The Agent Skills Directory

[SAFE]: The skill incorporates secure development practices, such as using werkzeug.security for password hashing (generate_password_hash and check_password_hash) and retrieving sensitive configuration from environment variables rather than hardcoding them, as seen in the configuration and authentication guides.
[INDIRECT_PROMPT_INJECTION]: The templates include endpoints that process data from external sources via JSON bodies and query parameters, which is a standard surface for indirect prompt injection if the API is later integrated with LLM agents.
Ingestion points: Data is ingested via request.get_json() and request.args.get() in references/authentication-and-jwt.md and references/blueprints-for-modular-api-design.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are included in the code templates.
Capability inventory: The skill facilitates database writes, user authentication, and profile management.
Sanitization: The skill provides a validate_json decorator and validate_email function in references/request-validation.md to ensure data conforms to expected formats.

flask-api-development