The Agent Skills Directory

[SAFE]: The skill is primarily educational, providing best practices for advanced TypeScript features such as Stage 3 decorators, variadic tuples, and recursive conditional types. It contains no executable code or scripts that perform malicious actions.
[INDIRECT_PROMPT_INJECTION]: The skill discusses patterns for ingesting external data, such as route parameters, environment variables, and API payloads, which represents a potential injection surface. However, the skill provides specific remediation guidance to mitigate these risks.
Ingestion points: External data enters through route strings (dsl-route-param-inference.md), environment variables (impl-env-config-loader.md), and HTTP request bodies (impl-schema-derived-api-client.md).
Boundary markers: The instructions consistently advocate for the use of Zod or Valibot schemas to create strict boundaries for untrusted data.
Capability inventory: The skill facilitates static type checking and structural transformations; it does not include capabilities for arbitrary command execution or network exfiltration.
Sanitization: The patterns emphasize schema validation, data coercion, and type-level assertions (assertNever) to ensure all data conforms to expected shapes before processing.
[EXTERNAL_DOWNLOADS]: The skill references established and trusted libraries and tools, including Zod, Valibot, tRPC, and official TypeScript documentation. These are well-known services and organizations, and their mention is standard for the documented development workflows.

typescript-advanced-patterns