Skip to content

auto-claude 1.3.0: wait on compute without ending the run - #41

Merged
unohee merged 1 commit into
mainfrom
feat/auto-claude-wait
Nov 15, 2026
Merged

unohee merged 1 commit into
mainfrom
feat/auto-claude-wait

Conversation

@unohee

@unohee unohee commented Nov 15, 2026

Copy link
Copy Markdown
Contributor

Why

The ArtifactNet run on vela (session 287dfa7e) stopped because of how it handled compute waits, not because anything hung:

  • Milestones had prose start conditions such as (start only after … shows DONE). The driver could not read them, so it started each line at once, and the model deferred it to [!].
  • Once every line was deferred, the final check failed and a round was spent. On 11-09, all 3 rounds went in 25 min.
  • On 11-15 the run stopped at round 3/3 (03:17Z). The cache finished at 06:19Z, and nothing moved until the person asked at 08:27Z.
  • The Monitor refusal told the model the driver would prompt it. That was false for compute.

What

  • New queue fields. :: after: / :: alive: / :: within: go after the check, with base last. The driver polls after every 3 min from the repository root and builds other lines meanwhile.
  • Waiting is not the end of the run. While lines wait, no final check runs and no round is spent, and the run does not count as idle. Waiting time is left out of maxHours, up to maxWaitHours (72).
  • Dead or late jobs. If alive fails twice, or a wait passes its within (12 h by default), the model is asked once. A wait past twice its within is held.
  • wait_until tool. It puts the open milestone back to wait on its own branch. A stacked milestone keeps its onto.
  • Plan checks and prompt text. validatePlan refuses parenthesised prose start conditions. Every text that steers the model now points compute waits to wait_until / :: after:, never to defer_goal.

Tests

  • scripts/ci.sh is green: auto-claude 149 pass (was 135), tsc OK, and the shared copies are in sync.
  • The ArtifactNet incident is replayed in a test: only waiting lines are left and the final check fails, yet the run stays on through 5 h of polling and starts the line when after passes.
  • Mutation checks:
    • Removing the "no final check while waiting" guard fails 5 tests.
    • Removing the onto hand-off fails the stacked test.
  • Not exercised in a live session yet. The first real check is the ArtifactNet run after deploy.

Measured in ArtifactNet on vela (2026-11-09 to 11-15). Milestones carried
prose start conditions ("(start only after ~/data/.../cache.log shows
DONE)"). The driver could not read them, so it started each line at once
and the model deferred it. Once every line was deferred, the final check
failed and a round was spent. On 11-09 three rounds went in 25 min. On 11-15
the run stopped at round 3/3, and nothing resumed it when the cache
finished three hours later. The guard refused Monitor and said the driver
would prompt the session, which was not true for compute.

Changes:
- New queue fields after the check: `:: after: <cmd>` must exit 0 before
  the line starts, `:: alive: <cmd>` detects a dead job, `:: within: 12h`.
  The driver polls `after` every 3 min from the repository root (fetching
  first for origin/ refs) and builds other lines meanwhile.
- Lines that still wait keep the run going: no final check, no round spent,
  not idle. Waiting time is left out of maxHours, up to maxWaitHours (72).
- When `alive` fails twice, or a wait passes its `within` (12 h without
  one), the model is prompted once with the last poll output. A wait past
  twice its `within` is held.
- New tool mcp__auto-claude__wait_until { until, alive?, within, reason }.
  At the next stop it puts the open milestone back to wait on its own
  branch. A stacked milestone keeps the commit it was stacked on, so the
  restarted branch is still rebased off its parent.
- validatePlan refuses parenthesised prose start conditions. The plan
  prompt, POLICY, the loop policy, the kickoff and the wait refusal now
  point to `:: after:` / wait_until for compute, never to defer_goal.
- /auto-claude status lists each wait; the status line piece shows ·waitN.
@unohee
unohee merged commit 3e84679 into main Nov 15, 2026
1 check passed
@unohee
unohee deleted the feat/auto-claude-wait branch November 15, 2026 13:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant