- Home
- Skills
- Code Quality & Review
- pr intelligence code review expert
pr intelligence code review expert
Senior-level code reviews anchored in diff analysis with automated verification of security, logic, and performance.
$6
Works with the AI Agents you already use
pr intelligence code review expert
Example session with this skill installed
Review the attached diff.patch for a Python backend service. Focus on security and database performance.
- Read your context and instructions
- Compiled the pr intelligence code
The review of diff.patch is complete. The PR introduces a new user search endpoint. Triage identified one high-risk security flaw and one medium-risk database performance issue. Both findings have been verified against the diff.
PR Review Findings
| File:Line | Title | Severity | Confidence | Scenario | Recommendation |
|---|---|---|---|---|---|
app/users.py:42 | SQL Injection Vulnerability | High | High | User input from search_term is interpolated directly into the query string, allowing arbitrary SQL execution. | Use parameterized queries with the execute() method. |
app/db.py:115 | N+1 Query Pattern | Medium | High | The loop fetches user profiles one-by-one inside a list comprehension instead of using a JOIN. | Refactor to use a single query with a WHERE IN clause or a JOIN. |
Verified Findings
- SQL Injection: Quote
f"SELECT * FROM users WHERE name = '{search_term}'"confirmed at line 42. - N+1 Query: Quote
[get_profile(u.id) for u in users]confirmed at line 115.
Questions for Author
- Why was the existing
QueryBuilderclass bypassed for the search endpoint? - Are there plans to add pagination to the search results to prevent memory exhaustion?
Release Readiness
- Database: Migration script is missing for the new
last_loginindex. - Security: The search endpoint lacks rate limiting.
Next steps
- Implement parameterized queries in
app/users.py. - Batch the profile lookups in
app/db.py. - Add a migration file for the
userstable index.
Connects securely to your tools. The creator never sees your data.
About this skill
Code reviews often suffer from two extremes: shallow nitpicking or false positives that erode developer trust. This skill enforces rigorous, diff-anchored analysis to identify high-impact defects in pull requests while ensuring every finding is physically present in the added code.
What it does
- Diff-anchored validation ensures every comment maps to a specific added line to eliminate ghost issues.
- Risk-based triage prioritizes reviews based on security, performance, and architectural impact.
- Multi-profile analysis routes code to specialized logic for security, databases, concurrency, and API contracts.
- Automated verification runs findings through a verifier script to check for quote accuracy and impact scenarios.
- Secret redaction automatically masks sensitive data like keys or PII found within the diff before reporting.
How it works
- Scope and Triage: Analyze the PR diff using
pr_diff.pyandpr_risk_rules.pyto identify high-risk zones. - Deep Inspection: Review the code through specialized profiles like
security-reviewerordatabase-reviewer. - Fact-Checking: Validate every finding with
review_verifier.pyto confirm the line exists and the quote is exact. - Calibration: Consolidate findings, adjust severity, and remove duplicates or low-confidence noise.
- Reporting: Deliver a structured review including impact scenarios, recommendations, and production readiness checks.
Frameworks & tools
This skill utilizes specialized Python scripts (pr_diff.py, review_verifier.py, pr_risk_rules.py) for deterministic analysis. It aligns with OWASP Top 10, ASVS, and Semantic Versioning standards across backend, frontend, and infrastructure code.
Why this beats prompting it yourself
Generic prompts often hallucinate bugs or comment on unchanged code. This skill uses a deterministic verification pipeline that rejects any finding not strictly present in the diff, maintaining a high signal-to-noise ratio.
Use cases
- Auditing security-critical PRs for injections, broken access control, or exposed secrets.
- Reviewing database migrations and N+1 query patterns in backend updates.
- Checking API contract compatibility and breaking changes in microservices.
- Verifying test coverage and assertion validity for mission-critical logic.
Known limitations
Does not understand complex business logic beyond pattern matching. Cannot execute code to verify if tests actually pass. Does not make final merge or block decisions.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.
Cannot execute code or verify test results. Limited to pattern matching on added lines. Does not make final approval or merge decisions. Requires specific script execution for verification.
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 buying a skill.
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 2 days ago
- Software
Git