The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs on how to install official React development tools and bootstrap projects using npm and npx. These resources are fetched from established and trusted package registries.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides examples of components that process data from external sources, such as user inputs via FormData, API responses, and local storage. These ingestion points define a potential attack surface for indirect prompt injection if the agent applies these patterns without sufficient data isolation or sanitization.\n
Ingestion points: Handling of FormData in Server Actions, data fetching from APIs, and reading from browser localStorage.\n
Boundary markers: The instructional code snippets do not explicitly show the use of boundary markers or instructions to ignore embedded commands when processing untrusted data.\n
Capability inventory: The skill configuration includes access to Bash and file-editing tools, which increases the potential impact of an injection vulnerability.\n
Sanitization: While some snippets include basic validation logic, the primary focus is on feature demonstration rather than comprehensive input sanitization.

react-patterns