teach
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources tracked in
RESOURCES.mdand user-provided input inMISSION.md. This creates a vulnerability surface where maliciously crafted data from external resources or user input could influence the agent's output or behavior. - Ingestion points:
RESOURCES.md,MISSION.md, and thelearning-records/directory. - Boundary markers: The skill uses markdown headers and specific file formats (e.g.,
MISSION-FORMAT.md) to structure data, but does not implement explicit "ignore instructions" delimiters for untrusted content. - Capability inventory: The agent writes and reads markdown and HTML files and is instructed to execute shell commands to open files.
- Sanitization: No explicit sanitization or validation of the ingested external content is defined in the instructions.
- [COMMAND_EXECUTION]: The skill instructions advise the agent to use CLI commands to open generated lesson files for the user, which utilizes the agent's underlying shell execution capability.
Audit Metadata