The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Node.js scripts (.claude/hooks/scripts/extract-session-patterns.js and .claude/hooks/scripts/validate-content.js) to analyze project data and validate generated content. These are internal maintenance scripts located within the skill's directory.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes session history and file changes to identify patterns, which could potentially ingest untrusted content from the project files.
Ingestion points: Tool call history and file changes are scanned by the extraction script.
Boundary markers: Not explicitly defined in the provided text.
Capability inventory: The skill can execute local scripts and write new files to the .claude/rules/ and .claude/skills/ directories.
Sanitization: The process includes a validation script (validate-content.js) and requires the user to manually approve, edit, or reject each suggested pattern before it is saved.

learn