The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation provides instructions for using npx to execute utilities such as apidog-mcp-server, openapi-typescript, and orval. These are industry-standard developer tools used for facilitating Model Context Protocol (MCP) communication and automating the generation of TypeScript types from API specifications.
[DATA_EXFILTRATION]: The instructions demonstrate standard practices for client-side authentication, including the retrieval of tokens from localStorage and their inclusion in Authorization: Bearer headers. These patterns are documented for educational purposes within a frontend development context and do not represent malicious exfiltration.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an ingestion point for external data by having the agent read OpenAPI specifications from remote URLs (e.g., https://api.example.com/openapi.json).
Ingestion points: Remote OpenAPI JSON/YAML URLs used for type generation.
Boundary markers: The skill explicitly advocates for the use of Zod schemas to perform runtime validation (UserSchema.parse(response.data)), which acts as a robust boundary for incoming data.
Capability inventory: The agent utilizes npx for static code generation and local file writing for types.
Sanitization: External content is sanitized via Zod validation logic to ensure the application only processes data matching the expected schema.

api-integration