The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local bash commands to retrieve metadata and line counts for specific agent configuration files.
Evidence: The script uses wc -l to analyze MEMORY.md, CLAUDE.md, and global Claude rules.
Evidence: The script uses ls and grep to identify topic-specific markdown files and rule files in the .claude/rules/ directory.
[COMMAND_EXECUTION]: The skill performs automated validation of file references found within memory logs to identify stale entries.
Evidence: Step 3 extracts potential file paths from MEMORY.md and uses the shell test operator [ ! -f "$f" ] to verify their existence.
Evidence: The extraction process uses a strict regular expression whitelist ([a-zA-Z0-9_/.-]+\.(ts|js|py|md|json|yaml|yml)) which prevents shell metacharacters from being passed to the existence check, mitigating command injection risks.

status