The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Recommends downloading the eas-cli global package and the submit-expo-feedback utility via npx for deployment and feedback tasks.
[REMOTE_CODE_EXECUTION]: Instructs the use of npx --yes submit-expo-feedback@latest to execute remote code from a package registry, which is a common but sensitive pattern for developer tools.
[COMMAND_EXECUTION]: Contains instructions for the agent to execute shell commands for exporting web bundles, deploying to EAS, and interacting with the environment.
[INDIRECT_PROMPT_INJECTION]: The example implementation in app/api/ai+api.ts creates a vulnerability surface by accepting a prompt from an untrusted request body and forwarding it to an LLM provider without sanitization.
Ingestion points: app/api/ai+api.ts via request.json().
Boundary markers: No delimiters or instructions are used to isolate the user-provided prompt from the system context in the OpenAI API call.
Capability inventory: The skill demonstrates the ability to make external network requests (fetch) to third-party APIs.
Sanitization: No input validation or content filtering is implemented for the data processed by the API route.

eas-hosting