adversarial-machine-learning

Pass

Audited by Gen Agent Trust Hub on Feb 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONNO_CODE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (LOW): The skill directs the agent to fetch data from an external GitHub repository (gmh5225/awesome-ai-security). This source is not on the trusted organizations or repositories list.
  • [PROMPT_INJECTION] (LOW): Risk of Indirect Prompt Injection (Category 8 surface). (1) Ingestion Point: Remote Markdown content from the specified GitHub URL. (2) Boundary Markers: None specified to differentiate data from instructions. (3) Capability Inventory: The skill contains no scripts, so risk depends on the agent's inherent capabilities. (4) Sanitization: No sanitization or verification of the fetched content is required.
  • [NO_CODE] (SAFE): The skill consists only of a markdown file with metadata and instructions; no executable scripts, binaries, or configuration files are present.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 23, 2026, 07:21 AM