archive-claude-skill-governance-experiment

Deploys application code to production environments via shell commands. Handles building, testing, pushing container images, and updating live services. Use when the user wants to deploy, release, ship to production, or update a live environment.

Reads markdown files in a project directory and produces a structured summary of their contents. Use when the user wants to summarize documentation, get an overview of markdown files, or understand a directory of docs.

Reformats source code files to match a consistent style guide. Reads code files across the project and applies formatting corrections using Edit. Use when the user wants to enforce code style, reformat files, or fix formatting inconsistencies.

Evaluates Claude Code skill directories for security and governance risk across 7 dimensions (permission scope, data exposure, prompt injection, blast radius, reversibility, semantic overlap, dependency risk). Produces a scorecard with overall LOW/MEDIUM/HIGH rating and remediation recommendations. Use when auditing a skill, reviewing a skill for security, or assessing skill risk before deployment.