The Agent Skills Directory

[SAFE]: The skill provides structural guidance for building frontend applications using a feature-based architecture. All provided code snippets demonstrate legitimate use of established libraries like Axios, Zustand, and TanStack Router.
[INDIRECT_PROMPT_INJECTION]: The skill defines patterns for ingesting data from external APIs and URL search parameters. It mitigates injection risks by mandating the use of Zod schemas and TypeScript for strict validation and sanitization of all incoming data before it enters the application state.
[COMMAND_EXECUTION]: The instructions reference using Bun for package management and project tasks (e.g., bun run craft). These are standard developer tools and do not involve arbitrary or malicious shell execution.
[CREDENTIALS_UNSAFE]: The code includes a placeholder token (fake-token-123) within an AuthFakeGateway class. This is explicitly documented for use in unit tests and mock environments, posing no risk to real-world credentials.

frontend-engineer