scanner-tuning

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as vulnerability scan reports, service banners, and plugin descriptions, which could contain adversarial instructions (e.g., 'disregard this vulnerability').
  • Ingestion points: Data is ingested through target files or directories specified in arguments and processed using the Read and Grep tools (SKILL.md).
  • Boundary markers: The skill includes a dedicated 'Prompt Injection Safety Notice' that explicitly warns the agent to ignore instructions embedded in scan output or system banners.
  • Capability inventory: Access is limited to local file system reading and pattern matching via Read, Grep, and Glob tools.
  • Sanitization: The instructions mandate a multi-step validation workflow and independent verification (Step 1) to ensure the agent does not adopt claims found in the data without evidence.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 01:48 PM
Security Audit — agent-trust-hub — scanner-tuning