github-pr-review
by fvadicamogithub-pr-review is a GitHub PR review skill for collecting inline comments, PR-level review bodies, and replies, then organizing feedback by severity so you can fix blockers first. Use it to resolve PR comments, respond to reviewers, and update the branch with targeted commits and thread replies. It fits the github-pr-review guide for authenticated GitHub CLI workflows.
This skill scores 84/100, which means it is a solid listing candidate for Agent Skills Finder. The repository gives users enough workflow detail to decide on installation: it clearly targets PR review comment resolution, explains when to use it, and shows a structured process for fetching comments, classifying severity, applying fixes, and replying to threads. Directory users should still note that it depends on GitHub CLI and would benefit from a more explicit install/run path, but it is strong enough to be useful and understandable.
- Strong triggerability: the frontmatter description names specific use cases like resolving PR comments, triaging feedback, verifying PR readiness, and responding to reviewers.
- Good operational depth: the SKILL.md includes a stepwise workflow for fetching inline comments and PR-level reviews, classifying severity, and applying fixes with user confirmation.
- Helpful supporting references: separate guides for CodeRabbit parsing and severity levels improve agent leverage when dealing with automated review comments.
- No install command or setup instructions are provided, so users must already know how to wire the skill into their agent workflow.
- The skill appears specialized to GitHub PR review handling and GitHub CLI usage, which limits fit for broader code-review tasks.
Overview of github-pr-review skill
github-pr-review is a GitHub PR review skill for turning scattered review comments into a prioritized fix plan, then helping you resolve them with less guesswork. It is best for agents or developers who need to review a pull request, understand what is blocking merge, and respond to reviewer feedback in a structured way. If your goal is github-pr-review for PR Review, this skill is more useful than a generic “summarize comments” prompt because it is built around comment fetching, severity grouping, and reply-oriented workflow.
What this skill is for
Use the github-pr-review skill when you need to:
- collect inline comments and PR-level review bodies from the current PR,
- separate original comments from replies,
- classify feedback by severity,
- decide what to fix first,
- and update the branch with targeted commits and thread responses.
What makes it different
The main differentiator is the workflow design: it does not just read review text, it cross-checks multiple review surfaces and treats severity as an action signal. That matters for PRs with automated reviewers like CodeRabbit or Gemini, where important feedback may be split across inline comments, review summaries, and grouped sections.
When it is a good fit
The github-pr-review guide is a good fit if you already have a PR open, have GitHub CLI available, and want an installation-oriented workflow that can inspect the live PR context. It is less useful if you only want a high-level code audit with no GitHub access, or if you need one-off prose feedback without actually touching the repository.
How to Use github-pr-review skill
Install github-pr-review
Install the github-pr-review skill with:
npx skills add fvadicamo/dev-agent-skills --skill github-pr-review
This install makes sense when the agent can run gh against the target repo and current branch. The workflow assumes authenticated GitHub CLI access and a PR that is already open or discoverable from the checked-out branch.
Give the skill the right input
Strong github-pr-review usage starts with a concrete prompt. Include:
- the PR link or confirm you are on the PR branch,
- what kind of feedback you want resolved,
- whether you want only blockers fixed or all comments addressed,
- and any constraints like “do not change public APIs” or “keep the patch minimal.”
Better input:
- “Use github-pr-review to inspect the current PR, prioritize all review comments by severity, fix only HIGH and CRITICAL items, and leave LOW comments as follow-up notes.”
- “Run github-pr-review against this branch and summarize which comments came from inline review vs PR-level automated review.”
Suggested workflow and files
Start with SKILL.md, then read references/coderabbit_parsing.md and references/severity_guide.md before acting on comments. Those two files explain the cases that most often block adoption: CodeRabbit’s grouped review format and severity labels that should change your fix order.
Practical workflow:
- Confirm the current PR with
gh pr view. - Pull inline comments and PR-level reviews separately.
- Remove replies so you only act on original review items.
- Map comments to severity and decide the merge-critical subset.
- Apply fixes, then respond in threads only after the code change is in place.
Prompting tips that improve output
For github-pr-review, “fix review comments” is too vague. Say what the agent should optimize for:
- “minimize diff size,”
- “preserve existing design,”
- “resolve only actionable comments,”
- “treat style nits as optional,”
- or “explain any comment you intentionally do not apply.”
That gives the skill enough context to choose between patching code, replying with rationale, or deferring lower-value suggestions.
github-pr-review skill FAQ
Does github-pr-review require GitHub CLI?
Yes, the github-pr-review skill is designed around gh commands, so it works best when GitHub CLI is installed and authenticated. If gh cannot access the repo or PR, the skill cannot reliably fetch review data.
Is github-pr-review only for automated review tools?
No. It handles human review comments too. The advantage over a normal prompt is that github-pr-review can distinguish inline comments, review bodies, replies, and severity tiers, which is especially helpful when automated and human feedback are mixed together.
Can beginners use the github-pr-review skill?
Yes, if they can identify the current branch and understand basic PR workflow. The skill reduces manual sorting, but you still need to review suggested fixes before applying them. It is not a push-button merge tool.
When should I not use it?
Do not rely on github-pr-review if you are not working inside a GitHub PR, if you lack CLI access, or if you only need a quick plain-language summary of code quality. In those cases, a simpler prompt or a general code review workflow may be enough.
How to Improve github-pr-review skill
Give the skill sharper review goals
The best github-pr-review results come from stating what “done” means. Say whether the priority is merge readiness, reviewer satisfaction, or minimal churn. If the PR is large, specify the files or issue types that matter most, such as correctness, security, or API compatibility.
Provide context the skill cannot infer
If a review comment depends on business rules, mention them up front. For example, tell the agent whether a suggested change is blocked by backward compatibility, performance constraints, or an intentional product decision. That prevents unnecessary back-and-forth on comments that should be explained rather than changed.
Watch for common failure modes
The usual mistakes are treating replies as new feedback, missing PR-level review bodies, and over-fixing low-priority nits before blockers. Another common issue is responding before the code is actually updated. The github-pr-review guide is strongest when comments are triaged first, then acted on in severity order.
Iterate after the first pass
After the first review pass, ask for a second pass that checks only unresolved items and newly introduced issues. If a fix changed behavior, ask the skill to verify that the original reviewer concern is now addressed and that no adjacent regression was introduced. This is the fastest way to improve github-pr-review usage without expanding the patch unnecessarily.