Skip to content
Development
Skill

/resolve-pr-feedback

Evaluate, fix, and answer GitHub PR comments, and fix what failing checks report. Use when the user asks to \"resolve PR feedback\", \"resolve PR comments\", \"answer review questions\", or \"fix PR checks\".

From plugin
turbo
42074 skills

Claude Code

$ npx -y skills add tobihagemann/turbo --skill resolve-pr-feedback --agent claude-code

SKILL.md

resolve-pr-feedback.SKILL.md
name: resolve-pr-feedback
description: "Evaluate, fix, and answer GitHub PR comments, and fix what failing checks report. Use when the user asks to \"resolve PR feedback\", \"resolve PR comments\", \"answer review questions\", or \"fix PR checks\"."

Resolve PR Feedback

Fetch unresolved review comments from a GitHub PR (inline threads, review-body observations, and issue-comment observations from the PR conversation), evaluate each one, fix or skip based on confidence, answer reviewer questions using the recalled rationale behind the change, and reply. Inline threads are answered with thread replies; issue-comment findings from people are answered with new PR conversation comments. Review-body findings, AI-reviewer issue comments, and the findings behind the head commit's failing checks flow through the same evaluate-and-fix pipeline; their outcomes land only in the summary.

Task Tracking

At the start, use `TaskCreate` to create a task for each step:

1. Fetch comments and checks 2. Triage already-addressed feedback 3. Run `/interpret-feedback` skill 4. Split questions and change requests 5. Run `/evaluate-findings` skill 6. Resolve ambiguities 7. Run `/implement-findings` skill 8. Verify fixes 9. Answer reviewer questions 10. Wait for AI re-reviews and checks 11. Run `/reply-to-pr-threads` skill 12. Run `/reply-to-pr-conversation` skill 13. Summary

Step 1: Fetch Comments and Checks

Auto-detect owner, repo, and PR number from current branch if not provided. Then run `${CLAUDE_SKILL_DIR}/scripts/fetch-pr-data.sh`, which handles full pagination (reviews, review threads, inner comment pages for long threads, issue comments, commits, the head commit's checks) and emits a single merged JSON document:

bash ${CLAUDE_SKILL_DIR}/scripts/fetch-pr-data.sh <owner> <repo> <pr_number>

Output shape:

{
  "meta":          { "title", "url", "headRefName", "headRefOid", "baseRefName", "body" },
  "reviewThreads": [ { "id", "isResolved", "isOutdated", "comments": { "nodes": [ { "author", "body", "path", "line", "originalLine", "diffHunk" } ] } } ],
  "reviews":       [ { "author", "body", "state", "submittedAt" } ],
  "issueComments": [ { "author", "body", "createdAt", "updatedAt", "url" } ],
  "commits":       [ { "commit": { "oid", "abbreviatedOid", "message", "committedDate" } } ],
  "checks":        [ { "name", "app", "state", "conclusion", "title", "detailsUrl" } ]
}

Filter review threads to unresolved only. Filter reviews to those with a non-empty body, excluding `PENDING` state (unsubmitted drafts). Filter issue comments to those with a non-empty body. Filter checks to those whose `state` is `failing`.

Read what each failing check reports, by its `app`:

  • **`github-actions`** — read the failed steps' log with `gh run view --job <job id> --log-failed`, taking the job id from the end of `detailsUrl`. Record each distinct build error or failing test as one check finding, with the file and line the log names. When the log cannot be read or names no failure, record the check as unreadable.
  • **`sonarqubecloud`** — run `bash ${CLAUDE_SKILL_DIR}/scripts/fetch-sonarcloud-findings.sh <detailsUrl>`, which emits `{ "qualityGate", "failedConditions": [ { "metric", "actual", "threshold", "comparator" } ], "issues": [ { "rule", "path", "line", "message", "severity", "type" } ] }`. Record each issue as one check finding, and each failed condition that no issue accounts for, such as coverage on new code, as one more. When the script exits non-zero, the analysis is not readable: record the check as unreadable.
  • **Any other app** — record the check as unreadable.

Tag each check finding with the source `check` and the name of its check. Check findings skip Steps 2 through 4 and feed Step 5. Record each unreadable check's name, title, and `detailsUrl` for the Step 13 summary.

Step 2: Triage Already-Addressed Feedback

Review bodies and PR conversation comments (issue comments) often pack multiple distinct concerns into one comment. Split each non-empty, non-PENDING review body and each non-empty issue comment into atomic observations, one per paragraph or bullet, so each can be evaluated on its own merits. When a review body nests observations inside collapsed `<details>` blocks, as AI reviewers do for nitpicks that name a file and line range, extract every observation inside them, keeping its file and line range in its text, and drop only blocks that carry no observation, such as review metadata, commit lists, and prompts for AI agents.

For every observation, check whether a subsequent commit already addresses it. Compare the source timestamp (`submittedAt` for review bodies, `createdAt` for issue comments) against each commit's `committedDate`; only commits after the source was posted can address it. Start with commit messages; read `git show <oid>` only when the message is ambiguous. A commit addresses an observation when its changes clearly resolve that specific concern. Touching the same area is not enough for an observation.

When an observation asserts a requirement or an intended behavior, compare that assertion against `meta.body`. The PR description is the author's own current statement of intent, so a description stating the opposite leaves a premise for Step 5 to settle. Record the contradiction against the observation and carry it forward, so Step 5 settles that premise once instead of re-deriving it for every finding the premise produced.

Classify each observation:

  • **Addressed**: A subsequent commit resolves it. Record the commit SHA for the Step 13 summary.
  • **Unaddressed**: No subsequent commit resolves it. Carry into Step 3, tagged with its `source` (`review-body` or `issue-comment`), the author, the observation text, any recorded contradiction with the PR description, and (for review bodies) the review state.

Review-body and issue-comment findings have no `diffHunk`, file path, or line reference unless their text names a file and line. The

Read more
@tobihagemann
Ships withturbo

A human-gated agentic software factory for Claude Code and Codex: you decide what to build and judge the result; skills shape, build, polish, ship, and learn.

Stars on GitHub

Momentum

stars in total
420
Star history of tobihagemann/turbo on GitHub, a chart by star-history.com

Other skills on turbo.

apply-turbo-handoffs

apply-turbo-handoffs

Audit the turbo lessons /self-improve wrote to .turbo/handoff/ and apply approved ones. Use when the user asks to \"read…

@tobihagemann@tobihagemannInstall Skill
apply-findings

apply-findings

Apply accepted review findings as code changes, escalating ambiguous ones to the user. Use when the user asks to \"apply…

@tobihagemann@tobihagemannInstall Skill
assess-technical-debt

assess-technical-debt

Assess project-wide technical debt (hotspots, deprecated APIs, duplication, low-value tests) into a ranked report. Use when the…

@tobihagemann@tobihagemannInstall Skill
backlog

backlog

Capture an out-of-scope improvement or accepted simplification in .turbo/backlog.md. Use proactively mid-task, or when the user…

@tobihagemann@tobihagemannInstall Skill
code-style

code-style

Enforce existence, reuse, mirror, and symmetry principles so new code stays minimal and consistent. Use when writing, changing,…

@tobihagemann@tobihagemannInstall Skill

More skills from other plugins.

brainstorming
Auto-invoked

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or…

@obra@obraInstall Skill
ask-matt

ask-matt

Ask which skill or flow fits your situation. A router over the skills in this repo.

skills2w309k
@mattpocock@mattpocockInstall Skill
accessibility
Auto-invoked

accessibility

Design, implement, and audit accessible UI to WCAG 2.2 Level AA across Web, iOS, and Android — semantic ARIA roles and labels,…

ecc2w293k
@affaan-m@affaan-mInstall Skill
compiler-commit

compiler-commit

Use when you want to verify compiler changes and commit with the correct convention. Runs tests, lint, and format, then commits…

react6d252k
@facebook@facebookInstall Skill

Questions about resolve-pr-feedback.

Which coding agents can use resolve-pr-feedback?
resolve-pr-feedback installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.