red-team-tactics
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE] (SAFE): No malicious patterns or instructions detected. The content is purely theoretical and educational.
- [NO_CODE] (SAFE): The skill contains no executable scripts, binaries, or external dependencies. Permissions are restricted to read-only tools (Read, Glob, Grep) which are appropriate for its documentation-based purpose.
- [Indirect Prompt Injection] (LOW): Evidence Chain: (1) Ingestion points: Local file access via Read and Grep. (2) Boundary markers: Absent. (3) Capability inventory: Read-only file access; no network, write, or command execution capabilities. (4) Sanitization: Absent. The risk is negligible because the agent lacks the necessary tools to perform harmful actions even if it encountered malicious data.
Audit Metadata