teach
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external URLs and community forums to ground its teaching content, creating an attack surface for embedded instructions.\n
- Ingestion points: MISSION.md (user input), RESOURCES.md (external URLs), and external community resources.\n
- Boundary markers: None specified to prevent the agent from following instructions within external data.\n
- Capability inventory: File system write access (lessons, assets), CLI command execution (opening files), and network access.\n
- Sanitization: No defined sanitization or validation for content fetched from external sources.\n- [DYNAMIC_EXECUTION]: The skill generates interactive HTML files and JavaScript assets (widgets, simulators) at runtime, which are stored locally and executed in the user's browser environment.\n- [COMMAND_EXECUTION]: The agent is instructed to open the generated lesson files using CLI commands, which automates the local execution and rendering of its own generated code.
Audit Metadata