Skip to main content
The Problem: As individual productivity increases (3x–10x), the volume of Pull Requests (PRs) multiplies, creating a bottleneck for human reviewers. The Solution: Shift the AI’s role from just a “coder” to a “reviewer” to act as an automated first line of defense.

​
The “Living” Rulebook

Teams must codify tribal knowledge into a REVIEW_RULES.md file in the repository root. This turns unwritten preferences into a tangible asset the AI can enforce.

​
Example REVIEW_RULES.md

# Code Review Rules

## Architecture
- All API endpoints must include rate limiting
- Database queries must use the query builder (no raw SQL)

## Security
- Never log sensitive data (passwords, tokens, PII)
- All user input must be validated

## Testing
- All new features require integration tests
- Test coverage must not decrease

​
Agentic Review Workflows

Use a REVIEW_PROCESS.md to define a step-by-step agentic routine for the AI, such as:
  1. Checking out branches
  2. Triaging issues (P0 to P2)
  3. Summarizing findings before a human looks at the code

​
Example REVIEW_PROCESS.md

# Automated Review Process

## Phase 1: Setup
1. Checkout the PR branch
2. Read all modified files
3. Review the PR description

## Phase 2: Automated Checks
1. Run linter and report violations
2. Run type checker and report errors
3. Run test suite and note failures

## Phase 3: Rule Enforcement
1. Load REVIEW_RULES.md
2. Check each rule against changes
3. Categorize violations by severity:
   - P0 (Critical): Security, data integrity
   - P1 (Important): Architecture, performance
   - P2 (Minor): Style, documentation

## Phase 4: Summary
1. Generate review summary with:
   - Total issues found by priority
   - Pass/fail recommendation
   - Top 3 issues to address
Key Takeaway: Codify tribal knowledge into enforceable rules. AI handles the “tedious nits” so humans can focus on architecture and taste.