eas-workflows
Pass
Audited by Gen Agent Trust Hub on Nov 15, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches documentation and JSON schemas from
api.expo.devandraw.githubusercontent.com/expo/. These are official Expo domains and repositories, which are recognized as trusted vendor sources. - [COMMAND_EXECUTION]: The skill uses
npx eas-cliand a local Node.js fetch script. The commands are scoped to official tools (eas-cli,@expo/agent-cli) and local scripts provided with the skill. The fetch script includes basic local caching functionality and does not execute the data it downloads. - [REMOTE_CODE_EXECUTION]: While the skill downloads external content (Markdown and JSON schema), it treats them as reference data for the agent to read. No remote scripts are executed, and the
fetch.jsscript only returns text output. - [DYNAMIC_EXECUTION]: The
fetch.jsscript uses standard Node.js APIs for network requests and file system caching. It does not useeval()or other unsafe dynamic execution patterns. - [INDIRECT_PROMPT_INJECTION]: The skill ingests external data (schema and docs). However, this data comes from trusted Expo sources, and the skill includes instructions to prioritize the official schema as the source of truth, reducing the surface for injection via third-party documentation.
Audit Metadata