- Home
- Skills
- Code Quality & Review
- github pr comments
Works with the AI Agents you already use
github pr comments
Fetch, triage, and address GitHub pull request review comments with automated classification and local code fixes.
$27
github pr comments
Example session with this skill installed
Get the PR comments for the current branch and tell me what I need to fix.
- Read your context and instructions
- Compiled the github pr comments
I have fetched the review comments for the current pull request using the GitHub CLI. There are 5 comments across 3 threads in 2 files.
PR #12: Refactor auth middleware
3 review comments in 3 threads across 2 files.
Must fix
src/middleware/auth.ts:24(@octocat): Missing error handling for expired JWT tokens.src/middleware/auth.ts:45(@dev-lead): The permission check logic is inverted, allowing unauthorized access.
Suggestions
src/utils/logger.ts:12(@octocat): Use a structured logging format instead of plain strings.
Questions
src/middleware/auth.ts:30(@dev-lead): Why are we bypassing the cache here? — suggested answer: The cache is bypassed to ensure we always validate the latest session state from the DB.
Next steps
- Run "Address the must fix comments" to automatically apply the security fixes.
- Review the suggested answer for the question on line 30 before replying on GitHub.
- Note: The
get-github-pr-commentsMCP server is not configured; using GitHub CLI fallback.
Connects securely to your tools. The creator never sees your data.
About this skill
Reviewing pull requests manually is slow when feedback is scattered across dozens of threads. This skill automates the retrieval and classification of GitHub PR review comments into an actionable triage report.
What it does
- Automated retrieval fetches all inline review comments using the GitHub CLI or a dedicated MCP server.
- Thread grouping consolidates replies and discussions by file and line number to prevent duplicate tasks.
- Smart classification categorizes feedback into Must fix, Suggestion, Question, Nit, or Resolved based on the code context.
- Contextual analysis reads the current file state to determine if a comment has already been addressed.
- Code implementation applies requested changes to the local workspace when explicitly instructed.
How it works
- Identify PR detects the pull request for the current branch or uses a specific PR number provided by the user.
- Fetch comments pulls raw JSON data from the GitHub API via
ghorget_pr_comments. - Triage groups threads and assigns priority labels based on the reviewer's intent and the current code.
- Report generates a structured Markdown summary of all pending feedback.
- Address modifies local files to satisfy the review requirements if the user requests a fix.
Frameworks & tools
This skill requires the GitHub CLI (gh) authenticated on the local machine. It is optimized for use with the get-github-pr-comments MCP server but falls back to standard CLI commands.
Why this beats prompting it yourself
Standard prompts often struggle to distinguish between a "nit" and a "must fix" without seeing the surrounding code. This skill explicitly reads the referenced lines to verify if a comment is still relevant, saving you from re-fixing already resolved issues.
Use cases
- Summarizing feedback on a large PR with multiple reviewers.
- Identifying specific comments left on a single high-traffic file.
- Automatically applying "Must fix" changes to a local branch.
- Generating a status table of what was changed versus what needs manual input.
Known limitations
This skill is read-only on GitHub. It cannot post replies, resolve threads, or create new commits. It does not support general issue comments or top-level PR conversation.
Read-only on GitHub. Cannot post replies or resolve threads. Requires GitHub CLI (gh) installed and authenticated. Does not support general issue comments.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 12 days ago
Requires