The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions and provided code samples facilitate the ingestion and processing of external, untrusted data from API responses and OpenAPI specifications.
Ingestion points: Data is ingested via loadOpenAPISpec and loadJSONSchema in schema-loader.ts from the file system, and through API response bodies in validate-responses.spec.ts and schema-validation.spec.ts from remote network endpoints.
Boundary markers: The skill does not define specific boundary markers or instructions to the agent to disregard potential instructions embedded within the API response data or specification files.
Capability inventory: The skill's environment includes file system read access (fs.readFileSync) and network request capabilities via Playwright and REST Assured.
Sanitization: The skill employs the Ajv library for schema validation, which enforces that ingested data matches a predefined structural contract, providing a layer of validation against malformed payloads.
[DYNAMIC_EXECUTION]: The skill uses the Ajv library to validate data against JSON schemas, which involves runtime code generation.
Evidence: The validateAgainstSchema function in contract-helpers.ts calls ajv.compile(schema), which dynamically generates validation functions from schema definitions to perform checks at runtime.

api-contract-validator