resolve-review-notes
Find the GitHub pull request for the current branch, collect and validate all review comments against current code, apply valid fixes, resolve addressed…
Skill by veverkap
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/veverkap-resolve-review-notes
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: productivity
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/resolve-review-notes
Tools it pre-approves 7
From allowed-tools in SKILL.md: the agent may use these without asking you each time.
ReadEditWriteGlobGrepBashAgent
What's inside
- Skill name
resolve-review-notes- Description
- Find the GitHub pull request for the current branch, collect and validate all review comments against current code, apply valid fixes, resolve addressed threads, and produce a completion summary.
- License
- MIT
- Allowed tools
- Read, Edit, Write, Glob, Grep, Bash, Agent
- Instructions
- 83 lines in SKILL.md · view on GitHub
Before installing
- Pre-approves any shell command
Files (1)
- SKILL.md
SKILL.md (first lines)
# Resolve Review Notes
## Mission
Resolve review feedback on the GitHub pull request for the current branch end-to-end.
## Prerequisites
- Run inside a Git repository with a non-detached current branch.
- GitHub CLI (`gh`) must be installed and authenticated for the repository host.
- The current branch must have an open pull request unless the user explicitly identifies another PR.
## Feedback Scope
Collect all feedback, not only unresolved inline threads:
- inline pull request review threads, including replies, outdated comments, and resolution state
- submitted reviews and their top-level review bodies
- pull request conversation comments
Treat a thread and its replies as one feedback item. Ignore only empty review bodies and purely administrative events; record bot feedback and human feedback alike.
## Workflow
1. Discover the pull request.
- Verify authentication with `gh auth status`.
- Read the current branch with `git branch --show-current`; stop on a detached HEAD.
- Resolve the PR with `gh pr view --json number,url,title,state,headRefName,baseRefName,reviewDecision`.
- Confirm the returned `headRefName` matches the current branch. If no open PR exists, report that clearly and stop; do not guess another PR.
2. Collect the complete feedback inventory.
- Use `gh api --paginate` for REST collections and `gh api graphql` with cursor pagination for review threads.
- Fetch review threads with thread IDs, `isResolved`, `isOutdated`, file path, line/original line, and every comment's ID, author, body, URL, and timestamps.
- Fetch submitted reviews from `repos/{owner}/{repo}/pulls/{number}/reviews`.
- Fetch conversation comments from `repos/{owner}/{repo}/issues/{number}/comments`.
- Preserve comment URLs and IDs so every outcome is auditable.
- Deduplicate review bodies or replies already represented in a review thread. Build an inventory before making edits so pagination gaps cannot silently omit later comments.
3. Validate every feedback item against the checked-out branch.
- Read the current code and relevant tests at the referenced location; for outdated comments, locate the corresponding current code by symbol or nearby context.
- Determine the reviewer's concrete claim or requested change and test it against current behavior, repository conventions, and available checks.
- Classify each item as:
- `valid`: the issue still exists and the requested outcome is appropriate
- `already-addressed`: current code already satisfies the comment
- `not-applicable`: the claim does not hold for the current code, with evidence
- `ambiguous`: intent or desired behavior requires a human decision
- Record a concise rationale and evidence for every classification. Do not treat resolved or outdated status as proof that a comment is addressed.
4. Fix valid items.
- Apply minimal edits.
- Group related fixes.
- Re-run relevant checks (lint/test/build where available).
5. Update GitHub review threads.
- Re-fetch thread state before mutating it to account for comments added during the run.
- For fixed or already-addressed inline threads, post a concise evidence-based reply when useful, then resolve the thread with GitHub's `resolveReviewThread` GraphQL mutation.
- Leave `not-applicable` and `ambiguous` threads unresolved unless the user explicitly asks otherwise; report the rationale or blocker.
- Do not resolve a thread when checks for its fix are failing.
- Do not edit, delete, minimize, or dismiss reviewers' comments or reviews.
6. Produce summary.
- Write `reports/resolve-review-notes-summary.md` with:
- PR number and URL
- total processed
- fixed
- already addressed
- not applicable
- ambiguous/pending
- comment URL and validation rationale for every item
- files changed
- checks executed and results
- threads resolved and threads left open
## Guardrails
- Do not fabricate reviewer intent when ambiguous.
- Do not make speculative logic changes without evidence.
- Do not skip comments because they are old, resolved, generated by a bot, or attached to outdated code.
- Do not expose authentication tokens or private PR content outside the repository and GitHub host.
- Do not approve, dismiss, merge, close, or submit the pull request.
- Do not push changes unless explicitly requested.
- Prefer one coherent commit plan, but do not commit unless explicitly requested.From veverkap/skills, MIT.
More from veverkap/skills
- 0.524score
duplicate-code-detector
veverkap
Detect significant duplicate code patterns in recently changed files using semantic analysis and open focused remediation issues (one per pattern, up to 3).
SkillBFreshness: BClaude Code +9 - 0.521score
ci-coach
veverkap
Analyze local CI workflow definitions, identify low-risk efficiency improvements, implement focused updates, and generate a data-backed optimization report.
SkillBFreshness: BClaude Code +9 - 0.521score
code-simplifier
veverkap
Analyze recently modified local code and simplify it for clarity and maintainability while preserving behavior.
SkillBFreshness: BClaude Code +9 - 0.521score
grumpy-reviewer
veverkap
Perform a critical, grumpy senior-style review of local diffs, focusing on correctness, security, performance, and maintainability.
SkillBFreshness: BClaude Code +9 - 0.521score
nitpick-reviewer
veverkap
Perform a detailed local diff nitpick review focused on subtle naming, structure, and consistency issues that linters often miss.
SkillBFreshness: BClaude Code +9 - 0.521score
principal-engineer-review
veverkap
Review local diffs and design changes from a principal engineer standpoint: MVP scope, correctness, maintainability, security, reliability, and reuse of…
SkillBFreshness: BClaude Code +9 - 0.521score
change-fix
veverkap
Diagnose and fix local branch failures or requested issues, verify with project checks, and produce a clear fix summary.
SkillBFreshness: BClaude Code +9 - 0.496score
improve-codebase-architecture
veverkap
Scan a codebase for deepening opportunities, present them as a visual HTML report, then drill into the selected option with concrete interface and migration…
SkillBFreshness: BClaude Code +9 - 0.281score
babysit-changes
veverkap
No description from the publisher yet.
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Aug 8, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/veverkap/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from veverkap
- 0.524score
duplicate-code-detector
veverkap
Detect significant duplicate code patterns in recently changed files using semantic analysis and open focused remediation issues (one per pattern, up to 3).
SkillBFreshness: BClaude Code +9 - 0.521score
change-fix
veverkap
Diagnose and fix local branch failures or requested issues, verify with project checks, and produce a clear fix summary.
SkillBFreshness: BClaude Code +9 - 0.521score
ci-coach
veverkap
Analyze local CI workflow definitions, identify low-risk efficiency improvements, implement focused updates, and generate a data-backed optimization report.
SkillBFreshness: BClaude Code +9 - 0.521score
code-simplifier
veverkap
Analyze recently modified local code and simplify it for clarity and maintainability while preserving behavior.
SkillBFreshness: BClaude Code +9 - 0.521score
grumpy-reviewer
veverkap
Perform a critical, grumpy senior-style review of local diffs, focusing on correctness, security, performance, and maintainability.
SkillBFreshness: BClaude Code +9 - 0.521score
nitpick-reviewer
veverkap
Perform a detailed local diff nitpick review focused on subtle naming, structure, and consistency issues that linters often miss.
SkillBFreshness: BClaude Code +9
Cite this page
resolve-review-notes skill. Agent Mart. https://agentmart.dev/skills/veverkap-resolve-review-notes (last updated 2026-10-24)