expo-cicd-workflows-v55
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches JSON schema and Markdown documentation from official Expo infrastructure (api.expo.dev) and their public GitHub repository (github.com/expo). These sources are recognized as official documentation providers for the Expo framework.
- [COMMAND_EXECUTION]: Utilizes npx to execute eas-cli, the official command-line interface for Expo Application Services, for workflow validation. It also includes a command for submitting feedback via a specific package (submit-expo-feedback).
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external documentation to guide the agent's behavior in generating CI/CD configuration files.
- Ingestion points: SKILL.md identifies several remote URLs (schema and syntax docs) fetched via a Node.js script (scripts/fetch.js).
- Boundary markers: The documentation content is processed without explicit boundary markers or directives to ignore instructions inside the retrieved text.
- Capability inventory: The skill has permissions to write files to the project directory and execute shell commands (node, npx).
- Sanitization: Content fetched from remote sources is used to inform agent output without explicit sanitization or filtering steps.
Audit Metadata