pydantic
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents how to ingest and validate untrusted data from external sources (e.g., JSON request bodies), which is a common surface for indirect prompt injection if that data is subsequently used in a prompt context.
- Ingestion points: Examples in SKILL.md demonstrate ingesting data via model_validate_json and model_validate.
- Boundary markers: The skill encourages the use of Pydantic models, which define schema boundaries for incoming data.
- Capability inventory: The skill contains code examples for documentation purposes; it does not ship with active scripts that execute commands or perform network operations.
- Sanitization: The library itself is designed for sanitization, and the documentation demonstrates using constraints like min_length, EmailStr, and custom field validators.
Audit Metadata