Staff Engineer Reviewer

Overview

This agent simulates the role of a Staff Engineer operating in 'paranoid reviewer mode.' It is designed to perform deep, structural audits on code differences (diffs) submitted for pre-landing review—meaning the implementation is complete, but before it ships to production.

The goal is not merely to check for style issues or bugs caught by standard CI/CD pipelines. Instead, this agent hunts for subtle, high-impact flaws that commonly survive automated testing but could cause catastrophic failures in a live production environment.

Capabilities

• Structural Analysis: Examines code changes against the main branch to identify deep architectural weaknesses.
• Performance Auditing: Specifically looks for N+1 query patterns and missing database indexes.
• Concurrency & Safety Checks: Detects potential race conditions, stale reads, and issues related to transaction integrity.
• Security Review: Scrutinizes trust boundaries, LLM interaction points, and SQL safety vulnerabilities (e.g., escaping bugs).
• Logic Validation: Identifies broken invariants, faulty retry logic, and conditional side effects that might not be covered by existing test suites.

Example Use Cases

1. Pre-Deployment Gatekeeping: Before merging a major feature branch into main, run this agent to ensure all structural integrity checks pass.
2. Complex Logic Review: When implementing multi-step workflows involving external APIs or database writes, use it to validate that failure modes are handled robustly.
3. Security Hardening: Use it specifically on code touching authentication or data persistence layers to check for boundary violations.