Repository navigation
feat: guard unregistered Stakpak TUI pending native lifecycle fixes - #6550
nikhilachale wants to merge 5 commits into
Conversation
🏆 Review leaderboardOct 4–11, 2026 · UTC
Ranked by distinct external PRs reviewed, then review rounds, then PR comments. Self-activity and bot activity are excluded. Show 5 more reviewers
|
|
Stakpak native qualification is BLOCKED, not an AO integration pass. Immutable curated report and receipts at PR6509 evidence head Official v0.3.88 / Z.ai glm-5.3-flash native proof established correct-workspace edits, additive private instructions, preserved AGENTS.md, native manual approvals, real-history restore and fresh token consumption. The exact submitted initial task occurred once in durable history; positional argv alone did not submit it. FAIL cancellation: manually approved shell PID932367/kernel-start78005410 and sleep survived Escape, reparented to PID1092, and wrote the delayed side effect. FAIL instruction deduplication: native restored checkpoint had2 system messages/2 original standing-token occurrences. Native foreground cancellation calls immediate child.kill; documented background cancel_task does not qualify this boundary. Non-restorable scope does not solve the shell leak. No monkeypatch/security disablement. AO sessions/screenshots/Kill/control/activity admission: NOT_RUN / unavailable because production admission was withheld. No fabricated native ID/ACP/reviewer/handoff. All owned native/child PIDs were absent after the witnessed deadline and exact tmux teardown; profiles/history and production PID625660/service preserved. Routine verification is current-head GitHub CI only; no local suites run. Leave draft pending a released native fix and actual AO proof. Skill reference synced to the installed VPS skill. |
|
Final stage receipt: source Status remains BLOCKED / draft / unregistered. Released Stakpak0.3.88 Escape left a witnessed shell/sleep alive and its delayed artifact appeared. Exact native restore consumed new private instructions but accumulated2 persisted system messages. Supported background cancel_task and non-restorable scope do not resolve foreground cancellation. Immutable native evidence, report/JSON download/hash verified. Actual AO UI screenshot and AO activity/admission/Kill/control gates remain NOT_RUN/unavailable because native admission failed. This is not an integrated harness claim. Owned native/child PIDs were confirmed absent; native profiles/history, prior evidence, dirty/preparation worktrees and production service/PID625660 preserved. Installed skill docs synced; runner unchanged. Supervisor may proceed to the next candidate. Requalification requires a released native fix, then actual AO/Electron proof; do not merge or deploy. |
|
Recovery source analysis and exact native patch/regression proposal: https://github.com/nikhilachale/agent-orchestrator/blob/ae5d4052b54ae0d316d49e3585f9c7f0cd97ac92/docs/research/stakpak-recovery-proposal-2026-10-12.md The official latest release is still0.3.88. Its foreground executor creates a separate group but kills only the immediate child; the CLI returns cancelled before cleanup acknowledgement. Upstream main has no repair. Supported tool advertising is not enforced exclusion, reserved native MCP replacement is prohibited, ACP shares the executor, and fresh-only scope does not fix ownership. The proposal covers owned job lifecycle, both cancel branches, timeout/drop/disconnect, PID reuse, explicit cleanup acknowledgement and same-session regression controls; instruction renewal/ACP restore are separate proposals. Prior witnessed survivor/late-side-effect failure is preserved. No unchanged Escape rerun, vendor monkeypatch, native permission weakening or fabricated AO proof. Stakpak remains unregistered and PR6550 remains draft. Actual AO screenshots and independent Kill are unavailable for this guarded candidate; Kill cannot substitute for same-session cancellation. Installed skill lessons updated; published report downloaded and SHA256 verified: e19fe26fe9763ca308563334db6d37cacf685cb75ad891db3aa9119300c6ed63. Current-head remote CI is pending; prior green checks are historical. No local suites, new native/provider process, native profile/credential edit or production activity. |
|
Recovery exact-head CI: PR6550 Recovery result and exact remaining upstream prerequisite are recorded privately at harness-recovery-20261012/reports/stakpak/result.json; published source patch/regression proposal: https://github.com/nikhilachale/agent-orchestrator/blob/ae5d4052b54ae0d316d49e3585f9c7f0cd97ac92/docs/research/stakpak-recovery-proposal-2026-10-12.md . No AO screenshot available for the guarded/unregistered candidate. Stock binary hash unchanged, prior owned PIDs absent, no new process to clean. No merge/release/deploy/upstream post. |
|
Failure report and fix implementation started (2026-10-12). Stock Stakpak v0.3.88 remains unqualified: after a witnessed foreground shell started, native Escape left descendants alive and a delayed side-effect file appeared. The pinned executor creates a process group but the cancellation/timeout paths kill only the immediate child. Cancellation forwarding/future-drop ownership also needs correction. The existing source trace and reproducer preserves the failure. The earlier document was a proposal, not an implemented fix. A durable VPS implementation worker is now applying the native source correction in an isolated upstream checkout, with focused executor/transport regression tests and a real fix PR. Relevant tests and native build will run in GitHub CI only; exact-head artifacts will be used for native and AO qualification. This AO PR stays draft until its required proof is obtained. No stock-native success or AO screenshot is claimed yet. Fresh-only scope will continue to reject restore; the distinct history/duplicate-instructions defect is not silently waived. Actual fix PR, CI results and labelled prerelease AO screenshots will be linked here as available. No Mac source changes/tests/builds; no merge or release. |
|
Concrete native source fix is now open for review: stakpak/agent#767 (draft; The patch owns the foreground command child, signals its Unix process group while its leader is unreaped on cancel/timeout/drop, bounds explicit cleanup, and preserves successful-command behavior. Real MCP/executor CI regressions cover both cancellation branches, timeout, coordinated disconnect, execution-future abort, nested shell/sleep identities, unrelated control and subsequent requests. CI also requires the stock executor to fail the cancellation regression. Validation and exact CI-artifact Escape/AO qualification are still pending. Stock0.3.88 remains FAIL; PR6550 stays unregistered/draft. No restore/history fix or Windows tree guarantee is claimed. All routine validation runs in GitHub Actions; no local suites or native compilation were run on the VPS. |
|
Concrete Stakpak native foreground fix is implemented in stakpak/agent767, source Current qualification job: 9/9 PASS, compiled stock negative control, affected lint and release build. Full upstream workspace/feature suite: CANCELLED/NOT_VERIFIED after repeated hosted stalls/timeouts and unavailable logs; native PR stays draft. No routine local suites/builds/native compilation ran. Exact artifact passed22 native checks in the actual AO Electron standalone terminal: initial leading-dash task once in history, real cwd, private/project instructions, AGENTS preservation, initial/second/third active→settled turns, observed foreground shell/sleep PID/start witnesses, Escape with immediate physical PID absence, no25-second delayed artifact, unrelated PID/start/heartbeat control, same native identity and third turn after cancellation. Tool timeout35s exceeded the cancellation time. This is supplementary native/runtime evidence; AO6550 remains unregistered/draft, with no manager/agent-session Kill, restore, Chat, handoff, catalog or registered-auth claims. Immutable detailed report and receipt. All four PNG URLs were downloaded and SHA-256 verified after publication; images inspected for credentials. Backend/UI source UTC2026-10-12T05:56:59Z, native command executing; PNG SHA UTC2026-10-12T06:03:07Z, same native identity after Escape and third count3 turn; PNG SHA Retained failure: earlier16e63ed startup after forced renderer reload inserted OSC11 reply text before the prompt (FAIL, UTC05:17:23Z); actual failed screen. Clean fresh launch is a separate passing control; reconnect behavior remains unqualified. Default security/permissions were preserved. Original stock Escape failure and restore-system duplication remain failures; original AO screenshot unavailable because no registered session existed then. Cleanup: both final shell closes204; shell list200/empty; runtime hosts/native/control PIDs absent; owned Electron main978778 and launch members stopped; scratch daemon979004 shutdown202 and absent; mux helpers stopped and compute lock released. Production625660/ao-self-hosted, profiles/credentials/source/shared git/prior evidence untouched. Runner SHA8887aa2e... unchanged. No merge, release publish, deployment, signup or billing change. |
|
Harness-specific reports, JSON receipts, screenshots and provenance now live on this integration PR branch, rather than the skill PR #6509. Evidence index with embedded screenshots and JSON links. These screenshots show native/prerelease controls in AO standalone terminals. The harness remains gated/unregistered; this is NOT registered AO integration or manager Kill/restore proof. This is evidence relocation only. Existing tested source/native hashes and failures are preserved; the documentation commit is not a fresh runtime pass. Applicable checks run in PR CI. |



Code changes: +175 -0
Tests: +172 -0
Others: +688 -0
Binary assets in this source snapshot: 4 (excluded from line counts).
Draft qualification boundary; unregistered and launch/hooks/restore guarded.
Current status
Current source snapshot:
ae5d4052b54ae0d316d49e3585f9c7f0cd97ac92(basea7860405e4d48333204b6627ec3e83b20a99cc00). Draft; GitHub mergeabilityMERGEABLE, merge stateBLOCKED. Exact-head CI: all 20/20 reported checks successful. CI run 1 · CI run 2 · CI run 3. Green source CI does not establish current native/UI runtime proof or full integration readiness.This fix created only clean isolated source worktrees and temporary review reports; it created no AO/native sessions, provider profiles or owned runtime processes. No local build, gofmt, test, fixture, provider or UI run was performed.
Historical runtime build: No registered AO runtime build is qualified. Historical stock Stakpak0.3.88 failed foreground cancellation and accumulated standing system messages on restore. Supplementary native/prerelease standalone-terminal capture used AO/UI
fb60bd4097ebd4109bbe7ee68e908f722e6ee03f, upstream source62bd8833600b0f906ff841930433264ece36dbab, executable SHA256882e7dce06ce64afe76249befff732a5f17b86c5594291526b5302a323965ac2(reports0.3.88; unreleased patched source). Report · evidence index, original receipts, image/provenance hashes. Receipt SHAs/bytes are unchanged. Linux/VPS proof is not Windows/macOS native proof. Runner SHA2568887aa2e7101b9e6d3950bfca08ffddf2e1e6c1c3f1d030c8dec2676443aecearemains historical where reported; manual/standalone proof is not a strict runner PASS.Changes since runtime proof: Source remains an intentionally guarded draft/unregistered qualification boundary. Evidence relocation did not admit stock Stakpak or rerun native conformance.
The matrix separates current verification from recorded historical passes;
NOT_RUNdoes not erase those passes. Excluded interfaces/manual model input are deliberate supported-scope decisions.Historical implementation and evidence
Released0.3.88 Escape settled the native TUI while the witnessed shell/sleep descendant survived and wrote its delayed side effect. Exact native history continuation also accumulated two standing system messages. Private/project instructions, cwd edits and manual approvals were independent passing controls. Positional argv did not submit the task: after-start input was required. Fresh-only, background-task and tool/MCP alternatives do not cure the foreground ownership defect. The guarded adapter refuses launch/hooks/restore; no production registry/domain/API/migration/UI/reviewer changes are advertised. Historical AO source bb6b356 and recovery42a94816f54a074b56df23c8754cd35277e66125 each had 20/20 green source checks; earlier lint failure is retained.
historical receipt 1 · historical receipt 2
Reproduction and retained-original limitations
The pinned reports/JSON retain action/result, file/token effects, process identity, cancellation, native-history continuity and capture assertions. A complete sanitized executable reproduction bundle is not established for every manual helper: exact helper bytes/hash and invocation, applicable local-contract bytes/hash with probe argv and interrupt/readiness/draft-cleanup settings, and complete native/AO fingerprints are unavailable where not recorded in the linked receipts; recoverability of those original inputs is unknown. No missing command, contract, hash or result is invented. Later reproductions must use a new attempt and cannot replace original failed/missing evidence.
Recorded owned-process/session/daemon/Electron cleanup is historical. Source worktrees, native profiles/history, dirty scratch workspaces, binaries and raw logs were reported retained privately; a complete current inventory/location/hash and recoverability of every original class are unavailable in the public record. Current recoverability is unknown; this reporting fix reconstructs no missing input and silently substitutes none. Share sanitized artifacts only; exclude credential stores, raw DBs and provider-profile directories. Preserve the retained failures and original screenshots.