PR Fixup

Use this workflow directly in the user-started primary conversation. Do not launch a verifier, implementer, or other remediation subagent. A read-only pr-poller is the sole exception: launch it only when the user explicitly asks to wait for or monitor PR updates. For a cost-controlled workflow, the user may switch the same conversation to the lower-cost implementation/test model before starting CI remediation.

Use gh by default; auth or transport errors leave state unknown, never clean. If connector tools are available, use structured PR/check/thread data; avoid dumping full HTML/diffs. Map GraphQL thread IDs to REST comment IDs before replies, and refresh current-head state after pushes and review aggregation.

Pipeline

Create a visible checklist:

1. Gather PR state
2. Resolve an authorized PR merge conflict
3. Fix failing CI checks
4. Triage every review thread
5. Apply valid fixes and record every disposition
6. Commit, rerun affected checks, and push
7. Re-check the new head
8. Report

1. Gather PR State

Before the first GitHub call, obtain any network approval required by the runtime. If the runtime denies access, stop until the user authorizes access.

Run scripts/pr-state --summary <PR> and scripts/pr-resolve list <PR>. Load review-evidence.md for snapshot fields, review classification, hidden threads, and access fallbacks. For cross-repository PRs, use the snapshot's delivery fields as the push target.

Then run gh pr view <PR> --json state,baseRefName,headRefOid,mergeable,mergeStateStatus,reviewDecision. Require matching head SHAs before triage. Verify the base SHA against the current base-ref tip. Refresh mismatched metadata before attributing failures. If the PR is merged or closed, stop. For a merge, report mergedAt and mergeCommit.oid.

Treat peer alerts as advisory until their SHA matches the current head. If captured output is empty, truncated, or invalid, rerun through rtk proxy. Capture stdout and stderr separately and preserve the exit code. An empty capture cannot prove that the PR is clean.

If GitHub reports a conflict, load merge-conflicts.md before CI or review triage. An explicit fixup request authorizes conflict resolution within that request. Otherwise, report the conflict without changing the branch.

If the base advanced without conflicts, use that reference's advanced-base procedure. Base advancement alone does not require a branch update. Record the head, base, synthetic merge, and focused test results. If either input SHA changes, repeat the relevant validation. Keep the helper's exit code unchanged and report any separately verified base drift.

For related or stacked PRs, use the semantic-conflict procedure in the same reference. Reconcile changed contracts and stable IDs before delivery.

For pending checks, load waiting.md. Use scripts/pr-await and its final exit code. For queue membership or enqueue/dequeue requests, load merge-queue.md. For queue removals or failed synthetic checks, also load ci-troubleshooting.md. Ordinary PR-head checks do not validate a merge-group commit.

Directly supplied review findings need current-source validation even without a GitHub thread. Apply valid fixes within the user's authorization. Do not invent comment IDs or resolutions for findings without a thread. External replies and thread resolution still require user authorization.

If reviewDecision=REVIEW_REQUIRED and mergeStateStatus=BLOCKED, report the human approval gate. Green checks do not authorize self-approval or merging.

2. Fix CI Failures

Before changing code, confirm every reported failed check, its run_id, and the parent workflow/job status. A failed job can be visible while its workflow is still in progress; confirm its conclusion and failing step before treating it as reproducible code evidence. Use scripts/run-quiet gh-run -- gh run view <run-id> --log-failed so large logs do not flood the conversation. If it returns only GitHub request/transport lines or no failure text, treat logs as unavailable and, after terminal state, use scripts/pr-state --job-log <job_id>. For temporary gaps or aggregate-only logs, use the same fallback; it handles plain-text and ZIP responses and emits bounded context. Follow references/ci-troubleshooting.md. Reproduce the exact failed command where possible; CI-specific Go lint often needs golangci-lint run ./... --new-from-rev=<base> --timeout=5m.

If CI reports files or commits outside the PR diff, or a stale base SHA, resolve the authoritative base repository, ref name, and current base SHA from PR metadata. Fetch that ref from an explicit base remote, verify its tip matches the reported SHA, and compare git merge-base HEAD <base-remote>/<base-ref> with git diff <base-remote>/<base-ref>...HEAD; do not assume origin/<base> when origin points to a fork. Inspect the parent workflow/run to determine whether a newer base commit caused the failure before changing product or docs code. If the fix is already upstream, update or rebase the branch only when authorized, rerun affected checks, and invalidate all prior exact-head evidence. If the installed gh pr view --json does not expose the base OID, use BASE_SHA=$(gh api repos/<owner>/<repo>/pulls/<PR> --jq .base.sha) alongside gh pr view <PR> --json baseRefName,headRefOid for the ref names. Run this fallback under set -euo pipefail, require a non-empty BASE_SHA, verify the fetched base tip separately, and only then run git merge-tree --write-tree <base-remote>/<base-ref> HEAD.

For unfamiliar, infrastructure, or E2E failures, load references/ci-troubleshooting.md and, for transport or queue evidence, references/transport-troubleshooting.md before changing code. Also load it for unexpected zero-duration or no-op manual-review runs: event and workflow provenance can explain them without a product-code change.

An explicit request to run pr fixup owns every failed required check on the current head. Never stop by calling a failure "unrelated" only because its files are outside the PR diff. Reproduce each leaf failure with retries disabled, inspect its artifacts and shared fixtures/cleanup, and fix every valid product, test, fixture, cleanup, or CI-contract defect it exposes. A dependent aggregate failure does not replace its leaf failures: trace the aggregate to all failed jobs, fix the underlying failures, and wait for the aggregate checks to rerun. If concrete evidence proves a failure is external after this investigation, report the exact job/log/reproduction evidence and keep the PR blocked; do not call it ready while a required check is failed.

Fix with /tdd or /e2e as applicable, run focused checks, and keep each remediation scoped to the reported failure. Do not suppress a failure or mark a check clean without fresh evidence.

If a reproducible failure is outside the PR diff, compare the failing assertion with the current implementation and concurrent or sibling PRs before editing. If it is a stale test expectation, the smallest valid remediation may be a test-only assertion update: keep it limited to the reported failure, run the focused test, and call out that scope. Do not change unrelated production behavior or duplicate a larger sibling change; for request-count/dedup assertions covering multiple hydration or effect continuations, use a deferred response, assert the count while it remains pending, then resolve and drain it before unmount/return so an immediately resolved mock cannot let cleanup timers create a false duplicate. When a remediation changes a documented behavior or contract, update the authoritative spec/guidance, plan, and task file when present; refresh each Verification/Results section with exact post-fixup commands and outcomes; commit docs and code together, keep regression tests and verification commands aligned, and re-check the documentation before completion; record why no update is needed when the behavior remains internal.

3. Triage And Address Reviews

Use scripts/pr-resolve list <PR> to obtain unresolved threads. Before handling multiple threads, make a thread-to-finding map and one body file per thread. Its previews can be truncated, so run scripts/pr-resolve show <PR> <thread_id> immediately before each write; verify comment/thread IDs and that the body names that thread's finding, file, and commit. Use scripts/pr-state --comment <comment_id> only for a flat comment view when no thread context is available. Validate against the current head, spec, and architecture before editing or replying.

Every unresolved thread requires an explicit disposition, regardless of its author, bot identity, visibility, or apparent severity. Record exactly one of these dispositions:

When an automated review identifies a security-boundary risk and the user explicitly chooses the behavior, record the accepted risk, trust/revocation responsibility, and targeted regression coverage in the final response, PR body, and authoritative spec. Do not silently encode the compromise.

• actionable: make the valid code or test change, verify it, cite the resulting commit in the reply, and resolve the thread when writes are authorized.
• already addressed: verify the current implementation or an existing commit, explain where it is addressed, and resolve the thread when writes are authorized.
• informational or optional: make no code change, acknowledge the information or suggestion in a reply, and resolve it when complete cleanup is requested.
• invalid: do not make an invalid code change or silently ignore the finding; reply with concrete reasoning grounded in the code, spec, or architecture, then resolve it only after that pushback is posted and writes are authorized.

A classification is not a disposition. A thread is not complete merely because it was called optional, informational, already addressed, or invalid. GitHub replies and thread resolution are external writes. A request such as "complete cleanup", "clean up all review threads", or "leave no threads unresolved" explicitly authorizes a concise reply and resolution for every unresolved thread, including informational, optional, and invalid findings. A request to address selected comments authorizes replies and resolution only for those selected threads. A review-only request authorizes no writes. When writes are not authorized, record every disposition, report the still-unresolved thread, and do not declare the PR clean.

For code changes, push the fix and pass targeted verification before replying and resolving. For non-code dispositions, verify the current head first, then use the atomic helper path scripts/pr-resolve reply <PR> <comment_id> <thread_id> --body-file <path> to reply, resolve, and react in one operation when the body contains Markdown or shell metacharacters. For short plain-text bodies, a safely quoted argument is acceptable. Never interpolate review text into an unquoted shell command or use backticks in the command itself. After the helper returns, re-fetch the thread and verify the posted reply body and resolved state before treating the disposition as complete. Then rerun scripts/pr-resolve list <PR> and the exact-head scripts/pr-state --summary <PR> check before reporting.

For an ordering or concurrency finding, trace the complete producer → event-bus transport → gateway/client path. Sequential publishes do not prove delivery order when a remote bus uses separate subscriptions; consolidate one stream or add sequence-aware buffering when order is contractual, and cover both the transport boundary and local emulator.

When feedback says an action must remain reachable, add and run a regression at the legal minimum width. Verify the actual hit target (for example, elementFromPoint() at the control center) and clickability, not only toBeVisible, before pushing.

4. Commit, Verify, Push

Commit through /commit, then rerun only the unit, integration, or E2E command affected by the remediation. Push when that targeted check passes for the exact current HEAD. For a cross-repository PR, push the exact current HEAD to the summary's authoritative head repository and ref only when pr.maintainer_can_modify is true. Re-fetch the PR afterward and require its pr.head_ref_oid to equal local HEAD; an upstream remote comparison is not sufficient. Run broad /verify only if the user explicitly requests it or the PR/CI finding requires it. After every push, run fresh gh pr view <PR> --json baseRefName,headRefOid,mergeable,mergeStateStatus (or equivalent) at the pushed head; if GitHub briefly returns an older head, retry the query and exact-head pr-state with short bounded backoff before triaging or reporting. Require local HEAD, headRefOid, and checks_head_sha to match; confirm mergeable is not CONFLICTING and mergeStateStatus is not DIRTY; scripts/pr-state --summary does not include mergeability.

If a remediation changes rendered UI, invalidate screenshots captured before fixup and recapture and re-publish every affected viewport after the final commit. Never leave pre-fixup screenshots in the PR.

Immediately before a remediation commit or push—and again after long-running remediation—refresh PR state. Require the PR to remain open and its head ref to match the local branch. Before a push, compare the remote head OID with the local upstream tip; after the push, require the PR head OID to equal local HEAD. If the PR merged or closed, do not recreate its deleted branch with a stale push: preserve the local fix and ask before creating a clean follow-up.

After any rebase or force-push, fetch the PR base and compare local HEAD, the upstream tip, and pr.head_ref_oid; rerun affected checks. A merge-commit head requires --rebase-merges or a verified merge-only delta; after long hooks/tests, compare the latest authoritative base with the rebase base and reconcile if changed; otherwise a rebase invalidates prior evidence: rerun affected checks, scripts/pr-resolve list <PR>, and scripts/pr-state --summary <PR>; use --force-with-lease, never an unconditional force-push. If the rebase or conflict resolution touched AGENTS.md, CLAUDE.md, or a skill/reference file, run the shared harness validation in .agents/skills/harness-improvement/references/validation.md before pushing.

5. Re-check

After every push, re-fetch current-head state and run scripts/pr-resolve list <PR> (and show for any thread you may answer), then scripts/pr-state --summary <PR> again for the new head. For explicit consolidation, verify the target head before commenting/closing the superseded PR; preserve its branch unless deletion is requested. Automated reviewers may resolve or replace threads; do not reply to or resolve a thread that fresh state reports as resolved unless an authorized later code/policy change invalidates the rationale for that resolution. In that exception, re-fetch/show the thread, explicitly reopen it with scripts/pr-resolve reopen <PR> <thread_id>, then use the reply helper to reply/resolve/react and verify the final thread. Before an ordinary reply to a pre-push thread, run scripts/pr-resolve show <PR> <thread_id>; if it reports resolved: true (often with an Addressed in commit ... marker), record the thread as auto-resolved and do not post a duplicate reply. Continue replying/resolving only for resolved: false threads, including hidden unresolved threads. If a finding cites an obsolete commit, inspect the cited lines at the current PR head before editing. Apply only the missing portion, and do not duplicate an assertion that is already present or reply to a resolved or stale thread. Treat each fresh summary as a new review-evidence snapshot: inspect every non-empty body in review_evidence.exact_current_head_reviews[], even when unresolved_review_thread_count=0 and scripts/pr-resolve list is empty. Classify current-head review bodies and top-level bot/issue comments before declaring the PR clean; empty thread and issue-comment counts are insufficient. Treat a non-empty hidden_unresolved_threads value in that fresh snapshot as a mandatory hidden-thread gate: expand and disposition each hidden thread, then run scripts/pr-resolve list <PR> again after the refresh and immediately before reporting. Require checks_head_sha to match that head, report pending checks separately from failures, and rerun scripts/pr-resolve list <PR> before declaring the PR clean. The final predicate must also require hidden_unresolved_threads=[]; do not require filtered and unresolved counts to be equal because the filtered count includes resolved threads. Treat prior review evidence as stale. When the user authorized thread writes, every unresolved thread still needs an explicit reply and resolution matching its disposition. A duplicate or stale bot thread needs the same treatment once current source proves the finding is already fixed, including a thread surfaced only in hidden_unresolved_threads; only current-head actionable threads drive code changes. Declare the PR clean only when the exact-current-head review classification reports no unaddressed findings, checks_snapshot_complete=true, failed_checks=[], pending_checks=[], approval_required_runs=[], actionable_issue_comment_count=0, unresolved_review_thread_count=0, hidden_unresolved_threads=[], there is no merge conflict, and scripts/pr-resolve list <PR> is empty. Require either base_advanced_since_head=false or recorded merge-result validation for the current head and current base, as defined in references/merge-conflicts.md. An unknown base relationship is not sufficient. Within the user's monitoring limit, continue checking after resolutions until automated review jobs are terminal; otherwise report the exact pending check names.

When remediation changes tests or validation, reconcile any validation commands or counts claimed in the live PR description with the final verification before declaring fixup complete. Reuse /pr's live-body preservation and REST-fallback procedure, preserve intervening bot or maintainer text, and re-fetch exact-head state afterward; a body PATCH triggers workflows, so restart pr-state/pr-await before treating pre-PATCH CI as current.

If the user explicitly requested a persistent Kandev plan update and the task has an external Kandev plan, call get_task_plan_kandev before fixup and update_task_plan_kandev after fixup with the remediation commit, final exact-head check counts, resolved-thread state, and mergeability. Without that authorization, report the plan update as pending and do not invoke Kandev task or session APIs. Batch plan/task synchronization into the final documentation commit. Record the prior head's fixup evidence before a plan commit/push: that push restarts CI and invalidates the snapshot. For tracked docs/plans/** artifacts, keep prose head-agnostic and record remediation scope/local verification before that commit; then rerun scripts/pr-state --summary and scripts/pr-resolve list for the new head and report pending checks separately; keep exact-head verification/queue work externally pending, not complete, until pending checks are empty. Mark prior current-head claims historical/superseded when a new head replaces them; report only the latest head's SHA, CI/review counts, and mergeability; do not leave planned verification marked unstarted after it has run.

Before declaring fixup complete, verify git status --short is clean, git rev-parse HEAD equals git rev-parse @{upstream}, the PR head equals local HEAD, and the fresh mergeability state is not conflicting. Do not call the PR clean from CI/review counts alone when the worktree or remote tip still differs.

The phrase "ready to merge" is reserved for a fresh current-head snapshot with every required check successful or explicitly skipped, no pending or failed leaf or aggregate check, no unresolved review thread, no merge conflict, and the local, upstream, and PR head OIDs aligned. A clean local reproduction does not waive a failed remote check; push the remediation and re-check the new head first.

6. User-Requested Merge

Merge only after the user explicitly asks and the current-head state is clean. From a linked worktree, run gh pr merge <PR> --squash without --delete-branch: that flag can attempt a local checkout of the base branch and fail when another worktree owns it, even after the remote merge succeeds. Report the remote merge separately. Delete a remote or local branch only when requested and through a worktree-safe cleanup flow.

When branch protection requires a merge queue, after the exact current head is clean and all checks pass, run gh pr merge <PR> --auto --match-head-commit <SHA> without a merge strategy. In this mode the command requests queue admission, not a direct merge. Verify the added_to_merge_queue timeline event and pullRequest.mergeQueueEntry { id state position estimatedTimeToMerge headCommit { oid } } with GraphQL; treat "already queued" as verification, not a retry. Compare the expected SHA with pullRequest.headRefOid, not the queue commit; keep the PR open while queue checks run. Do not rely on a null autoMergeRequest or mergeStateStatus=UNKNOWN to determine queue status.

For "monitor until merge", poll queue/PR state with bounded cadence; require MERGED, mergedAt, mergeCommit.oid, and isInMergeQueue=false. QUEUED/AWAITING_CHECKS and queued CLEAN remain pending; inspect references/merge-queue.md for synthetic checks before reporting a blocker.

Guardrails

• Do not create Kandev subtasks unless the user explicitly asks for task tracking.
• Do not use native delegation or a full-history context fork to poll CI.
• Do not push, post comments, or resolve threads when the user asked for review only.
• Do not silently dismiss an invalid finding; every invalid thread needs a concrete pushback disposition, and an authorized complete cleanup must post that explanation before resolving it.
• Do not proceed with an unverified PR when mandatory verification is blocked.