The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends using npx to run @pymodel/react-frontend-skills, which downloads and executes the package from the registry. This is a legitimate utility provided by the skill author for project version detection and API scanning.
[COMMAND_EXECUTION]: The documentation includes commands for initializing the MSW environment and scanning the project for deprecated APIs using npx. These are standard development tasks consistent with the skill's purpose.
[INDIRECT_PROMPT_INJECTION]: The skill describes handlers that ingest and process external request data (JSON bodies, query parameters, and path parameters). While this creates a surface where untrusted data could influence agent responses if echoed back, the skill promotes standard mocking patterns without specific exploitation vectors.
Ingestion points: Request bodies (request.json()), query parameters (url.searchParams), and path parameters (params.id) in various resolver examples.
Boundary markers: None explicitly shown in the example mock handlers.
Capability inventory: The skill demonstrates network interception and command execution via npx for setup and scanning.
Sanitization: Examples focus on matching and returning mock data; explicit sanitization of request content is not demonstrated but mocking behavior is generally isolated.
[DYNAMIC_EXECUTION]: The 'GraphQL Schema-First' guideline demonstrates the use of the graphql package to dynamically resolve operations against a schema at runtime. This is a standard and recommended practice for maintaining schema-accurate mocks.

msw