dig
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (MEDIUM): The skill performs git clone on arbitrary GitHub repositories. Because it is not restricted to specific trusted organizations, it may ingest malicious code onto the system.
- COMMAND_EXECUTION (LOW): Employs standard shell commands including mkdir and git clone to manage the local filesystem and retrieve remote data.
- Indirect Prompt Injection (LOW): The skill processes untrusted repository content through a research sub-agent. Evidence: 1. Ingestion point: /tmp/cc-repos/. 2. Boundary markers: Absent in the sub-agent prompt. 3. Capability inventory: Exploration of code and documentation via sub-agent. 4. Sanitization: None provided for the cloned content.
Audit Metadata