Skip to content

fix: stage agent-resolved conflicts before checking unmerged files - #168

Open
cwbcheng wants to merge 6 commits into
jeremymcs:mainfrom
cwbcheng:codex/fix-conflict-resolution-staging
Open

fix: stage agent-resolved conflicts before checking unmerged files#168
cwbcheng wants to merge 6 commits into
jeremymcs:mainfrom
cwbcheng:codex/fix-conflict-resolution-staging

Conversation

@cwbcheng

@cwbcheng cwbcheng commented Aug 6, 2026

Copy link
Copy Markdown

Problem

When a tracked PR has merge conflicts, the conflict-resolution agent often resolves the files in the worktree without staging them (git add). PatchDeck then checks for unresolved conflicts with git diff --name-only --diff-filter=U, which reads the index's unmerged state — not the worktree content. So fully-resolved files still show as unmerged and the run is falsely declared failed with codex left unresolved merge conflicts: <files>, then falls into the code-owner fallback path. The resolution is never committed or pushed, so GitHub keeps showing "This branch has conflicts that must be resolved" forever.

Observed in production on PR cwbcheng/OpenAgent#433: the agent reported "All merge conflicts have been resolved … nothing staged/committed", then PatchDeck logged PR work error: codex left unresolved merge conflicts: tools/novelstudio/embodied-agent-runtime-control/src/contract-shapes.mjs … and the three conflict files stayed unresolved on GitHub.

Fix

After the conflict-resolution agent completes successfully, stage everything in the worktree (git add -A) before the unmerged-index check. If the agent actually left conflict markers behind, the existing marker check (git grep for <<<<<<< etc.) still catches them and fails with the more accurate "left merge conflict markers" reason, so genuinely-broken resolutions are still escalated.

Tests

  • Added regression test babysitPR commits agent-resolved conflicts even when the agent does not stage them. It simulates the agent editing the conflicted file without staging: the unmerged index check only clears after PatchDeck runs git add. Fails without the fix (run fails with "left unresolved merge conflicts").
  • Verified: node --import tsx --test server/babysitter.test.ts (111 tests pass), npm run check passes.

Rejected review-thread feedback previously fell through the GitHub
follow-up path silently: needsGitHubFollowUp only considered accepted
items, so a rejected review comment was never replied to and its
conversation was never resolved, contradicting the agent prompt that
says to reply with the rejection reason and resolve the conversation.

Now rejected review threads (excluding PatchDeck's own status/audit-trail
replies) are collected as follow-up tasks, get a 'Rejected - no code
change made' reply with the reason, and the conversation is resolved.
Audit-trail verification skips the code-change audit for these items and
only requires thread resolution, and the final status update keeps the
rejection decision while advancing the thread-resolved flag.
Localize the comments PatchDeck posts to GitHub:
- Agent prompt now requires Simplified Chinese for every comment,
  thread reply, and summary.
- Follow-up replies use Chinese headlines (已拒绝/已在提交/已在最新更新)
  with 原因 for rejected threads, and 正在回复 for quoted references.
- Status messages (accepted/in-progress/needs-attention/verifying/resolved)
  are localized.
- APP_STATUS_COMMENT_PATTERN now recognizes both English and Chinese
  status markers so recovery of existing status replies keeps working.

Tests updated to assert the Chinese wording.
The internal-reply exclusion in needsGitHubFollowUp only matched
'PatchDeck status comment' and 'Automation audit trail follow-up', but
classifyNonActionableAppFeedback also produces 'PatchDeck audit trail
comment' and 'PatchDeck agent command comment'. A rejected item carrying
one of those reasons was treated as real reviewer feedback, which made
PatchDeck reply to its own follow-up comment, producing a new audit
trail, which was then replied to again - an infinite loop.

Widen the exclusion to all PatchDeck-authored marker reasons and add a
regression test that a rejected audit-trail comment never triggers a
follow-up reply.
The code-owner fallback agent is asked to commit and push its own
changes, but agents sometimes finish without doing so (e.g. a truncated
model response). Previously PatchDeck marked the fallback run completed
and removed the worktree, discarding the agent's edits and leaving the
PR branch and review threads untouched.

After a successful fallback agent run, PatchDeck now:
- commits any uncommitted agent edits in the worktree,
- pushes the local head to the PR head branch,
- verifies the remote head matches the local head.

Adds a regression test that a fallback agent leaving uncommitted edits
results in git add/commit/push to the PR branch.
The DeepSeek-backed codex agent frequently times out on long review
comments: evaluation had a 3-minute cap and apply had a 15-minute cap,
which was not enough for multi-file fixes (e.g. 21 generated fixtures).
When apply timed out mid-run, the uncommitted worktree edits were
discarded.

Raise evaluation to 7 minutes and the default apply timeout to 30
minutes (matching the code-owner fallback timeout) so complex review
threads can be evaluated and fixed to completion.
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