learn

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates persisting session data into files used for long-term agent guidance, which creates an indirect prompt injection surface.
  • Ingestion points: Insights are extracted from active session data and written to AGENTS.md and LESSONS.md.
  • Boundary markers: No delimiters or warnings are used to separate user-originated content from system guidance.
  • Capability inventory: The skill performs file writing across the project workspace.
  • Sanitization: No validation or sanitization is performed on the data before it is stored in the agent's permanent context.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 05:37 AM
Security Audit — agent-trust-hub — learn