What does this skill do, and when should you use it?
This skill helps an agent locate the open GitHub Pull Request associated with the current branch and inspect comments that need attention. It first verifies GitHub CLI authentication, then runs the repository’s comment-fetching script. The agent numbers review threads and comments, summarizes the work required, and asks the user which items to address. Fixes are applied only after the user selects the items.
Runs gh auth status to verify GitHub CLI authentication; runs scripts/fetch_comments.py to print comments and review threads on the PR; numbers the comments and threads and summarizes the required work; asks the user which numbered items to address; applies fixes for the selected items. gh commands require network access and may need workflow/repo scopes.
- A developer needs to review outstanding feedback on the open PR for the current branch.
- A maintainer wants to choose specific review comments before any code changes are made.
- A contributor needs one view of both regular comments and review threads.
- A GitHub CLI session encounters authentication or rate-limit issues and needs reauthentication guidance.
How do you install this skill?
- Before use, verify gh authentication, granted scopes, and the PR associated with the current repository and branch to avoid accessing or exposing comments in the wrong GitHub context.
- The script does not define an actionable recovery path for no-PR cases, insufficient permissions, GraphQL schema changes, or pagination failures.
- Before applying fixes, confirm the files to change, run appropriate tests, and preserve a recoverable version-control state; the repository is marked deprecated, so maintenance status should be verified.
- Shell / CLI
- Network access
- Local filesystem
GitHub CLI
In Codex, use the built-in $skill-installer and run: $skill-installer gh-address-comments. The skill is located at skills/.curated/gh-address-comments in openai/skills. Restart Codex after installation so it can load the new skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/openai/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/.curated/gh-address-comments" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Inspect the comments on the open PR for the current branch and list the items I can choose to address.
Authenticate GitHub CLI first, for example with gh auth login. Then ask the agent: “Inspect the comments on the open PR for the current branch and list the items I can choose to address.” The agent verifies authentication, fetches comments and threads, and waits for your numbered selection. The source does not document a more specific trigger syntax.
What are this skill's strengths and limitations?
- Covers both regular comments and review threads, with user selection before changes.
- Uses GitHub CLI and the repository’s scripts/fetch_comments.py, giving the workflow a defined inspection step.
- Includes a documented recovery path for authentication and rate-limit problems.
- Requires GitHub CLI, valid authentication, appropriate workflow/repo scopes, and network access.
- The source does not specify the minimum required permissions.
- No test suite, supported-platform list, or implementation details for the fetching script are provided.
- The documented scope is the open PR for the current branch; support for other hosting platforms is not established.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| GitHub PR Comment Handler this page ✓ OpenAI · Official | 42 · Not recommended | ★ 28k | 4mo ago | — |
| PR Babysitter ✓ OpenAI · Official | 57 · Use with care | ★ 129k | 11d ago | Apache-2.0 |
| Ship a PR (e2e repo PR delivery workflow) | 59 · Recommended | ★ 10k | 11d ago | Apache-2.0 |
| CODEOWNERS PR Splitter ✓ NVIDIA · Official | 57 · Use with care | ★ 3.6k | 11d ago | Apache-2.0 |
| RenderCV PR Reviewer | 55 · Use with care | ★ 18k | 7mo ago | MIT |
How did FollowSkills review this skill?
The skill requires an authenticated gh CLI and may request workflow/repo scopes plus network access; it waits for the user to select comments before applying fixes, providing some consent boundary. Deductions apply because permissions are relatively broad, data flow and external GitHub effects are under-described, rollback is absent, sensitive-data handling is unspecified, and there is no explicit pre-change confirmation. No malicious or covert exfiltration behavior is shown.
The script includes authentication checks, GraphQL error handling, JSON parsing, and pagination, so the main path is plausible. Static evidence cannot confirm GraphQL field compatibility, relative-path reachability, behavior with no or non-open PRs, permission failures, or network errors; the fix phase lacks defined failure feedback and validation. The score remains below the static ceiling of 10.
The scenario and core inputs are clear: process comments on the GitHub PR associated with the current branch, number and summarize them, then ask which to address. Deductions apply because non-fit cases such as no PR, multiple PRs, non-GitHub hosting, or read-only requests are not defined, and Chinese-language interaction is unspecified. The core function depends on GitHub and gh CLI, creating mainland-China network reachability risk.
The SKILL.md has a concise structure with prerequisites and a basic workflow; script purpose and usage are reasonably inferable from the repository context. Deductions apply for unknown skill-level licensing, missing versioning, changelog, maintenance owner, update path, parameter documentation, example output, and troubleshooting. The repository README also marks the repository deprecated, increasing governance uncertainty.
The script can aggregate conversation comments, reviews, and inline threads, and the workflow has clear user-facing utility. Deductions apply because static reading cannot verify completeness or direct usability, while fix strategy, edit boundaries, tests, and final-summary format are unspecified. The result still requires user judgment and review, so the score stays within the static ceiling.
The source, GraphQL query, and error handling are auditable, and the command path is potentially reproducible, supporting limited verification. Deductions apply because there is no committed test suite, CI coverage, third-party execution evidence, or sample result; the assessment is based on static reading only and cannot reach a higher evidence tier.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →