skill-staged-review

A two-pass code review initially validates code against predefined intent contracts and success boundaries, then executes a secondary pass to evaluate overall code quality metrics and structural completeness.

You can automate code scanning by invoking the skill on your codebase to quickly detect TODOs, empty functions, suspicious returns, and mock data, providing immediate insights into code completeness.

You need specification compliance verification when preparing a major feature or PR, ensuring the code strictly adheres to predefined contracts and maintains high quality standards before merging.

Yes, external model-based assessments are integrated into the quality evaluation pass, allowing development teams to perform detailed external evaluations alongside automated checks for complex projects.

Structured code review explicitly separates specification validation from quality metrics by guiding users through loading intent contracts first, followed by targeted quality checks like stub detection.

Automated code scanning focuses primarily on detecting incomplete implementations like TODOs and mock data, meaning complex specification boundary violations may still require external evaluation to fully validate.