contracts. A two-successful-run control rejects the earlier receipt and accepts the later one. Negatives cover stale/wrong run, wrong agent/task, missing final proof, summary/unsettled/hidden output, false explanations and unread service data. - Product E2E TypeScript and `git diff --check` passed. - Live protected-action qualification is required after merging this and the companion private service/cleanup fixes. The previous failed result remains failed; this is a bounded task-clarity experiment, not a claimed model or completion-schema repair. - Current-head Greptile and Paperclip reviews are 5/5 with zero open threads. CI initially failed the unrelated agent-chat-projects cleanup POST with a PostgreSQL lock timeout after its assertions completed. One bounded retry passed on unchanged source; all current-head gates passed at `64bc169e147290ed4dbaa18221e0268362ce81ee`. The original failing attempt and its logs remain available. The task-clarity change is still a bounded experiment; the observer repair addresses a demonstrated storage-assumption defect. ## Risks A clearer task can still fail with the selected model. All acceptance checks and completion input validation remain strict; a saved briefing alone cannot qualify the complete journey. Only test prompts and test evidence code change. Missing or contradictory native evidence fails closed. ## Model Used OpenAI Codex, GPT-6 family. The exact serving model ID and context window are not exposed by this session. Used reasoning, source editing, shell execution and test inspection. ## 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
|
59b76777db |
test: clarify onboarding artifact completion instructions (#18258)
## Thinking Path > - Paperclip helps people manage AI agents and inspect their work. > - Product E2E journeys test that work through normal browser and API controls. > - The Cloud onboarding journey asks one native child to publish a file and save a document. > - Its task request did not explain the completion report format or make each output step distinct. > - A live attempt published the file, then submitted an invalid finish report and declared a blocker. > - This change orders the scripted request and explains the existing completion contract. > - The agent must still produce the result and pass the same independent checks. ## Linked Issues or Issue Description Refs #17940. Refs #18228. **What happened?** The scripted onboarding request left completion details implicit. A live child published its proof file, then submitted a finish payload without verification and with an invalid schema value. Its later blocked report also listed local paths as artifacts. This does not establish a Core validator defect. **Expected behavior** The request should explain the existing tool contract. The child must publish the file, save the document with its returned link, then submit a valid completion report. Argument correction must not replace artifact verification or bypass approval. **Steps to reproduce** Run the explicit Cloud onboarding Product E2E journey through its private target provider. Approve the proposed child through the UI. Inspect the native finish report and the independently checked attachment, document, and parent links. **Paperclip version or commit** The observed deployment used public Core `6a674716447e8fa268c5cdeb1db1020319368a77`. This fixture change starts from `a8e07500db68f187b19de83823eb7fabf1196021`. **Deployment mode** Cloud testing with the scripted Product E2E harness. ## What Changed - Order the child request: create files, publish the proof, save its document, then finish. - Require verification for actual checks and omit the optional schema field. - Explain receipt-derived references in evidence and file artifacts. Keep local input paths in verification details. - Document same-attempt argument correction and preserve historical failed results. ## Verification - Existing onboarding-first-task and artifact-evidence suites: 83 checks passed. - The maintained harness unit wrapper also passed 137 Node contract checks. - Independent source review found no contract, privacy, or oracle change. - Full `pnpm build` and `pnpm -r typecheck` passed at the preceding commit `c3900cf2`. The only later change names the filename required by the existing document assertion. - Current `d75180b4`: the 83 fixture/artifact checks, 137 contract checks, and `pnpm test:e2e:runner:typecheck` passed again. - Credential-free `--list --suite product-journeys` passed. - The local full repository Vitest run at `c3900cf2` was stopped before the filename correction. It is incomplete and is not reported as a pass. Current-head CI passed: 54 successful checks and two optional Storybook skips. The separate advisory review is still running and is not reported as passed. Initial fresh-worktree dependency and sandbox IPC failures remain recorded. - No paid model, Cloud workspace, release, or deployment was run for this change. These checks establish harness compatibility, not a live model pass. The prior failed campaign remains failed. ## Risks The model may still submit an invalid report or fail another step. A fresh live journey is required to measure the instruction change. This PR changes fixture text and documentation only. It supplies no result, checksum, real attachment ID, or successful status. Models, limits, validators, and independent assertions stay the same. ## Model Used OpenAI Codex, `gpt-6.1-sol`, ultra reasoning. The model identity comes from confirmed parent runtime metadata. Context window size is unavailable. Tools and code execution assisted the change and review. ## 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
|
2badf8a4a5 |
fix(evals): verify native QA replies on the reloaded task thread (#18248)
## Thinking Path > - Paperclip manages AI agents that do work for users. > - Product E2E journeys check user outcomes on real deployments. > - Native task threads can render saved final replies from a settled run. > - The new QA Plan and Stop checks looked only for agent comments. > - Both flows reached the requested terminal state and showed the correct reply, but that storage assumption rejected them. > - This PR checks the reloaded visible reply and binds it to the final successful run. Explicit presentation decisions remain authoritative. ## Linked Issues or Issue Description Refs #18222 and #18244. **What happened?** The deployed Plan journey completed all three native runs, including revision and approval. Stop completed native cancellation, resource release and a new direction. Their final-answer checks failed because the comments endpoint did not contain the rendered native reply. The original failed grades remain unchanged. **Expected behavior** Require exactly one correct visible final answer after reload, attributed to the final successful owned run. Reject an earlier run, progress text, hidden content, duplicate answers or a cancelled run. If its selected completion comment is projected, require that comment and matching author/run/text. Otherwise require a single durable final assistant message and its settled native final-response anchor. **Steps to reproduce** Run the Plan or Stop journey on a native deployment whose task thread renders the settled run before a completion comment is projected. ## What Changed - Add a narrow reply observer over the existing task-thread hooks. It reloads the page and refreshes saved run/comment metadata while waiting. - Bind native replies to a validated durable final assistant event and the successful run's settled final-response anchor. Require the selected comment when projected; support pending or suppressed projection without accepting a summary fallback. Accept payload and item-normalized finals, rejecting contradictory fields. - Use the observer for Plan completion and Stop's new direction. Keep existing run counts, deployment binding, cancellation and artifact checks. - Add negative grading tests, include the helper in the definition hash, and document its evidence contract. Classic protocol comment matchers are unchanged. ## Verification - 63 focused tests passed across native reply evidence, agent evidence and journey contracts. - Product E2E TypeScript passed. The reply observer uses the remaining journey deadline and backs off polling. - Live Plan and Stop attempts exposed this grading failure. Their original outcomes remain failed; fresh qualification is required after merge. - Full current-head CI passed at `40840ed742977853630d733292461ad7003fc7fb`; Greptile and Paperclip Review both scored 5/5, with zero unresolved review threads. ## Risks The observer depends on existing task-thread test hooks and run anchors. Missing or changed hooks fail closed. It reads rendered user-facing text; a finish summary alone cannot satisfy it. Native event provenance, run ownership and exact reply equality are required. No backend, model, provider, budget or cancellation policy changes are included. ## Model Used OpenAI Codex, GPT-6 family. The exact serving model ID and context window are not exposed by this session. Used reasoning, source editing, shell execution and test inspection. ## 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
|
8941db9ce6 |
fix(evals): accept editor escapes and cancel sensitive QA requests (#18244)
## Thinking Path > - Paperclip manages AI agents that do work for users. > - Product E2E journeys check real browser actions and saved results. > - A rich editor can store plain underscores as escaped Markdown. > - The QA reply check treated that valid saved message as a failure. > - A private target adapter also needs to cancel credential-safe HTTP requests before cleanup. > - This PR fixes reply comparison and adds abort and timeout options to the existing sensitive transport. ## Linked Issues or Issue Description Refs #18222. **What happened?** The artifact revision and Stop journeys rejected the correct single user reply because its underscore was stored as `\_`. Sensitive connection setup also lacked a way to pass its lifecycle signal to the actual HTTP request. **Expected behavior** The check accepts the editor's underscore escape and still rejects changed, duplicate, missing, or foreign replies. A cancelled sensitive request closes before setup cleanup proceeds. **Steps to reproduce** Run an artifact revision with an underscore in the reply. For the transport case, abort a pending sensitive request or let its deadline expire. ## What Changed - Compare the fixture reply after undoing only underscore escaping. Keep exact text, count, and author checks. - Let `ConnectionApi.postSensitive` accept a bounded timeout and abort signal. Check cancellation again after cookie loading and before dispatch. - Test the retained reply failure and both cancellation paths with an actual loopback HTTP server. Document the adapter contract. ## Verification - Focused Product E2E tests: 30 passed across two files. - Product E2E TypeScript check passed. After rebase, a final check first found four diagnostics from stale runner/plugin SDK declarations; rebuilding those two workspace packages cleared all four, and the unchanged E2E check passed. - The first sandbox run failed because loopback `listen` returned `EPERM`. The unchanged HTTP tests passed with loopback permission. The test now reports listener errors directly. - Both independent reviewers found a possible pre-arrival deadline-test hang. The follow-up races arrival against request settlement and adds one second of deadline headroom. All 30 affected tests passed after that fix. - Full current-head CI passed in run 40982914525. Greptile and Paperclip Review both scored 5/5 at `f0526f5fbbf13de5ea43e868eb012069ca3d8907`; all review threads are resolved. Live QA requalification follows the merged public and private fixes; no live pass is claimed here. ## Risks The text normalization applies only to plain fixture replies and only removes underscore escapes. It does not remove arbitrary Markdown. Sensitive calls retain their existing 30-second default and closed error messages. API wrappers must forward the optional signal and await the actual request. ## Model Used OpenAI Codex, GPT-6 family. The serving model ID and context window are not exposed by this session. Used reasoning, code editing, shell execution, and test inspection. ## 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 5.5
|
7d194e901f |
feat(ci): merge lockfile refresh PRs through a dedicated bot (#18238)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - CI owns `pnpm-lock.yaml`. Contributors do not commit it, so a poisoned lockfile cannot come in from a community PR. > - `refresh-lockfile.yml` regenerates the lockfile from the manifests on master and opens a PR. > - That PR waits on required checks and code owner review. A `GITHUB_TOKEN` push does not start the checks, so a maintainer must merge each PR by hand. > - The PR content comes only from manifests that maintainers already reviewed on master. A human review adds little. > - This pull request lets a dedicated GitHub App merge the refresh PR at once, without CI or review. > - A release-age gate holds the merge when the refresh adds a package version that is less than 3 days old on npm. > - The benefit is a lockfile loop that closes itself, without a bypass for any identity that other workflows can use. ## Linked Issues or Issue Description Refs #12115 (the same stuck-PR problem, fixed there by starting the checks). Refs #18219 (an example refresh PR that needed a manual approval and merge). **What happened?** Each lockfile refresh PR stays open until a code owner approves it and a maintainer merges it. The required checks `ci / verify` and `ci / e2e` also do not start, because the workflow pushes with `GITHUB_TOKEN`. **Expected behavior** The refresh PR merges with no human action and no CI run, unless it adds a package version that is too new to trust without review. **Steps to reproduce** Merge a manifest change to master. Look at the `chore/refresh-lockfile` PR that the workflow opens. It stays blocked. ## What Changed - `refresh-lockfile.yml` runs in a new `lockfile-refresh` environment. Only `master` can use this environment. - The workflow creates a token for a dedicated lockfile bot app after dependency resolution is complete. The token has only contents and pull request write access. - The workflow merges with `gh pr merge --squash --admin --match-head-commit <sha>`, not `--auto`. It merges only the commit that it pushed. - New `.github/scripts/check-lockfile-release-age.mjs` compares the base and refreshed lockfiles. It looks up the npm publish time of each added package version. The bot merges only when every version is at least 3 days old. A lookup error, a missing publish time, a non-registry version, or a resolution that is not a plain registry tarball also holds the merge. - When the gate holds the PR, the workflow posts or updates one report comment on the PR and does not merge. - A daily schedule runs the refresh again, so a held PR merges after its versions are old enough. - A held run uploads its lockfile as a `held-pnpm-lock` artifact. The artifact records the hash of the master lockfile it started from. The next master run resolves from it only while master still has that lockfile, so held versions stay while their manifests do not change, and a maintainer's lockfile correction replaces them. The workflow does not read the PR branch for this, because users with write access can push to it. - The generated PR body now describes the gate. - `tests/runner-e2e/OPERATIONS.md` now says that the bot can merge the report-lock correction without review, under the same gate. - New `lockfile-refresh-merge.test.mjs` pins the merge safeguards: the environment, the step order, the gate conditions, and `--match-head-commit`. - `doc/RELEASE-AUTOMATION-SETUP.md` §8.1 tells an admin how to create the app, the environment, and the ruleset bypass. It also tells why we do not use `github-actions` or `commitperclip` for the bypass. ## Verification - `node --test .github/scripts/tests/check-lockfile-release-age.test.mjs .github/scripts/tests/lockfile-refresh-*.test.mjs` passes (17/17). - I ran the gate against the real lockfile diff of #18219 and the live npm registry. With 3 days, all 7 added versions passed (`eligible=true`). With 60 days, it held all 7 `turbo` 2.11.10 versions (9.1 days old) and wrote the report. - The workflow YAML parses. actionlint did not run, because it is not installed locally and Docker was not available. - No live run yet. After the admin setup in §8.1, run the workflow with `workflow_dispatch` on master, or merge a manifest change. Then make sure that the bot opens and merges the refresh PR. ## Risks - **Setup order.** An org admin must do the steps in §8.1 before this PR merges. If not, the next refresh fails at the token step, and the PR stays open for a manual merge, as it does now. - **No CI on lockfile changes.** A bad resolution can land on master without a check. The next manifest PR or the canary release shows the problem. - **Release-age gate limits.** The gate stops only fast-detected attacks on new versions. A malicious version that stays undetected for more than 3 days still merges without review. Lockfile changes that add very new versions now wait for a maintainer or up to 3 days. - **Bypass scope.** The app bypasses the master ruleset only in pull request mode. It cannot push to master directly. Only the master-only `lockfile-refresh` environment holds its key. If the key leaks, the holder can merge any open PR past review. Rotate the key if you think it leaked. - **More master runs.** A merge by an app token starts push workflows on master. A merge by `GITHUB_TOKEN` does not. After each lockfile merge, the canary release runs, and the refresh workflow runs again with no change. ## Model Used - Claude Opus 5.5 (`claude-opus-5-5`), Anthropic, in Claude Code with tool use (shell, file edit, GitHub CLI). ## 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 5.5 <noreply@anthropic.com> |
||
|
DottaandPaperclip
|
ebed7f5237 |
test(evals): protect Antigravity setup and two-turn delivery (#18190)
## Thinking Path > - Paperclip manages agents and their work. > - A provider integration must work through ordinary agent setup and task controls. > - The recorded Antigravity walkthrough exposed setup, delivery, and continuation failures. > - A repeatable check must verify real work and retain failed attempts. > - This pull request converts all 47 recorded steps into one fixed Product E2E journey. > - Maintainers can rerun the case and inspect independent delivery, cleanup, and accounting evidence. ## Linked Issues or Issue Description Builds on merged #18187, #18188, and #18189, and consumes the shared repairs from merged #18230 and the equivalent manual-policy declaration from #18222. Refs #18214 for the CI task wrapper. Refs #12503 and #14938 for related legacy adapter work. **Agent or provider** Google Antigravity native Runner with a saved managed Google subscription and the exact Low model. **Why this adapter is useful** A user must be able to add an agent, connect its account, and receive real task results across a follow-up. Picker visibility alone is insufficient. **How the agent is invoked** A fixed Playwright driver uses ordinary product controls. The agent runs through native ACP. Independent checks use public APIs, actual downloads, owned process observations, and native lifetime fences. ## What Changed - Register one explicit-only case: `antigravity-agent-journey.runner-antigravity.local.saved-subscription-two-turn`. - Map all 47 source steps and declare saved-account prerequisites, finite limits, and same-host macOS ARM64 scope. - Create an agent through the ordinary picker, connect the saved account, select Low, run Test, and wait for readiness. Save an instruction marker, deliver a file, and update it through an ordinary follow-up. - Require each run's visible completed file activity, exact downloaded bytes, committed Done result, same tagged conversation, distinct physical owner birth, released lease, and free native fences. Verify reload persistence. - Stop failed native runs that still own a retry. Replay an existing Stop with its durable request identity and wait for the matching audited acknowledgement. A terminal reservation without a native intent remains unconfirmed. - Keep private evidence in the existing result, report, dashboard, and accounting pipeline. Retain the paused QA agent and financial history while preserving the shared account and budget policies. - Journal only the ordinary wizard's future ID, target, name, intended harness and expected creator before forwarding its POST unchanged. Bind the response and interrupted recovery to exact identity reads; a pending 404 stays unconfirmed. - Serialize private resource snapshots through temporary-file fsync, atomic rename and directory fsync. Abort forwarding on storage failure and preserve legacy receipts. If later journal or diagnostic writes fail, retain the primary error, always close the browser, and attempt the ordinary failed report. - Add production-component browser controls and harden task-menu readiness, alias selection, scroll restoration, failure packaging, and one hermetic auth-test budget. - Wait for the existing GitHub recovery fixture to register its next deadline before asserting the original range; preserve its nested-transaction and zero-request checks. In the decision-expiry fixture, verify the exact saved expiry separately from the earlier global wake of another queue. ## Verification - Fresh registered acceptance passed on eval source `9060f3cd01b7c3445ecdf703d72e9cf30fe42b59` and merged serving product `61bd4ac6919c2309f766a25e62273b074cee3261`. It completed ordinary setup and two turns with native `.5` and `gemini-3.8-flash-low`. All 16 matchers, 14 checkpoints, and cleanup passed in **165,613 ms**. Actual downloads were **61 bytes** and **119 bytes**. Both turns passed their own run-bound browser activity checks and reload persistence. - Independent read-back confirms Done, a paused retained agent, the same connected saved default account, and no active owned run. Both accepted results have the same full tagged conversation and distinct primary process births. All 21 observed processes retired; both leases released and all three native fences were free after each turn. The selected `.5` checkpoints also showed equal agent/backend session IDs; this is an observed case requirement, not a generic ACP invariant. - Receipts report 339,273 input tokens, 7,224 output tokens, and 202,846 cached input tokens. Subscription prices are unavailable. Billing coverage remains incomplete, not free. - The normal report CLI passed. All 310 files from nine original attempts and two separate recoveries retained their hashes. The twelve current screenshots are byte-identical, and 194 local report links are contained and valid. The eight earlier reports were copied without regeneration. - Delivery source `1bf4fb04df04710b82844b3c9355767f46b5f68f` branches from `d3f2a60e795fc60d64c8b19d1eed4d48657d5492` and merges cleanly with current master `1d571de5f9f886d66e520800baec8fc2f67d0948`. All 41 prior `c8b332` authored file blobs remain identical; the 42nd path is the decision-expiry fixture contract correction. Creation ownership, the strict durable writer, and the ordinary successful workflow remain unchanged. Of 43 measured file blobs, 30 remain identical. The Antigravity oracle, artifact, billing and report gates and its suite-specific validation block remain unchanged. The actual retained agent matches its original creator; its original request `setupId` was not captured and is not reconstructed. - Later master changes include model-visible workspace guidance and rejected-terminal-report handling. Their bounded source assessment is separate: the first actual finish failed, then a same-turn replacement was accepted and committed. Later read-only results match both original native-result hashes, and the complete event catalogs have no `session.failed` event. Master #18228 also expands detection of explicit JSON file requests. The measured create/update-and-attach `.txt` task already required publication under both parsers; the publication validation gate itself is unchanged. Master #18240 changes remote workspace cache staging and adds an OpenCode index refresh. The selected local lane returns before those paths. Master #18213 adds optional numeric turn timing and trace boundaries. Its selected Antigravity model, text, IDs, result and retirement semantics remain unchanged in the bounded source review. Local diagnostic rows and awaits can change event ordinals and latency; the oracle checks relative semantic order. Collector coverage and latency improvement remain outside this proof. This does not claim that the whole current runtime is byte-identical to serving `61bd`, or relabel the measured result as a new run. - The adopted upstream cache controls passed 3/3 with pinned real Turbo. Forced work disables local and remote reads and writes; ordinary work reuses local results. The original genuine-pnpm/Turbo controls remain historical to `9060`'s earlier write-only policy. The duplicate cache patch is superseded by #18230. Its qualification-only pnpm addition is removed because the adopted upstream controls do not require it. - The E2E typecheck, map checks and 277-control replay are carried from the unchanged eval source paths. That replay passed, including the 183 Antigravity controls, catalog/connection checks and the separate product-journey validator. The carried creation controls include 91 config/recovery cases. The final launcher suite passed 19/19, including five real campaign finalization fault cases; the report suite passed 7/7. These final checks did not rerun the earlier 277-control group. The canonical production-component browser spec passed in 27.4 s under its unchanged 60 s limit before the replay. Four paused-composer cases and the product base's build/typecheck plus 376 focused regressions were verified earlier. - The shipped wizard and actual API client cover persistence before forwarding, lost response JSON, foreign/malformed requests, wrong response identity/creator and cleanup-error preservation. Real filesystem controls failed 5/5 before the writer repair and passed 5/5 afterward, preserving the old journal on partial-write, temporary-file sync and rename faults. Three persistent-sync campaign controls failed before the finalization repair and passed afterward. All five final campaign fault cases preserve the primary failure, cleanup failure, usage and failed report while attempting browser closure. The strict journal still blocks the next mutation. These are provider-free ownership controls; the Google task proof remains the original `9060` → `61bd` acceptance. - The decision restart check passed in the local 20-case fixture run. A current-source real-PostgreSQL rollback control reproduced the exact 90,000 ms earlier global wake: the old equality rejected it, while the exact durable expiry and new future-wake bound passed. The original 60 s open-state, exact-timestamp, continuation and quiet checks remain. The broader local run was **FAILED: 19 passed / 1 failed**, with a 15 s timeout and a 30 s cleanup timeout in an unchanged timestamp-pagination case. That case passed in canonical `c8b332` CI; the local timeout cause remains inconclusive. An extra local server typecheck failed against stale cached Runner declarations; it did not check this excluded test file. Both limits remain recorded for fresh canonical CI. - A broader local GitHub recovery group passed 21/22. Its unchanged pagination case observed an unexpected config request; the log does not establish the cause. That result remains retained, and fresh current-head CI is the final broader gate. - Exact delivery-head CI completed with 55 successes, two skips, and one optional neutral review. Required `ci / verify` and `ci / e2e`, build, typecheck, and Runner checks passed. Fresh Greptile is 5/5 with zero unresolved threads. Optional Paperclip is neutral 4/5; its carried recommendations are dispositioned against the public Stop contract and selected live identity evidence. Seven earlier converted attempts retain their failed grades and partial evidence. Historical Paid8 and its later read-only per-run UI observation remain separate from the fresh acceptance above. Unknown prices and the older partial startup accounting remain explicit. Full lineage and raw evidence remain private in the normal report bundle. ## Risks - This case requires an authorized saved personal default Google subscription and finite operator candidate admission. It does not test fresh OAuth or establish global qualification. - The physical observer supports a same-host loopback macOS ARM64 target. Cloud, Linux, and other models remain outside this case. - A known-cost stop cannot cap missing prices or in-flight subscription charges. Existing operator policies remain unchanged. - Raw account/native identities stay private. Missing retirement proof keeps resources retained. The selected cleanup mode does not test agent purge. ## Model Used OpenAI Codex (GPT-6). The host did not expose the exact model revision or context limit. The model used tools for code edits, tests, browser automation, and review. ## 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
|
d3f2a60e79 |
test(e2e): reuse core journeys across deployment targets (#18222)
## Thinking Path > - Paperclip manages agents and their work. > - Product E2E tests verify work through the browser and public APIs. > - The classic matrix already tests several common task stories. > - Its older stories depend on local fixture setup and workspace access. > - Deployment tests need the same story checks with a separate target adapter. > - This PR shares those checks and adds five explicit journey selections. > - A deployment pass still requires evidence from that deployment. ## Linked Issues or Issue Description **What existing behavior does this improve?** Product E2E journey reuse across managed instances and deployed Cloud instances. **Current behavior** The deployed-target catalog covers task creation, assignment, Ask, and fresh onboarding. Plan revision and Everyday stories run through the classic fixture driver. Several story steps assume a local workspace. **Proposed behavior** Add explicit Plan revision, artifact revision, protected-action approval and decline, Skill Studio, and Stop/redirect journeys. Share the main browser sequences and content checks with the classic driver. Let target adapters provide authentication, native agents, a reachable service fixture, and owned cleanup. **Reason and benefit** A local pass cannot verify the deployed application. Shared story code lets a deployed target exercise the same behavior without copying an entire test or starting a replacement server. **Breaking changes** No product API or default matrix selection changes. Passed deployment results must satisfy the declared story checks and work-run count. Unsupported target capabilities fail before a story runs. Related public work: #17262 and #18098. Upstream prerequisite: #18230. ## What Changed - Extract shared Plan, artifact download/oracle, protected service, Skill Studio, and Stop/redirect helpers. Keep classic callers on those helpers. - Add five explicit paid journey selections with owned-task observations and native company, agent, model, and environment checks. - Include classic prompts, delivery/decision code, and the decline grader in the source-definition hash. Keep classic transient-read retries and reject deployment binding failures immediately. Match cookie directory boundaries in sensitive fixture requests. - Require complete story checks and expected run receipts before retaining a passing result. Keep failed evidence and private downloads. - Add a secret-safe request method for deployment fixture credentials. Add the missing declaration for the existing manual evidence policy module so the E2E project can typecheck. - Document target responsibilities and promotion after a real deployment pass. This source catalog does not enroll candidates in a daily plan. ## Verification - E2E TypeScript check: passed, including the review follow-up. - Review follow-up: 30 tests passed across four files, covering transient Plan reads, immediate binding rejection, cookie-path scope, agent receipts, and artifact evidence. Stop now waits for a saved, still-active native turn before the browser click and retains its startup events. Counterexamples reject claimed-but-unstarted, already-completed, interrupted, and non-native work. Shared observation and API helpers also enter the definition fingerprint. Cancellation snapshots are captured before lease cleanup. A direct terminal-boundary assertion rejects native work after cancellation even if it is already in the first snapshot; five cleanup-window counterexamples are covered. The Stop-output regression first failed for both native `item.delta` and `item.completed`; the fixed check detects both while allowing accounting-only updates. A report-catalog test first failed because the sandbox denied the tsx child IPC pipe (`listen EPERM`); the unchanged test passed outside the sandbox. - Affected journey, catalog, API, service fixture, and delivery checks: 313 tests passed across 19 files. - Workspace build: passed after rebase. - Workspace typecheck: passed. The default repository run was attempted. Repeated embedded PostgreSQL startup failures prevented local database coverage; it was stopped without claiming a full pass. Linux CI must supply that coverage. - Broad E2E support run: 4,793 passed, 23 failed, 2 skipped. Failures include existing stale image/Pi assertions and process/startup timeouts under host load. The affected tests pass separately. This is not a claim that the broad suite is green. - Upstream CI prerequisites: this branch is rebased on the merged #18230 repair. The earlier conflicting Turbo flags no longer apply. The shell-hook fix is also supplied by that upstream change. Linux shell-hook investigation: 19 focused tests and adapter TypeScript passed. The exact staged-config sequence reproduced CI's 500 in a pinned Node 24 Linux container. Node's default stdin socket at shell level zero selects Bash remote-shell startup and reads `.bashrc` instead of `BASH_ENV`. Ignored stdin and a nested shell each produce 1000. The fixture now ignores stdin; the production hook and 1000 assertion are unchanged. See [Bash startup rules](https://www.gnu.org/software/bash/manual/html_node/Bash-Startup-Files). - GitHub CI on the earlier head passed most suites but failed two browser tests (a fixture cleanup lock timeout and a fixture run staying queued) plus the shell-hook assertion. The lock-timeout case received one retry, which was cancelled by a later source push and does not establish recovery. These failures remain part of the verification record. - Real Cloud qualification: the private target adapter is reviewed and merged. These candidates must pass there before daily-plan promotion. Current source review: Greptile 5/5 and Paperclip Review 5/5 on `a9df1d6e8d880710fce39b87e689e07e1c8e6edc`, with zero unresolved review threads. The full [GitHub CI run](https://github.com/paperclipai/paperclip/actions/runs/40974946081) passed on that exact head, including both required aggregate checks. ## Risks Shared helper changes affect classic callers. The independent ZIP oracle and visible delivery requirements remain in place. Stop requires native acknowledgement, terminal event evidence, resource cleanup, and no later original-turn activity. The managed local target still lacks paid-agent fixtures and reports these journeys as unsupported. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, terminal use, and code execution. The exact serving model ID and context window are not exposed to 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> |
||
|
github-actions[bot]andlockfile-bot
|
1843d27e09 |
chore(lockfile): refresh pnpm-lock.yaml (#18219)
## Thinking Path > - Paperclip uses a dependency lock for reproducible builds. > - Dependency changes on master need matching lock entries. > - Paid eval reports also pin a reviewed offline correction. > - Reports must install before paid model work starts. > - This automated PR keeps the lock and report proof together. ## Linked Issues or Issue Description **What happened?** Dependency manifests on master changed after the last committed lock refresh. **Expected behavior** Frozen installs must match those manifests. Eval reports must retain matching offline lock proof. **Steps to reproduce** Run a frozen install before and after this lock refresh. ## What Changed - Refresh the dependency lock from the manifests on master. - Include the matching offline report correction and hashes. ## Verification - Pinned pnpm resolution completed with lifecycle scripts disabled. - The report proof refresh accepted the reviewed committed baseline. - Required PR checks and reviews must pass before merge. ## Risks Dependency changes require review. Unknown report lock baselines remain blocked. This update changes no permissions or paid enable gates. ## Model Used None — generated by pnpm and the deterministic report-lock updater. ## Checklist - [x] Thinking path and problem description included - [x] Generated lock and matching report proof included - [ ] Required CI checks and reviews pass Co-authored-by: lockfile-bot <lockfile-bot@users.noreply.github.com> |
||
|
DottaandPaperclip
|
99d27fd5e7 |
feat(evals): add a metadata overview for the chat catalog (#18194)
## Thinking Path > - Paperclip is a control plane for AI agents that do work. > - Its evaluation reports separate recorded results from source coverage. > - A case without a main baseline measurement can still have a candidate run. > - A normal campaign card treats an expected execution without a result as missing evidence. > - Catalog planning data must not create those cards or change a recorded grade. > - This change adds a typed metadata view to the existing Evals dashboard. > - Readers can compare main outcomes, candidate execution, allowances and source review separately. ## Linked Issues or Issue Description **What existing behavior does this improve?** The existing Evals dashboard can show a campaign's retained results. It cannot show the full chat catalog as metadata without synthetic result entries. **Current behavior** An unused initial, an unknown allowance and a completed candidate run can all lack a main baseline result. Showing them as missing campaign artifacts hides these differences. **Proposed behavior** Add one validated metadata view in the existing renderer. Show one row per frozen scenario, separate execution and main outcome columns, filters, and the public request and expected outcome from static YAML. Keep normal campaign cards unchanged. **Reason and benefit** Readers can see which cases ran, which have no main measurement, and which still need preparation. Source review and candidate outcomes cannot add baseline pass credit. **Breaking changes** None to existing campaign inputs or outputs. The new metadata mode uses its own strict schema and requires a new private output directory. No matching open or closed PR or issue was found in the duplicate search. ## What Changed - Add a strict display-only catalog schema with closed enums and fixed catalog scope. - Validate the 244 row identities against static YAML without importing executable case callbacks. - Show main results, recorded execution, allowances and source review in separate columns. - Add search and filters, passing examples, qualification notes and expandable public test descriptions. - Add one dated immutable snapshot link in the existing shared report navigation. This does not change history data or update an already-published page. - Add a private metadata preview command to the existing dashboard owner. Keep existing campaign generation on its current path. - Keep allowance certainty independent of measurements and allow an optional later candidate outcome alongside a main outcome, without changing main counters. - Add focused pure, JSDOM and file-loader checks. Document the metadata command. ## Verification - The owning `catalog-overview.test.ts` Vitest suite passes all fourteen tests under the focused source configuration. The valid-main/unknown-allowance and main-fail/later-candidate-pass controls fail against the old renderer and pass with this repair. The never-run/unknown-allowance control fails against |
||
|
github-actions[bot]andlockfile-bot
|
4a6f6a65eb |
chore(lockfile): refresh pnpm-lock.yaml (#18185)
## Thinking Path > - Paperclip uses a dependency lock for reproducible builds. > - Dependency changes on master need matching lock entries. > - Paid eval reports also pin a reviewed offline correction. > - Reports must install before paid model work starts. > - This automated PR keeps the lock and report proof together. ## Linked Issues or Issue Description **What happened?** Dependency manifests on master changed after the last committed lock refresh. **Expected behavior** Frozen installs must match those manifests. Eval reports must retain matching offline lock proof. **Steps to reproduce** Run a frozen install before and after this lock refresh. ## What Changed - Refresh the dependency lock from the manifests on master. - Include the matching offline report correction and hashes. ## Verification - Pinned pnpm resolution completed with lifecycle scripts disabled. - The report proof refresh accepted the reviewed committed baseline. - Required PR checks and reviews must pass before merge. ## Risks Dependency changes require review. Unknown report lock baselines remain blocked. This update changes no permissions or paid enable gates. ## Model Used None — generated by pnpm and the deterministic report-lock updater. ## Checklist - [x] Thinking path and problem description included - [x] Generated lock and matching report proof included - [ ] Required CI checks and reviews pass Co-authored-by: lockfile-bot <lockfile-bot@users.noreply.github.com> |
||
|
DottaandPaperclip
|
cd0c86e5e1 |
fix(evals): support private evidence for the frozen chat catalog (#18170)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Product E2E tests need to retain evidence from both successful and failed work. > - Manual private runs use the existing protected AWS evidence route. > - Its fixed eight-case registry blocks other scenarios in the frozen chat catalog. > - Failed Playwright cells also produce static trace files that need exact owner and content checks. > - This pull request admits the fixed catalog, keeps eight cells per run, and retains the existing report identities. > - Exact source and run approval still controls each private operation. ## Linked Issues or Issue Description Refs: #18134, #17888. **What happened?** The manual private route accepts only its first eight execution IDs. It rejects other IDs from the frozen 244-scenario chat catalog before a run can start. The dotted `area.case` report identity also fails the retained report parser. A failed Playwright cell can produce canonical trace assets that the private evidence owner does not yet recognize. **Expected behavior** The trusted route must recognize exactly the frozen 244 OpenCode/local execution IDs. Each original run must still select at most eight cells and have exact protected approval. The evidence owners must retain complete scanned originals from the actual pinned report producer. Unknown carriers and credential contamination must still fail closed. **Steps to reproduce** Use a known frozen chat ID outside the first eight with the current manual classifier. The classifier rejects it. Parse a dotted chat execution ID with the current report parser. The parser rejects it. Generate a synthetic failed Node-only Playwright test and pass its actual trace/report tree through the private evidence owners. This change includes the narrow parser repair from the existing open #17888. It does not merge or close that PR. The related source work was checked for overlap before this branch was created. ## What Changed - Replace the small private registry with the exact frozen 244 execution IDs. Preserve the two existing undotted standalone task IDs. - Keep the private run limit at eight through classification, approval rows, cell binding and transfer sessions. Preserve nonprivate selector behavior. - Bind private cell report assets to the actual installed pinned producer, selected cell, canonical path and exact bytes before the existing scans and archive checks. Pass that verified owner into both real launcher packaging calls. - Exclude installed build staging from the private evidence walk. Retain the real standard-file cap and reject unknown nested carriers. - Add the existing dotted chat report grammar. Preserve other report families and legacy task IDs. - Use the existing public announcement dismissal control in one external-link/upload E2E test so a real announcement cannot cover its composer button. Preserve the open menu and existing timeout. - Document source compatibility and operator gates. No workflow, product, dependency, lockfile, case callback, query, clock or budget changes. ## Verification - 132 distinct focused controls are closed. The current affected owner suite passes 39 controls under a nondefault temporary directory. The other 93 unchanged controls retain their previous passes. The initial registry, map-order and test-tooling failures remain recorded. - Two actual launcher call-expression controls fail on the prior source because canonical trace files are dropped. Both pass after the private-owner wiring. These controls do not claim full launcher lifecycle or actual E2E qualification. - Temporary paths use the real operating-system temporary directory, including the handoff check. - The controls check all 244 IDs through real private routing, cell binding and original-run approval. They reject unknown IDs, aliases, wrong profiles, traversal, private duplicates and excess private cells. - The controls preserve actual nonprivate 33-ID and duplicate selector behavior. - A real synthetic Node-only failed Playwright test produces the report and automatic trace. The control uses actual package, seal, archive, receipt and restored validation paths with mock AWS transport only. - Owning TypeScript `--noEmit` check passes. No local product server, browser, model, database, SDK build or evaluation was run. - Prior-head CI run 40906120388 failed one E2E test because the real announcement footer intercepted its Add to composer click. The other 28 tests in that shard passed. The narrow fixture uses the existing UI dismissal pattern. The added handler passes an isolated noEmit check against the real Playwright Page type. An additional strict check of the whole browser test fails on five unchanged shared-fixture type errors; checking that unchanged shared file alone gives identical diagnostics. The initial shell exit was misreported and corrected in the validation record. Browser verification is deferred to the next normal CI run; no local browser or server was started. - Repository-wide local build and test commands were not run because this is a scoped source change. Normal required PR checks will provide the repository CI result. ## Risks A registry entry is not proof that a target or feature is ready. A future target must use the exact matching private helper and producer contract, retain any stronger target scanners, and produce fresh source fingerprints. Existing running jobs keep their pinned workflow and target sources. All original attempt limits, declared inputs, private storage checks, fatal scanners and exact administrator-controlled run approval remain in force. The new cell report contract accepts only canonical assets from the actual pinned Playwright installation. Other assets and layouts still fail closed. No private evidence or generated report is included in this PR. ## Model Used OpenAI GPT-6 through Codex, with tool use and code execution. The runtime does not expose a finer 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
|
cc6a3296e5 |
fix: keep agent YAML compact and preserve omitted settings (#18168)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The Advanced tab lets operators edit agent settings as YAML. > - Its first format embedded company catalogs, profile definitions, history, and instruction text. > - This made one agent document grow with unrelated company data. > - Removing those fields from the display also needs safe save behavior. > - This pull request renders a compact document and resolves sparse edits against saved state. > - Operators can change settings without copying shared data or deleting hidden values. ## Linked Issues or Issue Description Refs #2219. Follow-up to #18036. **Current behavior** The Advanced document expands shared profiles, embeds instruction files, and lists read-only state and unused connections. It also emits empty command lists and unset fields. **Proposed behavior** Show the agent's own settings and profile/instruction references. Omitted settings stay unchanged. Explicit empty values clear supported settings. A named instruction file changes only when supplied; `null` deletes that file. **Reason and benefit** An agent's YAML should remain readable when its company adds profiles or its instruction files grow. Every setting remains available through YAML or its owning settings tab. ## What Changed - Remove the read-only `observed` export and unused connection inventory. - Render profile-key references and instruction locations. Keep file contents and legacy prompt text out of the default editor. - Omit unset nullable fields, empty maps/lists, zero budget, standard permissions, and empty commands. - Add a sparse input schema and saved-state resolution. Preserve omitted fields, binding definitions, file contents, permissions, and channel settings. - Keep explicit clears and map/list replacement. Patch named instruction files and use `null` for deletion. Preserve hidden legacy prompt text unless explicitly changed. - Check profile references, shared edits, and file restrictions before writes. Keep revision guards, native authorization, and partial-save recovery. - Update documentation, verification maps, and regression coverage. Remove operating-system clipboard dependence from browser tests. ## Verification - All 48 jobs pass in the [final CI run](https://github.com/paperclipai/paperclip/actions/runs/40904124001) on `6dbb11d7a6`. Greptile is 5/5 on this exact commit. Both findings are resolved; there are no failed or pending checks. - Paperclip Review ended neutral because its agent finished without a complete assessment. It did not provide a code finding. Superagent Security Scan passed. - The focused suite passes 58 tests across YAML parsing, planning/saving, compact output, and the editor. - A large unit fixture with 30 profiles, 50 entries per profile, and a 1,000-line instruction file renders under 70 lines and is more than 95% smaller than the prior expanded representation. - All eight browser cases pass against real APIs and a dedicated PostgreSQL database. The compact-output case stays under 60 lines with large profiles and instruction text. It checks sparse saves and profile-key binding without changing shared definitions. The resource-save browser case also checks that unpinned skill shorthand does not trigger another sync during an unrelated save. - Browser coverage also checks resource saves/clears, secret redaction, key receipts through rename, stale drafts, lost responses, instruction relocation, and revision restore. The MCP fixture is a local target; no model run or external provider login is claimed. - Workspace build and typecheck pass. Map, Storybook catalog, and token gates pass. These checks and all eight browser cases passed after rebase onto current master, including its new Secrets regression. The merge conflict consisted of adjacent test additions; both were retained. The final skill no-op correction also passes a focused browser rerun. The mixed-resource channel regression failed before its fix and now passes, including revision restore with unpinned skills. UI typecheck passes after both corrections. - The first browser run hit a five-second test wait during an active multi-resource save. A focused rerun exposed shared clipboard interference. The helper now waits for save completion and uses scoped copy events plus keyboard input. The next full run passed. Earlier logs and screenshots remain outside the repository. Run the focused browser suite with: ```sh pnpm exec playwright test --config tests/e2e/playwright.config.ts tests/e2e/agent-configuration.spec.ts ``` ## Risks - Omitted fields now preserve saved values. Removing a line does not clear that setting. Supply its supported empty value instead. - Instruction file maps are patches. Old exports that removed a file by dropping its entry must now set that file to `null`. Entry files and read-only assets remain protected. - Explicit lists and adapter/runtime/metadata maps retain replacement behavior. Hidden legacy prompt text is preserved unless supplied explicitly; an empty string clears it. - Shared definition edits still require explicit consent. Saves still use sequential native APIs and can partially succeed. Existing unversioned resource race windows remain. - No database migration, new dependency, lockfile edit, or workflow change. ## Model Used OpenAI Codex, GPT-6 family. The exact serving model ID and context window are not exposed in this session. The work used reasoning, repository editing, code execution, and Playwright browser automation. ## 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> |
||
|
|
07c6df260b |
fix(server): schedule archived runner evidence from saved deadlines (#17925)
<!-- entire-trail-link-start --> https://entire.com/gh/paperclipai/-/trails/85/changes/paperclip/85 <!-- entire-trail-link-end --> ## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Retired native sessions leave continuity and quarantine evidence on disk. > - The old hourly loop mixes evidence retention with backup expiry, which has different database eligibility rules. > - Evidence already has saved creation timestamps and local publication signals. > - This pull request gives evidence its own registered worker and actual next deadline. > - The worker stops filesystem and SQL activity when its evidence obligations settle. ## Linked Issues or Issue Description Stack slice 65. Parent: #17923 (`codex/sleep-64-skill-version-retention`). Related foundation: #15625. **What happened?** At parent `7b94aa185c29673372646bbf2e07a2fbd16bc394`, an hourly backup/evidence tick still inspects native state roots. Evidence's private due time is checked by that recurring tick instead of the shared scheduler. Empty evidence and future retention windows have no reason to be rediscovered every hour. **Expected behavior** Reconstruct evidence once. Wake from the existing publication signal. Register the earliest saved retention deadline. Keep busy, failed, or incomplete evidence recoverable. Preserve backup eligibility as a separate duty. **Steps to reproduce** Start retention with an empty evidence root or archives inside the seven-day window. The old timer still visits the combined owner. The new evidence worker disarms when empty, or registers the saved expiry with no evidence filesystem or SQL calls between relevant changes and that deadline. ## What Changed - Add an evidence worker through the delivery coordinator and shared scheduler. Track startup reconstruction without delaying HTTP readiness. Unsubscribe local publication before delivery shutdown. - Select `backups` for the existing backup timer and `evidence` for the new worker. Existing callers that omit the selection retain combined behavior. Backup expiration remains the next separate migration. - Use hourly retries for concrete busy or failed evidence, with earlier saved expiries and one-second continuations independent. Restore candidate/removal/byte summaries. Schedule publication at the first removable instant; preserve legacy manual-tick semantics. - Expose the existing earliest evidence deadline. Preserve one-second bounded continuation, concrete failure/busy retries, cleanup-source archive exclusions, symlink protection, late activity checks, atomic rename, and interrupted removal recovery. - Preserve non-ENOENT inspection errors as recovery obligations. Retry a failed session root in a new evidence pass instead of continuing past it and forgetting it. - Register the first removable instant after the inclusive retention boundary. Keep the existing manual dry-run publication semantics. - Add real filesystem/isolated PostgreSQL lifecycle tests and disabled/held-startup wiring cases. Add no table or migration. The archive and quarantine writers already call `noteNativeRunnerEvidence`; this PR connects that existing signal to the scheduler. Shared code owns scheduling, while evidence rules stay with the retention service. See [`doc/plans/2026-10-27-runner-evidence-retention-sleep-audit.md`](doc/plans/2026-10-27-runner-evidence-retention-sleep-audit.md). ## Verification - Adopted actual prerequisite squash `b06ea111a254c62e0a55199b010d82e0adb33bf0` from #17923 through normal merges. Qualified runtime/package/UI/map trees remain unchanged. The final nine-path diff includes one browser-spec readiness repair. - Runtime qualification remains **652 passing TypeScript cases across 22 complete files**, plus **29 passing Rust cases** in native-backend, runtime-notification and configured-environment targets. One actual closed Pi snapshot launch is intentionally skipped because its external distribution root is not provided. Accepted source is `b80eee32f344ae0218d8b791c14f2ea174da09cc`; all runtime and earlier sleep-test bytes are preserved. - The complete outage-recovery browser file passes **all five journeys** on a fresh isolated database after four additive rendered-history readiness assertions. Every original assertion, outage control and timeout remains byte-identical when those new lines are removed. The intentionally unresolved first-session scenario keeps its pre-auth interaction and delayed-owner coverage. The prior CI/local timeout evidence and one rejected broader precondition remain in the audit. - Adapter/Runner/database builds, server/Runner typechecking, Rust formatting/workspace checks, server build, module boundaries, migration order and generated TypeScript/Rust profiles pass. PostgreSQL 17 fixtures apply all 413 journal entries through index 414; source hashes and temporary fixture imports restore exactly. - **Inherited validation exception:** Copilot profile 81's separate source-identity check fails on the merged Rust configured-environment hash. Its inputs and attestations equal upstream and stay unchanged here. Grok's identity check passes. No release attestation is regenerated or new live-provider qualification claimed. Required exact-head CI/security and Greptile 5/5 still must pass. - [The evidence audit](doc/plans/2026-10-27-runner-evidence-retention-sleep-audit.md) retains historical failures, counts and limitations. No full local repository-suite, live provider, external-alarm, measured suspension or fleet-savings claim. Fresh final-head CI/security, exact-head Greptile 5/5 and actionable Ponytail findings remain separate merge gates. ## Risks - Live publication is the existing server-owned executor in one process. Independent filesystem writers need an explicit signal contract. Startup reconstructs restart state. - Busy sessions and concrete inspection/removal failures retain retries. Unknown cleanup-source archives remain protected. - The existing scan enumerates directory names before applying its entry/removal budget. This PR preserves that behavior and does not claim a bounded total directory-listing allocation. - An in-process deadline cannot wake stopped compute. Cloud sleep-safety and external-alarm coverage stay unchanged. Physical evidence reclamation can wait for a supported host wake. - Backup expiration still has an hourly database eligibility scan. This slice removes only the archived-evidence duty from that loop and does not claim full instance suspension. ## Model Used OpenAI Codex (GPT-6-based coding agent), using repository inspection, normal Git merges, composition proofs and complete PostgreSQL/filesystem/startup suites. No subagents were used in this reconciliation. ## 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 --------- Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: Codex <noreply@openai.com> |
||
|
Devin FoleyandPaperclip
|
fd98075a22 |
fix(ui): show environment variables on agent Secrets pages (#18121)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Operators inspect agent configuration to understand which credentials an agent can use. > - Agent environment bindings can contain inline values, company secret references, and user secret references. > - The Secrets page did not mount the environment editor. Inline credentials were absent, and its empty message said no secrets were bound. > - The existing environment editor already handles these binding types and redacted values. > - This PR mounts that editor on the Secrets page in both UI layouts and clarifies the empty reference message. > - Operators can inspect and edit the configured bindings. Values remain masked, and the Runtime editor remains available. ## Linked Issues or Issue Description **What happened?** An agent can receive an inline credential through its environment while its Secrets page shows no matching entry. That page only rendered the secret-reference overview and API access controls. **Expected behavior** The agent Secrets page shows existing environment bindings as well as secret references. Saved inline values stay redacted. Saving another binding preserves existing credentials. **Steps to reproduce** 1. Save an agent with an inline environment credential. 2. Open the agent Secrets page. 3. The credential row is missing, although the runtime configuration contains it. **Paperclip version or commit** Reproduced on master `27270e9bf8b7a294cf03f8b84f4857f1b1ff05c7` before the fix. Related: [#12517](https://github.com/paperclipai/paperclip/pull/12517) proposes changes to secret assignment controls. This PR covers display of existing inline environment bindings through the current editor. ## What Changed - Show the environment editor on agent Secrets pages in both UI layouts. Keep the Runtime editor available. - Change the empty overview message to describe missing company secret references. Personal references remain visible in the environment editor. - Add native and legacy agent regressions through the actual page components. - Add a browser check for masked values, secret references, save/reload, and exact credential preservation through an opaque configuration revision. - Update the feature and agent access maps for the existing UI entry point. ## Verification - Red before the fix: the native and legacy page regressions could not find the saved environment credential row. - The focused UI suites pass: 207 tests, including all four layout/adapter page cases. - The server redaction and configuration revision suites pass: 42 tests. Native Runner environment assembly suites pass: 40 tests. - An isolated Chromium flow passes against the real local API and database. It checks a masked inline value, a company secret reference, and save/reload of an added variable. It then removes the new variable and checks that the opaque server revision matches its original value. This proves that the hidden credentials survive both saves without reading them. The initial browser attempt exposed a fixture ordering assumption; the test now compares variable-name sets because PostgreSQL JSONB does not retain insertion order. - `pnpm -r typecheck`, `pnpm build`, `pnpm check:token-gates`, and `node scripts/check-maps.mjs` pass. - The full local `pnpm test:run` attempt found three failures in unchanged server fixtures. The chat-recovery case passes in isolation and in its 22-test group. The two workspace cases inherit `init.defaultBranch=main`; all 179 workspace tests pass when the child process uses `init.defaultBranch=master`, as the fixtures expect. The long local invocation was then stopped. Later local groups were not reached. All corresponding CI test shards pass. - [CI passes on the final commit](https://github.com/paperclipai/paperclip/actions/runs/40809064907). The first canary dry run found an npm version collision. Its single retry passed without a source change. - Apex passes at 5/5 on the current commit with no unresolved findings. ## Risks Low risk. The change reuses the existing editor and save path. No API, secret delivery, permission, or database contract changes. The tests cover masked values and existing references during save/reload. ## Model Used OpenAI GPT-6 through Codex, with tool use and code execution. The session 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 — scoped checks pass; the full local command limits are recorded 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> |