tutor-pro
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (student level, subjects, learning goals, and exam answers) to generate personalized educational content and study plans.
- Ingestion points: Student data collected during the Intake interview (Step 1) and responses to the VARK learning style quiz (Step 3) in SKILL.md.
- Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings for the agent when processing student input.
- Capability inventory: The skill instructs the agent to generate and save output in markdown format using student-provided naming components (e.g.,
tutor-<student>-<subject>-...). - Sanitization: There is no explicit requirement for the agent to sanitize, escape, or validate the user-provided text before incorporating it into the final tutoring plan or analysis reports.
Audit Metadata