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
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:30 PM