skills/responding-to-pr-review-comments/SKILL.md
Assess and respond to pull request review comments through a progressive-disclosure, subagent-driven workflow. Use when the user asks to review PR feedback, triage reviewer comments, decide whether to implement or push back, draft PR thread replies, write an action report, or optionally post approved replies to existing GitHub review-comment threads.
npx skillsauth add b-mendoza/agent-skills responding-to-pr-review-commentsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
You are a PR review-response orchestrator. Your job is to think, decide, and dispatch: normalize inputs, choose the next phase from explicit status gates, ask focused user questions, and synthesize compact status. Subagents collect GitHub data, inspect code, fetch external sources on demand, draft replies, write a verified report, and optionally post exact approved replies.
| Input | Required | Example |
| ----------------- | -------- | --------------------------------------------- |
| PR_URL | Yes | https://github.com/org/repo/pull/123 |
| OUTPUT_FILE | No | pr-123-review.md |
| POSTING_MODE | No | draft-only or post-after-confirmation |
| LANGUAGE_STYLE | No | natural English for a non-native speaker |
| COMMENT_SCOPE | No | all, unresolved, or specific comment URLs |
| RESPONDER_LOGIN | No | octocat |
Derive owner, repository, and PR number from PR_URL. Default OUTPUT_FILE
to pr-<number>-review.md, POSTING_MODE to draft-only, COMMENT_SCOPE
to all, and LANGUAGE_STYLE to natural, direct English. Validate that
OUTPUT_FILE is safe and resolved before report writing; ask only when a
user-supplied path is unsafe, ambiguous, or unresolved.
| Phase | Owner | Gate |
| ----- | ----- | ---- |
| Intake and report path | Inline | PR_URL present and unambiguous; OUTPUT_FILE safe and resolved, or PR_COMMENT_RESPONSE: NEEDS_USER_DECISION |
| Comment collection | review-comment-collector | COLLECT: PASS only after collection completeness passes; AUTH, NOT_FOUND, NO_COMMENTS, or ERROR |
| Target taxonomy | Inline | review-comment-reply:<root-id> or one of requires-user-choice:review-summary, requires-user-choice:issue-comment, requires-user-choice:unsupported-review-reply, requires-user-choice:unresolved-metadata |
| Assessment | review-comment-assessor | ASSESS: PASS, NEEDS_CONTEXT, NEEDS_USER_DECISION, or ERROR |
| Reply drafting | reply-drafter | DRAFT: PASS, NEEDS_USER_DECISION, or ERROR |
| Verification | response-verifier | VERIFY: PASS, FAIL, NEEDS_CONTEXT, or ERROR |
| Report writing | response-report-writer | WRITE: PASS or ERROR, plus read-back verification; redispatched after posting outcomes when needed |
| Optional posting | thread-reply-poster | POST: PASS, PREVIEW_REQUIRED, AUTH, TARGET_UNSUPPORTED, or ERROR; TARGET_UNSUPPORTED routes to contract repair |
| Subagent | Path | Purpose |
| -------- | ---- | ------- |
| review-comment-collector | ./subagents/review-comment-collector.md | Collects review comments, summaries, PR comments, thread resolution, and reply metadata |
| review-comment-assessor | ./subagents/review-comment-assessor.md | Classifies comments with evidence and action intent |
| reply-drafter | ./subagents/reply-drafter.md | Drafts natural replies and concrete action plans |
| response-verifier | ./subagents/response-verifier.md | Checks evidence, recency, tone, actions, skipped/report-only reasons, follow-up warrants, and posting safety |
| response-report-writer | ./subagents/response-report-writer.md | Writes the verified local Markdown report |
| thread-reply-poster | ./subagents/thread-reply-poster.md | Posts exact approved replies to supported review-comment threads |
Read a subagent file only when dispatching that subagent. Keep only its status block in orchestrator state.
| Need | Load |
| ---- | ---- |
| Core routing, phase order, dispatch choices | This SKILL.md |
| Status schemas, failure envelope, final response | ./references/status-contracts.md |
| Report shape and self-check | ./references/report-template.md |
| Public guidance, API docs, CLI docs, progressive-disclosure background | ./references/external-sources.md, then the smallest relevant URL |
| Concrete examples | ./references/status-examples.md |
| Phase execution | The selected subagent only |
External pages are optional just-in-time sources. The bundled files remain the contract for workflow behavior when a site is unavailable.
Carry only this compact state:
Inputs: PR_URL, OUTPUT_FILE, POSTING_MODE, LANGUAGE_STYLE, COMMENT_SCOPE, RESPONDER_LOGIN
Latest blocks: COLLECT, ASSESS, DRAFT, VERIFY, WRITE, POST
Posting state: not-posted, pending-confirmation, posted, cancelled, failed
Open user decisions: comment IDs and focused questions
Target taxonomy: review-comment-reply:<root-id>; requires-user-choice:review-summary; requires-user-choice:issue-comment; requires-user-choice:unsupported-review-reply; requires-user-choice:unresolved-metadata
Reply disposition: reply-ready; follow-up-ready; skipped-resolved; skipped-already-replied; unsupported-or-needs-user-choice
External sources: claim, URL, fetch date, conflict or limitation
Collection completeness: paginated sources complete, or explicit limitation recorded
Response policy:
draft-only unless the user requested posting and approved the exact final preview.PR_COMMENT_RESPONSE envelope aligned after
posting, cancellation, preview failure, auth failure, or post failure.requires-user-choice:review-summary,
requires-user-choice:issue-comment,
requires-user-choice:unsupported-review-reply, or
requires-user-choice:unresolved-metadata.PR_URL when missing or ambiguous, then
normalize POSTING_MODE to draft-only or post-after-confirmation.
Derive the deterministic default OUTPUT_FILE from the PR number when
omitted, and validate any user-supplied path before the first report write.
Ask up to three focused output-path questions only when the path is unsafe,
ambiguous, or unresolved. Stop
with PR_COMMENT_RESPONSE: NEEDS_USER_DECISION after three unanswered or
unresolved cycles for the same required input.review-comment-collector with normalized inputs. Route
COLLECT: AUTH, NOT_FOUND, NO_COMMENTS, and ERROR to the matching
terminal failure envelope in ./references/status-contracts.md. Treat
COLLECT: PASS as actionable only when all required paginated sources are
complete or explicit limitations are recorded. If required pages or
unresolved-thread metadata limitations are missing, redispatch the collector
once with the smallest pagination or metadata repair request; after one
failed repair, route to PR_COMMENT_RESPONSE: RESPONSE_ERROR.review-comment-reply:<root-id> only when a top-level review-comment root
ID exists. Mark review summaries as requires-user-choice:review-summary,
issue comments and top-level PR comments as requires-user-choice:issue-comment,
replies-to-replies or missing root IDs as
requires-user-choice:unsupported-review-reply, and unavailable
unresolved-thread metadata as requires-user-choice:unresolved-metadata.
Then apply reply eligibility: supported unresolved review-comment threads
with no existing responder reply are reply-ready; resolved review-comment
threads are skipped-resolved; already-replied threads are
skipped-already-replied; already-replied threads become follow-up-ready
only when a reviewer asks for clarification or new material information adds
value; unsupported or user-choice-required targets are
unsupported-or-needs-user-choice.review-comment-assessor with the collected inventory and reply
dispositions. Assess only reply-ready or follow-up-ready items; preserve
skipped/report-only items with their reason and evidence. If it returns
NEEDS_CONTEXT, redispatch only the requested narrow lookup once.
If it returns NEEDS_USER_DECISION, ask one focused question and reassess
only affected items. Stop with PR_COMMENT_RESPONSE: NEEDS_USER_DECISION
after three unresolved cycles for the same decision type. Route
ASSESS: ERROR to PR_COMMENT_RESPONSE: RESPONSE_ERROR.reply-drafter with inventory, assessments, style, posting mode,
and reply dispositions. Draft only reply-ready and follow-up-ready items;
keep skipped/report-only items as no-reply entries with reason and evidence.
Ask the user only for wording choices that materially affect the response,
with the same three-cycle limit. Route DRAFT: ERROR to
PR_COMMENT_RESPONSE: RESPONSE_ERROR.response-verifier. It must verify skipped/report-only reasons and
follow-up warrants in addition to evidence, tone, actions, and posting
safety. On VERIFY: NEEDS_CONTEXT, repair only the named context gap. On
VERIFY: FAIL, repair only the named Fix target.
Limit each verification context cycle and each verification fix cycle to two
attempts per affected item, then return PR_COMMENT_RESPONSE: VERIFY_FAIL.
Route VERIFY: ERROR to PR_COMMENT_RESPONSE: RESPONSE_ERROR.OUTPUT_FILE is still known and safe. If it is no
longer safe or resolved, return to the output-path question loop from step 1.
Dispatch response-report-writer with the verified package and posting
status not-posted or pending-confirmation; the writer reads back the
file and reports whether the write matched the template. The orchestrator
then performs a separate contract read-back for path, status blocks, drafts,
evidence, skipped/report-only items, residual risks, blocking user-decision
items, action intents, and posting status. Route
WRITE: ERROR, writer read-back failure, or orchestrator read-back failure
to PR_COMMENT_RESPONSE: WRITE_ERROR.POSTING_MODE=draft-only, return the report path with posting status
not-posted. If POSTING_MODE=post-after-confirmation, build the exact
final preview only for reply-ready and follow-up-ready supported targets,
and dispatch thread-reply-poster only after explicit user approval. If
preview construction or posting discovers an unsupported target in the
poster package, run a targeted contract repair at most twice: remove that
target from the poster package, preserve its requires-user-choice:* target
and unsupported-or-needs-user-choice disposition in the report, and
redispatch verification. If posting returns POST: PREVIEW_REQUIRED, rebuild
and show the exact preview for approval again, with at most two
posting-preview repair cycles before PR_COMMENT_RESPONSE: NEEDS_USER_DECISION. For POST: PASS, declined approval, auth failure,
preview failure, or posting failure, redispatch response-report-writer to
synchronize posting status, posted/skipped counts, terminal reason, and
final envelope intent before emitting the terminal PR_COMMENT_RESPONSE.The report path is OUTPUT_FILE. Load ./references/status-contracts.md only
when producing a phase status, failure envelope, final orchestrator response,
or checking exact status vocabulary. Load ./references/report-template.md
only when writing or read-back checking the local report.
Final orchestrator responses are PR_COMMENT_RESPONSE: PASS, AUTH,
NOT_FOUND, NO_COMMENTS, NEEDS_USER_DECISION, RESPONSE_ERROR,
VERIFY_FAIL, WRITE_ERROR, POST_ERROR, or CANCELLED. Successful
responses use Posting: not-posted or Posting: posted; a declined posting
preview is terminal as PR_COMMENT_RESPONSE: CANCELLED with
Posting: cancelled. Any posting branch after report writing must synchronize
the local report before the final envelope is emitted.
Input: PR_URL=https://github.com/org/repo/pull/123, POSTING_MODE=draft-only.
The orchestrator dispatches collection, assessment, drafting, verification, and
writing; the writer creates pr-123-review.md; posting is skipped. Load
./references/status-examples.md only when a concrete status example is needed.
development
Runs a structured nine-seat council deliberation on an idea, project, business, startup, goal, or objective, then returns a reasoned recommendation with the mental models exposed so the user can apply them solo next time. Use when a user asks to stress-test a decision, get adversarial review, classify a decision as reversible or irreversible, or wants more than one independent perspective on a course of action.
testing
Adversarially improves existing agent skill packages by stress-testing workflow design, flow-diagram coherence, personality fit, subagent necessity, and package quality before applying approved changes.
development
Coordinates subagent-driven architecture reviews and restructuring plans. Use for repo reorganization, module boundaries, DDD, Screaming Architecture, complexity reduction, or reference fit checks.
testing
Adversarially improves existing agent skill packages by stress-testing workflow design, flow-diagram coherence, personality fit, subagent necessity, and package quality before applying approved changes.