fitness-coach
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data via OCR from images, creating an attack surface for instructions embedded within screenshots to influence agent behavior.
- Ingestion points: Screenshots and image crops from fitness apps specified in the 'Inputs' and 'OCR and Data Parsing' sections of SKILL.md.
- Boundary markers: While the skill includes behavioral boundaries (e.g., 'Do not diagnose'), it lacks specific instructions to ignore potential adversarial commands found within the extracted OCR text.
- Capability inventory: The skill identifies and interprets data to generate recommendations; no scripts were found that perform network requests, file writes, or command execution.
- Sanitization: No sanitization or validation logic is defined to handle or filter malicious content extracted from the images.
Audit Metadata