mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-12 06:10:44 +01:00
42bc238aa5a1208cabecba76773e0500fcd33959
5
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 DottaandPaperclip
|
a92f616cd0 |
fix(db): attribute bounded captures to execution owners (#18634)
## Thinking Path > - Paperclip manages agent work through shared server and database services. > - High query volume can spend main-thread time on SQL construction and result decoding. > - The existing bounded response capture counts rows and bytes, but it cannot name the operation that executed a query. > - Bundled callers can be missing, and measuring every parsed JSON value adds work to the diagnostic. > - This pull request carries a semantic execution owner into each measured query and bounds response sizing. > - Operators can rank query volume by owner before choosing a performance fix. ## Linked Issues or Issue Description Refs #17504, #17522, #17537, #17564. **What existing behavior does this improve?** The opt-in `PAPERCLIP_DB_RESPONSE_CAPTURE_WINDOW` diagnostic. **Subsystem affected** Database diagnostics and existing server execution boundaries. **Current behavior** The capture groups by table, operation, and builder caller. It cannot distinguish execution owners that reuse a prepared query. Bundled deployments need source maps for caller locations. The capture visits every returned value and serializes parsed JSON again to estimate its size. **Proposed behavior** Version 2 captures use fixed semantic owners for HTTP requests, delivery queues, execution phases, and budget work. They count completed calls, returned rows, errors, and completion wall time. They sample successful results 1, 17, 33, and so on, with fixed cell and string/byte limits. Complex values remain explicitly unmeasured. Semantic ownership works without source maps. **Reason and benefit** A short census can identify which operations issue the most queries or return the most rows. It preserves each query's execution owner through pooled connections, concurrent operations, and prepared-query reuse. It reduces diagnostic work without claiming per-query CPU attribution. **Breaking changes** Opt-in capture output moves from `entry.v1`/`summary.v1` to `entry.v2`/`summary.v2`. Sampled primitive byte lower bounds replace all-result byte estimates. Consumers must use the new fields and must not extrapolate them into a total. There is no database migration or application API change. ## What Changed - Added `withDatabaseOperation(owner, action)` to the DB package. The server owns semantic labels. Queries retain execution ownership in their completion closure. - Added fixed owners at existing HTTP, delivery, budget, and execution timing boundaries. Delivery work replaces the context of the producer that woke it. - Added complete call/row/error counters and completion wall durations within the existing bounded window. Durations include waits and are not CPU measurements. - Sampled and bounded primitive-value sizing. Complex and oversized values remain unmeasured. Known owners skip repeated stack extraction. - Preserved the deadline, group limit, paced output, query outcomes, and existing coverage exclusions. Updated the operator documentation. - Added real PostgreSQL owner-isolation and prepared-query tests, server scope tests, and plain bundled-runtime proof. - Moved projection-test and benchmark full-response sizing into an explicit test helper. Existing payload-size thresholds remain unchanged; these synthetic checks do not use the production sampler. ## Verification - Final-commit DB capture suites: 43 tests passed in 13.86 seconds against the actual PostgreSQL driver, including concurrent owners, prepared reuse, transaction failure, a dedicated client, sampling, and a query crossing the deadline. DB typecheck passed. - Server attribution and existing delivery/execution/budget suites: 53 tests passed. Final concurrent HTTP GET/JSON POST scope refinement: 5 tests passed. - Runtime bundle suite: 8 existing tests passed. The new no-source-map owner test passed after its fixture was corrected to use a driver thenable. An ordinary native Promise can bypass an overridden `.then`; the real PostgreSQL tests separately cover driver integration. - Observer-only experiment: three alternating pairs per case on Node 24.22.2, macOS arm64, 3,000 calls per run after warm-up. For an already-materialized synthetic JSON result, median capture CPU fell from 82.37 to 1.87 microseconds/call. For a 100-row primitive result, it fell from 8.10 to 4.40 microseconds/call. Every result retained identity and all calls were counted. The new diagnostic intentionally measures less byte data; these are observer costs, not database latency or a hosted application speedup. Disabled-case timings were below one microsecond/call and too small for a performance conclusion. - Final lifecycle regression: 28 unit tests and DB typecheck passed. Closing the last capture disables and replaces owner storage; overlapping captures retain it. - Final-commit full recursive typecheck passed in 38.38 seconds and full build passed in 40.81 seconds. The build stamp matches `83ed4ba9cebd68d48929c46f15ab32d5c0b9687b`. Focused server attribution tests also passed 5/5 before the test-helper-only migration. Full test acceptance will use the repository's parallel CI gates; the unsharded local command takes hours. - Verification failures retained: initial sandbox runs could not open local IPC/PostgreSQL listeners; permitted reruns exercised the real services. One old test expected JSON serialization and was updated to assert explicit unmeasured complex values. The first bundle fixture used the wrong promise boundary, as described above. A retained-callback regression also demonstrated that disabling and reusing the same AsyncLocalStorage instance could resurrect an old owner; replacing the instance after the last capture closes fixes that case. - The first CI run found older projection tests consuming the previous full-response fields. Those consumers now use the full-response test helper, preserving their byte thresholds. The two complete affected route suites plus helper tests passed 115/115 in 73.08 seconds. A temporary full-row query negative control then failed the unchanged 2,048-byte limit at 5,000,812 bytes and was restored; five other service suites passed 34/34 and the affected chat cases passed 3/3. All three previously failing CI shards passed on the corrected head. A local skill-list benchmark smoke check completed 255 captured queries with no in-flight queries or overflow; this was integration verification under concurrent test load, not a timing benchmark. - Two final-head runner jobs lost communication with GitHub before reporting a test result. One targeted retry of those jobs and their dependent aggregate passed. All 53 final-head checks passed, with 3 intentional skips; previously passing suites were preserved. - Hosted capture remains pending until the reviewed revision is deployed and a bounded window is explicitly configured. Reproduce the focused checks with Node 24 and the repository pnpm version: ```sh pnpm exec vitest run packages/db/src/response-capture.test.ts packages/db/src/response-capture-postgres.test.ts pnpm --dir server exec vitest run --config vitest.config.ts src/__tests__/database-operation-attribution.test.ts scripts/bundle-runtime.test.mjs pnpm -r typecheck pnpm build ``` ## Risks - Owner context is active while a valid future capture window is armed, up to 30 minutes before accounting starts. No-window and expired calls take the direct path. Diagnostics still add bounded work while enabled. - A nearest nested owner is reported. Broad `http.request` and unattributed groups remain visible; this does not identify every service callsite or SQL shape. - The existing census excludes tagged/raw explicit-execute/streaming paths and internal protocol commands. Its counts are not PostgreSQL transaction totals. - Completion wall time includes queue/database/network/scheduling waits. Concurrent durations overlap. Sampled bytes exclude complex values, large strings, and values beyond the sizing limits. - Activation is startup-only. Operators must choose the window after artifact preparation, allow rollout/warm-up time, and remove the expired setting afterward. ## Model Used OpenAI Codex, GPT-6 family, with code execution, repository inspection, and parallel agents. The exact deployment model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
|
df6156c2e3 |
fix(server): wake native status reassessment from committed changes (#17709)
## Thinking Path > - Paperclip manages agents, tasks and their durable native results. > - Changed evidence, task authority or review retirement can require a new status assessment. > - Periodic orphan recovery repeatedly reads history to find these changes. > - This change gives reassessment an owner driven by committed changes and saved deadlines. > - Existing result, assessment, decision and task rows remain durable truth. > - The status committer still verifies current sources and authority before writing. > - Settled or insufficient evidence becomes quiet until a supported producer changes it. ## Linked Issues or Issue Description Refs #17702 (merged), #17679 and #17680. This is slice 50, now based on master. It owns native status reassessment; workspace settlement, retained sessions and other orphan duties remain separate. **What happened?** Periodic recovery repeatedly inspects completed native results. Removing that scan without complete writer coverage or current authority checks could strand new evidence or commit an old result over newer work. **Expected behavior** Reassess from existing durable facts after relevant outer transactions commit. Retain real retry/lease deadlines and restart/drain recovery. Preserve current authority, quality-review and supersession checks. Make no recurring SQL when work is settled or dormant. **Steps to reproduce** Complete a native result awaiting evidence or review, leave it idle, then approve a real work product or retire the selected automatic review. Its saved result should advance once when authorized, or record the existing nonterminal business gate, without a periodic scan. ## What Changed - Extract reassessment and its pure status policy from finalization recovery into their owning modules. Add a scoped worker using the existing delivery coordinator and scheduler; no queue table or migration is added. - Retain the three durable discovery triggers: changed work-product history, authoritative task status/version, or retirement of the selected automatic completion review. Preserve selected-source, ownership, effects, retry and lease fences under the existing committer. - Notify from actual task/work-product, claim release, approval, interaction, OAuth, event evidence and workspace-settlement owners before their transaction writes. Keep provider calls outside short transactions and nested writes behind outer commit. - Scope ordinary milestone/finalization/coalesced-input wakes to an issue/run. Keep genuine company authority restoration explicit. Real Photon and coalescer regressions cover all three response/status queues. - Move the current parent's required quality policy into the owner and apply it to every reassessment that could complete a task. An unmet review records the existing in-progress decision and next action, avoiding an operational retry loop. Preserve conversation/session reset locks and the workspace omission-barrier signal. - Skip reassessment in periodic recovery while preserving explicit/startup recovery and the historical supersession guard before independent workspace work. Document the source contract, producer inventory, failures and qualifications in the linked audit. ## Verification - Broad source `9fddb3f0b1157df32b64fb1f22e9d6e9624470e8` passes **2731 distinct cases in 28 complete files**, with no skips: 2458 producer/owner/authority/workspace/quality cases in 394.43 seconds and all 273 governed-response cases in 53.74 seconds. Nineteen dedicated fresh PostgreSQL 17 databases migrate through 0413; eighteen embedded-only invocations preserve fresh-per-call isolation through the existing recovery helper. All 22 temporary imports restore exactly and all 28 hashes match. - Final runtime source `5ea3b0fd016bf7e63ba82a7044825a46a0cd2082` passes **1211 cases in five complete affected files** in 108.23 seconds after the narrowly reproduced evidence-trigger fix: all 43 worker, 145 corpus, 32 finalization-recovery, 968 native-session and 23 quality-review cases. Four fresh databases plus the quality helper's own fresh fixture use PostgreSQL 17. Its one adapted import restores exactly; all five hashes match. The other 27 broad-run test files are unchanged. These qualifications are source-scoped and not cumulatively double-counted. - Required/optional real regressions reproduce retry loops on retired-review and evidence-approval paths before repair; both become SQL quiet afterward with unchanged assertions/time limits. Controlled removal of only ordinary scope guards reproduces two coarse-hint SQL failures; exact guard restoration passes. Real producer, outer transaction, claim, authority, 103-item backlog, retry and drain cases pass. Photon transport is mocked; callbacks, locks and observed SQL are real. - Normal merges preserve 12661 parent/unrelated entries, 35 owned entries and 21 automatic shared results, with five original conflicts reviewed explicitly. The parent's actual squash reconciliation preserves the entire held child tree. The final quality fix changes only its owner and tests. Fourteen added comments are removed with identical nontrivia tokens; no directives are removed and the final comment check is empty. - Database build, server types/build, boundaries and exact-head migration order pass. The database tree is identical to merged #17702: no new SQL, queue table or snapshot is generated. Earlier benchmarks, host limitations, failed assertions and their repairs remain in `doc/plans/2026-10-26-native-status-reassessment-sleep-audit.md` and `doc/idle-native-status-reassessment.md`. - This qualifies targeted external PostgreSQL behavior, not embedded bootstrap on the semaphore-exhausted host, a full local repository suite, deployed sleep or measured fleet savings. Fresh exact-head CI and review remain required. ## Risks - Notifications retain the process-local supported-writer contract. Independent SQL writers and uncoordinated replicas are not covered by a new transport. - Reassessment competes with current authority changes. Existing row/source locks and retries remain required; arbitrary missing identities have no row to lock, so this is not a general predicate-lock guarantee. - Discovery pages 100 rows but retains outstanding candidates in memory. It remains proportional to backlog size, and cancellation occurs between queries. - Other orphan/workspace/session duties and Cloud sleep eligibility remain unchanged. Missing supported producer coverage could strand dormant work, so the producer audit and real callback/transaction tests remain part of acceptance. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, repository tools, shell execution and PostgreSQL tests. The runtime does not expose a more specific serving snapshot or context-window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (targeted scope is stated 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> Co-authored-by: Codex <noreply@openai.com> |
||
|
Naeem AkhtarandPaperclip
|
f9581e533e |
perf(server): find the native event cursor through the source seq index (#16140)
Find the contiguous native cursor in PostgreSQL while preserving gaps, late arrivals, source isolation, and event replay. Verify a long-history response budget against current master. Co-Authored-By: Paperclip <noreply@paperclip.ing> |
||
|
Devin FoleyandPaperclip
|
667c79ded2 |
fix: prevent retry-exhaustion events from exhausting attention-feed memory (#13451)
## Thinking Path > - Paperclip manages AI agents and their work. > - Its attention feed shows failed runs whose retry budget is exhausted. > - Startup retention builds this feed before startup completes. > - The feed joins every exhaustion event to the full run context, then removes duplicate runs in JavaScript. > - Recovery can revisit an exhausted run and append the same event again. This multiplies the data loaded into memory. > - This pull request selects one small row per run in PostgreSQL and makes exhaustion writes idempotent. > - Existing duplicate events can stay in the database without multiplying run contexts in server memory. ## Linked Issues or Issue Description Refs #13367. This fixes the repeated-event allocation path in the retention feed. Other full-feed sources and sweep cadence remain separate concerns. **What happened?** The attention query loaded one full run context for every matching exhaustion event. Deduplication ran only after the driver had loaded those rows. A run with thousands of exhaustion events therefore produced thousands of context copies. The startup retention sweep can exhaust the server heap while reading this result. **Expected behavior** The query should return one row per exhausted run and only the context fields that the feed needs. Repeated checks of the same exhausted retry budget should reuse the original event. **Steps to reproduce** 1. Create a failed run with a 32 KB context and 2,500 matching exhaustion events. 2. Build the attention feed, including dismissed items, as startup retention does. 3. Inspect the database result before JavaScript feed processing. The old join returns 2,500 copies of the run context. 4. Call bounded retry scheduling repeatedly for a run at its retry limit. The old writer appends another exhaustion event on every call. **Paperclip version or commit** The attention query was introduced by #9380 and is present in stable `v2026.831.1`. The retention caller is also present in that stable release. The later recovery callback added by #13075 provides a repeated path into the exhausted-budget writer. This change is based on `c0fda8fac` after rebasing onto current master. Related PR: [#12162](https://github.com/paperclipai/paperclip/pull/12162) changes retention cadence and newer-run suppression queries. This change addresses the exhaustion-event join and duplicate event writes. ## What Changed - Select the newest company-scoped exhaustion event per run with a PostgreSQL `DISTINCT ON` subquery before joining run data. - Project only `issueId` and `taskId` from run context. Preserve JSON types, fallback behavior, run ordering, company filters, and run/agent status filters. - Reuse an exhaustion event for the same run, reason, attempt, and retry limit under the existing run-row lock. Recognize historical events without a migration. - Skip sequence allocation and live publication when a receipt already exists. - Document the run-log behavior and add PostgreSQL regression tests. ## Verification - Focused attention, retry-scheduling, and event-sequencing suites: 72 tests pass on the rebased head `39ce77158`. - The large-history fixture verifies two database result rows under 4 KB, newest-message selection, task-ID fallback, and company/status filtering. This assertion runs before feed deduplication. - Concurrent event tests verify one receipt across two database clients, reuse of historical receipts, distinct reason/attempt/budget keys, and stable event sequences. - Repeated scheduling through a new service instance produces no extra run-log or live events. - `pnpm --filter @paperclipai/server exec tsc --noEmit`: passes. - `pnpm -r typecheck` and `pnpm build`: pass on the rebased head, including native runner checks. - Greptile: 5/5 on `39ce77158`, with no review threads. - [GitHub CI](https://github.com/paperclipai/paperclip/actions/runs/34929002304): all 25 workflow jobs pass, including all server/workspace test shards, serialized server suites, browser tests, typecheck, build, native-runner verification, and the release dry run. Security checks also pass. The branch has no merge conflicts. - `pnpm test:run`: stopped with unrelated failures. Two chat integration cases passed when rerun separately (2 passed, 993 unselected). Three skill-cache cases failed on both this branch and the unmodified parent commit, with `EACCES` during directory rename. The full suite is not reported as green. ## Risks - No schema migration or data cleanup is required. The query still scans matching event history in PostgreSQL; its result size now scales with exhausted runs. - This does not bound every source in the attention feed or change retention scheduling. - An exhaustion receipt is emitted once per retry decision. Consumers that observed repeated copies will now receive one event. - Native source-event replay handling stays on its existing path. - A deployed application boot has not been verified. ## Model Used OpenAI GPT-6, used through Codex with reasoning, repository inspection, code editing, and test execution. The runtime does not expose a more specific model snapshot 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 the focused tests locally and they pass (full-suite limitations are documented above) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandDev Agent
|
25cf079ec5 |
feat(runner): add Codex-native application integration (#12591)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The runner package is useful only when the application can start, observe, and recover a native Codex run safely. > - Existing direct adapters must keep their current execution and finalization paths. > - The application boundary therefore needs additive persistence, authorization, coordination, and recovery behind an explicit experimental adapter. > - This pull request adds that Codex-only boundary without activating generalized providers, remote environments, or the later task/SDK surfaces. ## Linked Issues or Issue Description **Subsystem affected** Shared contracts, database persistence, adapter utilities, server native-runtime services, and the experimental Paperclip Runner adapter. **Problem or motivation** The already-landed runner package has a qualified Codex path, but the application needs durable native-run state, guarded runtime selection, authenticated coordination, tool security, finalization, and recovery before the experimental adapter can be exercised safely. **Proposed solution** Add a Codex-only `paperclip_runner` application path behind the existing default-off native-runner setting. Bind native state and coordination to company/run identity, preserve persisted-run recovery, and leave every direct adapter on its existing legacy execution path. **Alternatives considered** The earlier stack boundary introduced a generalized executor and remote-environment lifecycle here. That made this PR depend on implementations in higher PRs and changed reusable sandbox behavior globally. Those pieces are now deferred together to #12592. **Roadmap alignment** ROADMAP.md does not list a conflicting native-runner integration project. This change adds the application boundary for the existing Runner architecture. ## What Changed - Added native run/result/finalization/provider-trace persistence, shared validators, and idempotent migration/replay coverage. - Added guarded Codex-only runtime selection, authenticated PRP coordination, recovery, finalization, and interaction services. - Added run/company-bound tool-gateway authorization, credential redaction, SSRF protections, and replay-safe behavior. - Added the explicit `paperclip_runner` adapter behind the default-off rollout setting. - Preserved legacy answered-question wake projection and direct-adapter execution/finalization paths. - Hardened cancellation so only owned in-memory child processes are signaled; persisted recycled PIDs/process groups are never trusted. - Retained the narrow Claude ACPX isolated-context security follow-up discovered after #12590. - Deferred the generalized executor, provider ingress, remote lifecycle, SDK/lab/eval work, release-process changes, and lockfile. ## Verification - Changed-file delta against `master`: 133 files. - GitHub Actions is the authoritative verification environment for this PR. - Full CI, security, and Greptile review will run on this lowest unmerged stack PR. - Local tests/build/typecheck were not run because this checkout is resource constrained. - Static diff/reference checks pass, and `pnpm-lock.yaml` is unchanged. ## Risks - This touches central heartbeat and agent-route code, so legacy compatibility is the primary risk. - Runtime selection remains Codex-only and explicit; direct Codex, Claude, OpenCode, process, HTTP, and plugin adapters remain on their existing paths. - Fresh native starts fail closed while the rollout flag is off; persisted native records remain readable and recoverable. - Cancellation, company/run binding, tool calls, status decisions, and completion writes are guarded or replay-safe. > 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. ## Model Used OpenAI Codex, GPT-5.6, with repository tools, code execution, and parallel 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 linked existing issues or described the issue in-PR following the relevant issue template - [x] I have not referenced internal or instance-local Paperclip issues or links - [x] My branch name describes the change and contains no internal Paperclip ticket id - [ ] I have run tests locally and they pass — GitHub Actions is authoritative for this resource-constrained checkout - [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 risks above - [ ] All Paperclip CI and security gates are green - [ ] Greptile is 5/5 with no open actionable findings - [x] I will address all Greptile and reviewer comments before merge ## Stack - Position: 3 of 5 overall; lowest of 3 currently unmerged - Base: `master` - Previous: [#12590](https://github.com/paperclipai/paperclip/pull/12590), qualified Claude ACPX runtime — merged - Next: [#12592](https://github.com/paperclipai/paperclip/pull/12592), generalized Codex executor, task experience, and developer SDKs --------- Co-authored-by: Dev Agent <dev@paperclip.ing> |