- Home
- Skills
- Code Quality & Review
- PR Review Crew
PR Review Crew
A staff-level PR reviewer that audits diffs for security, performance, and logic errors with evidence-backed reports.
$12
Works with the AI Agents you already use
PR Review Crew
Example session with this skill installed
Review the staged changes in this repo and check for any critical security risks or performance bottlenecks.
- Read your context and instructions
- Compiled the pr review crew
- Generated the document
I have completed a Staff Engineer review of the staged changes across the security and performance lenses. Stack pack detected: Next.js/React.
PR Review Report
Verdict: REQUEST CHANGES
One HIGH severity security issue and one MEDIUM performance bottleneck identified.
Scope Drift
Intent: Staged changes in local repository
Delivered: Auth middleware updates and new dashboard component
IN SCOPE: ✅ Auth hardening
DRIFT: ⚠️ Added 'DashboardStats' component (Not mentioned in staged intent)
Critical & Specialist Findings
| Severity | Category | Finding | Evidence | Confidence |
|---|---|---|---|---|
| HIGH | Security | Missing CSRF protection on new POST endpoint | app/api/profile/route.ts:14 | 9/10 |
| MEDIUM | Performance | Unnecessary re-renders; unstable object literal in dependency array | components/Dashboard.tsx:42 | 8/10 |
| LOW | Maintainability | TODO comment without owner or ticket reference | lib/utils/auth.ts:89 | 10/10 |
Adversarial Pass
How to break this: The profile/route.ts endpoint accepts JSON without validating content-type headers, potentially allowing cross-site request forgery if the session cookie is not SameSite=Strict. (Status: FIXABLE)
Next steps
- Apply the missing
CSRFtoken validation to the profile API route. - Memoize the
chartOptionsobject inDashboard.tsxto prevent unnecessary sub-component updates. - Run
npm auditto verify the three new dependencies added topackage.json.
Dashboard.tsx
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Reviewing pull requests is often a choice between a shallow skim that misses bugs or a deep dive that stalls the sprint. This skill acts as a Staff Engineer in your terminal, performing a multi-lens audit of git diffs to catch production-breaking defects before they merge.
What it does
- Multi-lens auditing performs focused passes for security (OWASP), performance (N+1 queries), testing coverage, and maintainability.
- Stack-specific analysis loads specialized rule packs for Next.js, Python, and Go to detect framework-level anti-patterns.
- Automated safety checks identifies critical risks like SQL injection, race conditions, IDOR, and unvalidated LLM output boundaries.
- Intent verification compares the diff against PR descriptions or tickets to detect scope drift or missing requirements.
- Evidence-based reporting assigns a confidence score to every finding and links it to specific
file:linelocations usinggrepandrg.
How it works
- Targeting identifies the review scope by checking a PR ID, staged changes, or a branch diff against the merge base.
- Detection identifies the repository stack and loads the corresponding rule pack and local audit tools.
- Execution runs the critical safety pass followed by five specialist lenses and an adversarial pass to try and break the implementation.
- Reporting generates a structured report with findings categorized by severity and confidence.
Frameworks & tools
Works with any git-based workflow. Leverages gh CLI, rg, and stack-specific auditors like npm audit, pip-audit, and govulncheck. Includes specialized rules for Next.js, React, Python, and Go.
Why this beats prompting it yourself
Generic prompts often produce style nits and vague "best practice" advice. This skill uses a systematic staff-level rubric, executes real CLI tools to verify dependencies, and demands file:line evidence for every claim, ensuring findings are technical rather than subjective.
Use cases
- Catching SQL injections and race conditions in a complex backend PR.
- Auditing new dependencies for known CVEs before they enter the lockfile.
- Ensuring a frontend refactor doesn't introduce N+1 component re-renders.
- Verifying that all acceptance criteria from a spec are actually implemented in the diff.
Known limitations
Requires a git environment. Automated dependency auditing requires the respective package managers and audit tools to be installed in the environment.
Requires git environment. Automated dependency audits only run if tools like npm-audit or pip-audit are present. Findings are scoped strictly to the provided diff.
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 1 month ago
Requires
Permissions
Allowed Hosts