Pre Ship Review

What it does

The Pre Ship Review skill performs a structured quality review of code before shipping, whether that's through pull requests, releases, or milestones. It focuses on identifying failures at integration boundaries – areas where contracts, examples, constants, and tests must align. The core principle is that AI-generated code often excels in isolated components but struggles with interactions between them; this skill systematically checks those interaction points to prevent shipping errors, missing documentation, or compliance issues.

When to use it

This skill is valuable in the following situations:

• Pull requests introducing multiple new modules that need to work together.
• Releases combining code from different contributors or branches.
• Milestones where quality gates must be passed before proceeding.
• Any checkpoint involving code with examples, constants across files, or interface extensions requiring validation.

Key capabilities

• Automated checks using external tools like Pyright, Vulture, and Semgrep.
• Orchestration of other AI skills (cc-skills) for deeper analysis, including code auditing and dead code detection.
• Verification of function parameter usage, configuration mappings, and domain constant correctness.
• Test quality assessment to ensure tests accurately reflect behavior.
• Identification of architectural boundary violations and implicit dependencies.
• Provides a compiled report with findings, severity levels, and suggested fixes.
• Offers three templates: New Feature Ship, Bug Fix Ship, and Refactoring Ship, each tailored for specific scenarios.

Example prompts

Here are some example prompts you might use with an AI agent equipped with this skill:

• "Run a Pre-Ship Review on this pull request."
• "Perform a 'New Feature Ship' review of the code changes in milestone X."
• "Execute the 'Bug Fix Ship' template for these modified files."

Tips & gotchas

• Template Selection is Mandatory: Always select and load the appropriate template (New Feature, Bug Fix, or Refactoring) before starting a review.
• Self-Evolving Nature: The skill improves with use; update it immediately when encountering issues or needing workarounds.
• Not for Minor Changes: This skill isn't necessary for single-file cosmetic changes, documentation updates, or dependency bumps.