The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard Unix utilities (wc, ln, ls, cat) to monitor file size and manage symbolic links for project configuration files (AGENTS.md, CLAUDE.md). These operations are limited to local filesystem metadata and file organization.
[SAFE]: The skill references internal workflow tools and skills using the cipherpowers namespace (e.g., cipherpowers:api-patterns, cipherpowers:commit-workflow), which are contextually appropriate for the developer's platform and the skill's maintenance purpose.
[SAFE]: The instructions include a "Critical Infrastructure Warning" that advises against blindly trusting auto-generated content, advocating for manual curation and security review of AI instructions, which is a defensive best practice.

maintaining-instruction-files