refactor
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill operates on codebase files which are external and potentially untrusted data. While it includes verification and safety steps, it lacks specific delimiters to separate file content from the agent's internal logic. • Ingestion points: Project source code and configuration files. • Boundary markers: Not explicitly defined in the instructions. • Capability inventory: File modification and deletion. • Sanitization: Relies on user verification and characterization tests rather than programmatic sanitization.
- [COMMAND_EXECUTION]: The skill suggests running industry-standard analysis tools and test suites. • Evidence: Mentions knip, ts-prune, vulture, cargo-udeps, and git grep.
Audit Metadata