Repository navigation
Record requeue source in attempt results and count automatic continuations #201
Description
What to build
Every attempt-result comment records how its implementation attempt was requeued: by a human (the issue carried ready-for-agent at claim time) or automatically (the issue carried only round-finished at claim time). From an issue's comment history, the tracker can then tell how many automatic continuations the issue has had in a row since its last human requeue.
This is groundwork for automatic continuation of round-finished issues. It does not change the runnable queue yet: today every claim is a human requeue, so every new comment records "human".
Design decisions:
- The requeue source is a hidden HTML marker in the attempt-result comment, next to the existing
<!-- agent-attempt: <started_at> -->marker. The count lives on GitHub, not in local state, so it survives restarts. - Only comments written by the agent's own GitHub identity count. Any other comment is ignored, even if it contains the marker.
- The count is the number of automatic-requeue result comments in a row, starting from the newest, up to the most recent human-requeue result comment. A human requeue resets it to zero.
- Result comments written before this change have no requeue marker and count as human requeues. An issue with old history therefore starts at zero.
- The requeue source is decided once, from the claim snapshot (the labels the issue had before claiming removed
round-finished), so the comment reflects the actual reason for the claim.
Acceptance criteria
- Every attempt-result comment the agent posts (all attempt outcomes, including recovery/emergency paths that post results) carries a requeue-source marker: human or automatic
- The requeue source is derived from the claim snapshot's labels:
ready-for-agentpresent means human, otherwise automatic - The tracker exposes the automatic-continuation count for an issue, computed only from the agent's own comments
- Count semantics are covered by tests: no history returns 0; only legacy unmarked comments return 0; human, auto, auto returns 2; auto, human returns 0; markers in other users' comments are ignored
- The runnable queue,
next issuevalidation and the release sequence behave exactly as before -
CONTEXT.mddefines Automatic continuation (an implementation attempt claimed fromround-finishedwithout a human requeue) and the requeue source it is counted by
Blocked by
None (can start immediately).
Activity
lbacik commentedon Oct 11, 2026
Agent Attempt Result: handoff
Branch: agent/issue-201
PR: none
Handoff note commit:
c6c2c0e4705efc94e5ecce21b2602ef608e7c348Evidence
UV_PROJECT_ENVIRONMENT=/home/agent/venv-py311 uv run --python 3.11 --extra dev ruff check && UV_PROJECT_ENVIRONMENT=/home/agent/venv-py311 uv run --python 3.11 --extra dev pytest && UV_PROJECT_ENVIRONMENT=/home/agent/venv-py313 uv run --python 3.13 --extra dev pytest→ exit code NoneDetails
Handoff note: issue #201
Summary
The requeue-source groundwork spec is fully implemented and committed as
b0f5044 on
agent/issue-201. Every attempt-result comment now carries ahidden
<!-- requeue-source: human|automatic -->marker next to the<!-- agent-attempt: … -->marker (normal publication for all outcomesplus the recovery-release path); the source is decided once from the claim
snapshot and stored on the attempt checkpoint so recovery keeps the
claim-time decision; and
GitHubTracker.automatic_continuation_count()counts automatic-requeue results in a row from the agent's own comments,
treating legacy unmarked results as human requeues.
CONTEXT.mddefinesAutomatic continuation and Requeue source. Full suite: 896 passed,
9 skipped;
ruff checkpasses. An inline code review (Standards + Spec)found no violations or scope creep; the parallel review sub-agents could
not be launched (token soft threshold).
Remaining work
reviewed inline, and committed. A continuation only needs to publish:
push the branch, open the PR, and post results per the driving process.
("human", "automatic")literal isrepeated in
recovery.pyandattempt_state.pyto keepattempt_statedependency-free; a shared constant could be introduced if the team prefers.
Evidence
git show b0f5044 --stat).tests/test_requeue_source.py(claim-snapshot semantics,recovery-retry keeps claim-time source); count semantics in
tests/test_github_tracker.py; marker tests intests/test_publication.pyand
tests/test_recovery.py; checkpoint tests intests/test_attempt_state.py.simple_coding_agent/recovery.py(marker helpers + counting),simple_coding_agent/publication.py(PublicationRequest.requeue_source),simple_coding_agent/attempt_state.py(checkpoint field),simple_coding_agent/lifecycle.py(thread-through at all 7 publish sites),simple_coding_agent/github_tracker.py(
automatic_continuation_count),CONTEXT.md(definitions).Handoff note: issue #201
lbacik commentedon Oct 11, 2026
Agent Attempt Result: complete
Branch: agent/issue-201
PR: #205
Evidence
UV_PROJECT_ENVIRONMENT=/home/agent/venv-py311 uv run --python 3.11 --extra dev ruff check && UV_PROJECT_ENVIRONMENT=/home/agent/venv-py311 uv run --python 3.11 --extra dev pytest && UV_PROJECT_ENVIRONMENT=/home/agent/venv-py313 uv run --python 3.13 --extra dev pytest→ exit code 0Details
Claude SDK completed the implementation workflow.