The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes the @zap-studio/fetch library, which is a resource owned by the vendor (zap-studio). It also uses zod for data validation, which is a well-known and trusted library for schema definition.
[PROMPT_INJECTION]: The skill describes an interface for ingesting data from external APIs via HTTP requests. This represents a surface for indirect prompt injection if the remote data contains instructions. However, the skill mitigates this by enforcing structured data validation via Zod schemas.
Ingestion points: Remote data fetched via $fetch and api.get/post methods defined in SKILL.md.
Boundary markers: Explicit Zod schemas (e.g., UserSchema, PostSchema) define the expected structure of external data.
Capability inventory: Network access via fetch implementation.
Sanitization: Uses Zod validation to ensure received payloads conform to expected types before processing.
[CREDENTIALS_UNSAFE]: The documentation includes a placeholder for an authorization header (Authorization: 'Bearer token'). This is a standard illustrative example and does not expose real credentials.

zap-fetch-typed-http