teach
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill recommends using shell commands to open the HTML lesson files generated during the learning process. This is a functional requirement for the user to view the lessons.
- [INDIRECT_PROMPT_INJECTION]: The teaching framework involves gathering information from external resources linked in
RESOURCES.md. This represents a potential surface for indirect prompt injection if external materials contain adversarial instructions. - Ingestion points: External resource links in
RESOURCES.mdand summaries of external content stored inlearning-records/(SKILL.md, LEARNING-RECORD-FORMAT.md). - Boundary markers: The skill formats do not explicitly require the use of delimiters or 'ignore' instructions for external content.
- Capability inventory: The skill performs file system operations (read/write) and shell execution for opening files.
- Sanitization: There are no instructions provided for sanitizing or escaping content retrieved from external sources before it is written to the workspace.
Audit Metadata