fastapi-development
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The file
references/security-and-jwt-authentication.mdcontains a hardcoded fallback value forSECRET_KEY('dev-secret-key'). While labeled for development, hardcoded secrets in authentication logic are a security risk if inadvertently used in production environments.\n- [COMMAND_EXECUTION]: The skill includes a shell scriptscripts/validate-api.shwhich accepts external input via command-line arguments. While currently limited to echoing the input, such scripts represent a potential path for command injection if the logic is expanded without strict sanitization.\n- [DATA_EXFILTRATION]: FastAPI setup examples inreferences/fastapi-application-setup.mdconfigure the server to listen on0.0.0.0. This binds the application to all network interfaces, which can lead to unintended exposure of the service to external networks.\n- [PROMPT_INJECTION]: Indirect prompt injection surface identified through API data processing.\n - Ingestion points: User-provided registration and login data (email, password, etc.) in
references/api-routes-with-async-endpoints.md.\n - Boundary markers: None present to differentiate between user-controlled data and system-level instructions.\n
- Capability inventory: The skill implements database writes (SQLAlchemy) and security token generation (JWT).\n
- Sanitization: Pydantic models provide basic type validation and length constraints, but do not specifically filter for prompt injection patterns.
Audit Metadata