The Agent Skills Directory

[DYNAMIC_EXECUTION]: The shared library scripts/check-lib.mjs uses dynamic import() to load Node.js packages from computed file paths within its importPackage function.
[DYNAMIC_EXECUTION]: The applyMutation function in scripts/check-lib.mjs interprets a JSON-based domain-specific language (DSL) to perform various file system operations, including writing, deleting, renaming, and creating symbolic links, which is used to manage test fixtures.
[COMMAND_EXECUTION]: The skill's scripts utilize child_process.spawnSync to run local processes, specifically Node.js for self-testing and the project's Prettier instance for code formatting.
[INDIRECT_PROMPT_INJECTION]: The skill scaffolds feature components that process untrusted data via Expo Router's useLocalSearchParams. Although the skill generates and enforces validation using valibot schemas, the resulting code possesses an inherent attack surface for indirect prompt injection if validation logic is weak or bypassed.
Ingestion points: Route parameters are ingested via useLocalSearchParams() in the generated model hooks (e.g., use-note-detail-model.ts).
Boundary markers: Enforces the use of parse<Feature>Params validation functions.
Capability inventory: Generated views are pure components, while model hooks handle application logic and navigation.
Sanitization: Parameters are validated and transformed using valibot schemas before being passed to the view model.

scaffold-feature-screen