continuous-learning
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to evaluate full session transcripts for pattern extraction, creating a surface for indirect prompt injection attacks.
- Ingestion points: The script evaluate-session.sh extracts the session transcript path from the agent hook input and facilitates the reading of that transcript file for analysis.
- Boundary markers: The skill does not implement or specify boundary markers or instructions to isolate the transcript content, which may lead the agent to follow instructions embedded within the session data during the evaluation process.
- Capability inventory: The skill uses a bash script to perform file system checks, directory creation, and messaging to trigger agent evaluation of local session files.
- Sanitization: No sanitization, filtering, or escaping is performed on the transcript content before it is processed by the agent.
Audit Metadata