socratic-tutor

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as code diffs, pull request content, and session history to generate tutoring content and quizzes.
  • Ingestion points: The agent reads external code, diffs, and pull request content provided by the user or from the environment (SKILL.md).
  • Boundary markers: The instructions lack explicit delimiters or guidance for the agent to treat the ingested data as untrusted or to ignore embedded instructions within that data.
  • Capability inventory: The skill uses interactive tools like AskUserQuestion and is encouraged to "drive the debugger," creating a vector where injected instructions could potentially affect tool usage or user communication.
  • Sanitization: There are no mentioned mechanisms to sanitize or filter the ingested data for malicious prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 02:02 PM
Security Audit — agent-trust-hub — socratic-tutor