The Agent Skills Directory

[SAFE]: The skill performs routine software development lifecycle tasks, such as checklist generation and codebase scanning, without any detected malicious patterns.
[DATA_EXPOSURE]: The skill reads project-specific configuration files including CLAUDE.md and milestone files in production/milestones/ to gather context. This access is limited to the local project environment and intended for checklist generation.
[COMMAND_EXECUTION]: The skill uses the Grep and Glob tools to scan the repository for TODO, FIXME, and HACK comments to assess codebase health. These are read-only operations restricted to the local filesystem.
[DATA_WRITE]: The skill generates a markdown document and explicitly asks for user permission before using the Write tool to save the checklist to the production/releases/ directory.

release-checklist