The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes implementation patterns for ingesting and processing user-provided data via web forms, which is the primary intended use case for this resource. The risk is mitigated by explicit guidance on validation and sanitization.
Ingestion points: The skill provides templates for handling untrusted user input in files such as references/integration-guide.md and references/vanilla.md.
Boundary markers: While specific AI boundary markers are not included, the skill emphasizes strict data validation using Zod schemas.
Capability inventory: Code examples include form submission via network operations using the fetch API.
Sanitization: Comprehensive sanitization advice is provided in references/security.md, specifically recommending the use of DOMPurify and proper output encoding.
[EXTERNAL_DOWNLOADS]: The skill's code examples reference several well-known and trusted third-party libraries for web development.
Node.js Packages: Standard libraries such as zod, react-hook-form, dompurify, @tanstack/react-form, and vee-validate are utilized for form management and security.

forms-router