review

/review — Pre-Landing Code Review

Analyze the current diff with the rigor of a principal engineer on a critical PR.

Step 0: Preamble

Run templates/PREAMBLE.md. Detect base branch.

Step 1: Gather Context

git diff --stat $(git merge-base HEAD origin/main)..HEAD
git log --oneline $(git merge-base HEAD origin/main)..HEAD

Step 2: Review Checklist

SQL Safety

• No string interpolation in SQL queries
• All user input parameterized
• LIKE patterns escaped (escape_like utility)
• No SELECT * in production code
• Migrations are reversible

LLM Trust Boundaries

• LLM output is never used in SQL, shell commands, or code execution
• User-provided prompts are sanitized
• LLM responses are validated before acting on them
• Token limits enforced on inputs

Conditional Side Effects

• No side effects in if-condition evaluation
• No mutations in logging statements
• No network calls in property getters
• No database writes in validation logic

Security

• No secrets in code (API keys, tokens, passwords)
• Auth checks on every endpoint
• CSRF protection on state-changing operations
• Input validation at system boundaries

Tests

• New code has corresponding tests
• Edge cases covered (empty, null, boundary values)
• Error paths tested
• No tests disabled or skipped without explanation

Dependencies

• New deps justified and maintained
• No known CVEs in added packages
• License compatible

Step 3: Findings

For each issue found:

### REVIEW-NNN: [CRITICAL|HIGH|MEDIUM|LOW] — {description}
**File:** {path}:{line}
**Evidence:** {code snippet}
**Fix:** {specific remediation}

Agent References

• Dispatch code-reviewer for deep code analysis
• Dispatch vulnerability-explorer for security focus
• Dispatch adversarial-reviewer for attack surface analysis

Self-Eval

Run templates/SELF-EVAL-PROTOCOL.md. Was every file in the diff reviewed? Were findings evidence-based?