readability
Pass
Audited by Gen Agent Trust Hub on Apr 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes bundled Ruby scripts that invoke the
git showcommand via theOpen3module to compare file content across different branches. - [PROMPT_INJECTION]: The skill processes untrusted user-provided text files, which constitutes a surface for indirect prompt injection.
- Ingestion points: Text files analyzed by scripts in the
scripts/directory. - Boundary markers: None; file contents are analyzed as raw text.
- Capability inventory: Subprocess execution (
git) and file system access. - Sanitization: None; normalization is performed for scoring accuracy but not for security sanitization.
Audit Metadata