The Agent Skills Directory

[SAFE]: The skill provides legitimate architectural guidance and code templates for the Litestar framework. It does not contain any malicious code, hidden instructions, or unauthorized data access patterns.
[INDIRECT_PROMPT_INJECTION]: The skill documents methods for ingesting untrusted client data through HTTP request components. 1. Ingestion points: Path, query, header, cookie, and body parameters as shown in references/parameter-and-body-patterns.md and references/request-object-and-upload-patterns.md. 2. Boundary markers: Recommends the use of typed parameters, Pydantic-style models (dataclasses), and explicit metadata (Parameter, Body) to structure incoming data. 3. Capability inventory: Includes examples for reading file content (data.read()) and accessing raw request bodies in references/request-object-and-upload-patterns.md. 4. Sanitization: SKILL.md includes a 'Validation Checklist' and 'Core Rules' that require explicit source identification, type enforcement, and constraint validation for all inputs.

litestar-requests