fastapi

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill implements a web API template that ingests untrusted data via JSON and Form inputs, establishing a surface for indirect prompt injection.\n
  • Ingestion points: Data enters the system via schemas.UserCreate, schemas.ItemCreate, and OAuth2PasswordRequestForm in router files (e.g., src/auth/router.py).\n
  • Boundary markers: The skill emphasizes the use of Pydantic schemas with explicit validation constraints (e.g., Field, min_length, EmailStr, gt=0) to validate and delimit input data.\n
  • Capability inventory: Capabilities are restricted to standard API operations: database interactions via SQLAlchemy, JWT token operations, and password hashing; no arbitrary command execution or dynamic code evaluation is present.\n
  • Sanitization: Input validation and sanitization are handled via Pydantic's type system and custom validators before processing.\n- [EXTERNAL_DOWNLOADS]: The skill manages dependencies using the standard uv package manager and official Python registries.\n
  • Packages: It utilizes well-known libraries including fastapi, sqlalchemy, aiosqlite, python-jose, passlib, and pydantic-settings.\n
  • Source: All referenced dependencies and tools are industry-standard for the FastAPI ecosystem and do not originate from unknown or suspicious sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:49 AM