optimize-claude-md
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs standard Unix utilities (Bash, wc, awk, fd) locally to measure file sizes, token counts, and identify project structure. These commands are used for analytical purposes within the project scope.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes and processes user-controlled documentation files (CLAUDE.md). This ingestion surface is managed through a workflow that requires manual user approval for all proposed text modifications, preventing automated execution of potentially malicious embedded instructions.
- [SAFE]: The skill implements strict safety protocols defined in its 'Hard Rules,' including the preservation of content during extraction, mandatory diff previews, and a prohibition on editing canonical skill metadata directly.
Audit Metadata