trailsnap-agent

Pass

Audited by Gen Agent Trust Hub on Oct 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from a user's photo library, such as OCR text and descriptions, which may contain untrusted content. The instructions mitigate this by directing the agent to separate observed metadata from generated prose and to treat library information as evidence rather than absolute facts. Ingestion points: Data is retrieved via query tools like trailsnap_search_photos and trailsnap_investigate_memory described in references/workflows.md. Boundary markers: The agent is instructed to present memories and descriptions with uncertainty and to clearly distinguish metadata. Capability inventory: The skill can generate responsive HTML documents based on search results. Sanitization: The workflows explicitly prohibit the use of scripts that could mutate library data in generated HTML documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 28, 2026, 02:54 AM