review-process

name	review-process
description	Use when an agent must review code changes. Defines the universal review sequence — identify files, read context, check against domain criteria, score, filter. Compose with a domain skill and review-output-schema.

Review Process

Base review protocol for all code reviewers. Every review follows this sequence.

Steps

Step 1: Identify Changed Files

• If given a git range: git diff <range> --name-only filtered to relevant extensions
• If given file paths: use those directly
• If given a feature description: search for recently modified files

Step 2: Read Full Context

For each changed file:

1. Read the complete file — not just the diff
2. Read any project headers it includes (not engine/framework headers) — you need to see the types being used
3. If a changed function takes or returns a pointer/reference, Grep for that symbol to understand its lifecycle across call sites

Adapt depth to your mandate: style review may be self-contained per file; safety and correctness review require cross-file tracing.

Step 3: Check Against Domain Criteria

Systematically check every rule from your loaded domain skill(s). For each file, for each rule category — be thorough but precise. Only flag things that are clearly violations, not judgment calls.

Step 4: Score and Filter

Rate every potential finding on a confidence scale. Apply the scoring thresholds and output format from your output schema. Only report issues you can substantiate with specific code evidence.