teach-me
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and executes external code or data provided by the user to provide runtime-based explanations. This creates a surface for indirect prompt injection if the ingested content contains instructions designed to manipulate the agent.
- Ingestion points: User-specified files, functions, concepts, or error messages (SKILL.md).
- Boundary markers: None explicitly mentioned for separating untrusted input from agent instructions.
- Capability inventory: Execution of Python/Node.js code, platform-specific tools (
/run,/verify), and temporary file modification for instrumentation. - Sanitization: No specific sanitization or filtering of the input content before execution or analysis is described.
- [COMMAND_EXECUTION]: The skill is explicitly instructed to execute code at runtime using
python3,node, and platform-specific tools (/run,/verify) to observe state and verify logic. While this is the primary functional purpose, it involves running potentially untrusted code found in the user's environment.
Audit Metadata