The Agent Skills Directory

[SAFE]: The skill uses local shell scripts (scan-skills.sh, scan-rules.sh) to gather data from the user's own .claude/skills and .claude/rules directories. This is consistent with its stated purpose of rules maintenance.
[SAFE]: The instructions explicitly state: 'Never modify rules automatically. Always require user approval.' This prevents unauthorized persistence or modification of the agent's operating environment.
[SAFE]: Remote code execution or untrusted downloads are not present. The scripts use standard utilities like grep, awk, sed, find, and jq to process local files.
[SAFE]: Data processing is contained within the agent's context for analysis and results are stored locally in a results.json file within the skill directory.

rules-distill