pair-mode
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill contains purely behavioral instructions and programming examples. No malicious code, remote downloads, persistence mechanisms, or obfuscated patterns were detected in the instructions or metadata.
- [PROMPT_INJECTION]: The skill defines a broad surface for indirect prompt injection due to its intended use case and tool permissions. Ingestion points: User-provided source code files are read through the file-read tool. Boundary markers: No specific delimiters or safety instructions are used to distinguish user data from agent instructions. Capability inventory: The skill enables 'bash: all' and 'file: read, write, edit' permissions. Sanitization: No input validation or content filtering is implemented for the code being analyzed.
Audit Metadata