teach
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from files like MISSION.md, RESOURCES.md, and learning records to guide the agent's logic.
- Ingestion points: MISSION.md, RESOURCES.md, and learning-records/*.md.
- Boundary markers: None provided to separate instructions from data.
- Capability inventory: Writing HTML files and executing CLI commands.
- Sanitization: No process for sanitizing external or user-provided resource content.
- [COMMAND_EXECUTION]: The instructions direct the agent to "open the lesson file for the user by running a CLI command." This presents a risk if file names, which are generated based on mission topics, are manipulated to include command injection sequences.
- [DYNAMIC_EXECUTION]: The skill requires the agent to generate interactive HTML lessons with simulators and widgets. This involves dynamic generation of JavaScript based on external knowledge sources, which could lead to malicious code execution in the user's browser context.
Audit Metadata