The Agent Skills Directory

[SAFE]: The skill is designed to manage internal documentation and configuration. It operates exclusively on local files within the repository context (such as .claude/rules/ and CLAUDE.md) and does not engage in network activity or access sensitive system paths.- [COMMAND_EXECUTION]: The skill instructions specify the use of a standard shell pipeline (cat, shasum, cut) to generate a SHA-256 hash. This hash is used to ensure that the generated REVIEW.md stays synchronized with the source rules files, which is a common and safe development practice.- [DATA_EXPOSURE]: The skill documentation explicitly forbids the inclusion of secrets, credentials, or internal paths in the generated REVIEW.md file, showing a proactive approach to preventing accidental data leaks.

review-steering