openclaw - 💡(How to fix) Fix ACP task maintenance accepts stale session-store rows as valid backing sessions [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#74074Fetched 2026-04-30 06:28:55
View on GitHub
Comments
1
Participants
2
Timeline
4
Reactions
0
Author
Timeline (top)
cross-referenced ×2closed ×1commented ×1

Root Cause

Evidence:

  • tasks list can still show the running task from registry state
  • tasks cancel <id> fails in dev/linked installs because task-registry.ts tries to load task-registry-control.runtime.{js,ts} and no built artifact was found under dist
RAW_BUFFERClick to expand / collapse

ACP task maintenance trusts session-store row existence too blindly. When a stale/invalid session-store entry exists for an ACP task, maintenance incorrectly treats it as a live backing session, preventing proper reconciliation and "lost" marking.

Observed Behavior

  1. ACP run leaves a stale session-store row (e.g., after crash or failed cancellation)
  2. ACP task maintenance sees the row and assumes a valid backing session exists
  3. Task/flow is not marked "lost" despite no live ACP process
  4. Operator must manually clear the stale row before maintenance can reconcile

Evidence:

  • tasks list can still show the running task from registry state
  • tasks cancel <id> fails in dev/linked installs because task-registry.ts tries to load task-registry-control.runtime.{js,ts} and no built artifact was found under dist

Expected Behavior

ACP task maintenance should only recognize a backing session if:

  • It is explicitly an ACP-owned session (identity/type marker)
  • It is in a non-terminal state (not failed/cancelled/lost)
  • (Ideally) The ACP runtime confirms the session is still live

Fix Direction

  1. Strengthen backing-session detection in ACP maintenance to require ACP identity + non-terminal state + runtime liveness check
  2. Ensure task-registry-control.runtime is included in build artifacts so openclaw tasks cancel <id> works in dev/linked installs

extent analysis

TL;DR

Strengthening the backing-session detection in ACP maintenance to require ACP identity, non-terminal state, and runtime liveness check is likely to fix the issue.

Guidance

  • Review the task-registry.ts file to ensure it correctly handles the loading of task-registry-control.runtime artifacts, especially in dev/linked installs.
  • Modify the ACP task maintenance logic to verify the existence of a valid backing session based on ACP-owned session markers and non-terminal states before assuming a live session exists.
  • Consider adding a runtime liveness check to confirm the session is still active, as ideally suggested in the expected behavior.
  • Ensure that task-registry-control.runtime is included in the build artifacts to enable successful cancellation of tasks using openclaw tasks cancel <id>.

Example

No code snippet is provided due to the lack of specific implementation details in the issue.

Notes

The provided guidance assumes that the issue lies within the ACP task maintenance logic and the handling of task-registry-control.runtime artifacts. Further investigation into the specific implementation details may be necessary to provide a complete solution.

Recommendation

Apply the workaround by strengthening the backing-session detection in ACP maintenance, as this directly addresses the observed behavior and aligns with the expected behavior described in the issue.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix ACP task maintenance accepts stale session-store rows as valid backing sessions [1 comments, 2 participants]