socratic-tutor
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines pedagogical behavior (Socratic tutoring) without requesting or using sensitive permissions, network access, or system-level tools.
- [INDIRECT_PROMPT_INJECTION]: The skill includes a 'resume from' command that allows the agent to ingest a previously saved state block provided by the user. 1. Ingestion points: User-provided text following the 'resume from:' command in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: No subprocess, file system, or network capabilities are defined or requested in the skill. 4. Sanitization: No explicit validation or filtering of the state block is defined.
- [DATA_EXPOSURE]: The 'save state' functionality is a local convenience for the user to copy their progress and does not involve exfiltration to external systems or access to credentials.
Audit Metadata