teach
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external resources to generate its teaching content.
- Ingestion points: External websites, community forums, and URLs provided in
RESOURCES.md. - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external content as untrusted or to ignore instructions embedded within it.
- Capability inventory: Extensive file-write capabilities across the workspace (lessons, reference docs, glossary) and the ability to execute CLI commands to open files.
- Sanitization: No mechanism is specified for sanitizing or escaping content retrieved from external sources before it is used in prompt interpolation or file generation.
- [COMMAND_EXECUTION]: The skill instructions suggest running a CLI command to open lesson files for the user. While intended for benign use (e.g.,
open), this creates a surface for potential command injection if the agent is manipulated into using malicious filenames or arguments. - [DYNAMIC_EXECUTION]: The skill generates interactive HTML files containing 'quiz widgets, simulators, and diagram helpers.' This involves writing code (likely JavaScript) to the local file system, which is then executed when the user opens the file in a browser environment.
Audit Metadata