review

Reviews pull request diffs for SQL safety, race conditions, and LLM trust boundary violations.

Updated Aug 9, 2026
One-click install
npx skills add https://github.com/raghavbadhwar/rstack --skill review-raghavbadhwar
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review
Source: https://github.com/raghavbadhwar/rstack/tree/main/review
Command: npx skills add https://github.com/raghavbadhwar/rstack --skill review-raghavbadhwar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Catching structural bugs like SQL injection, race conditions, and unsafe LLM output handling before code lands is hard to do consistently by eye. This Skill automates pre-landing PR review by analyzing the diff against the base branch with a structured two-pass checklist and parallel specialist reviewers. ## Core Features & Use Cases - Two-Pass Diff Review: Runs a CRITICAL pass (SQL safety, race conditions, LLM trust boundaries, shell injection, enum completeness) followed by an INFORMATIONAL pass (async mixing, type coercion, CI/CD pipeline issues). - Parallel Specialist Subagents: Dispatches dedicated reviewers for testing gaps, maintainability, security, performance, data migrations, API contracts, and adversarial red-team analysis on large diffs. - Fix-First Workflow: Auto-applies mechanical fixes (dead code, N+1 queries, magic numbers) while batching judgment calls into a single structured user question. - Greptile Triage: Fetches, classifies, and replies to Greptile bot comments on GitHub PRs with evidence-backed responses and per-project suppression history. - Use Case: Before merging a feature branch, ask for a pre-landing review to get a terse report of critical and informational findings, with safe fixes already applied and risky ones queued for your decision. ## Quick Start Ask the agent to review this PR against the main branch before merging.

Frequently Asked Questions about review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a pre-landing code review on my PR?

Invoke the review skill with a phrase like "review this PR" or "check my diff". It analyzes git diff against the base branch, runs critical and informational checklist passes, and reports findings with auto-applied fixes and items needing your input.

What issues does an automated PR review checklist catch?

It catches SQL injection and N+1 queries, race conditions, unvalidated LLM output written to databases, shell injection, incomplete enum handling, async/sync mixing, type coercion bugs, and CI/CD pipeline misconfigurations, each cited with file and line.

Does the review skill auto-fix the problems it finds?

Yes, mechanical fixes like dead code removal, missing eager loading, and magic numbers are applied automatically. Riskier changes such as security fixes, race conditions, or behavior changes are batched into a single question for your decision.

Can it triage Greptile bot comments on GitHub PRs?

Yes, it fetches Greptile line-level and top-level comments via the gh API, classifies them as valid, already fixed, or false positive, and posts evidence-backed replies. Per-project history suppresses known false positives.

When does the design review checklist run during a PR review?

The design checklist runs only when the diff touches frontend files, detected via gstack-diff-scope. It flags AI-slop patterns, typography and spacing issues, and missing interaction states, calibrated against the repo's DESIGN.md if one exists.

What are the limitations of automated diff review?

It reviews only the diff against the base branch plus directly related consumers, so systemic issues outside the change may be missed. Low-confidence design findings require visual verification, and it depends on git, the gh CLI, and gstack binaries being installed.