expo-overview
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external documents such as PRDs, requests, and specifications to determine routing and setup behavior. It lacks explicit boundary markers or sanitization procedures to prevent instructions embedded within these inputs from influencing agent operations.
- Ingestion points:
SKILL.mdreferences reading user goals from requests, PRDs, or specs. - Boundary markers: None observed.
- Capability inventory: Shell command execution via
npxandeas(SKILL.md). - Sanitization: None observed.
- [COMMAND_EXECUTION]: The instructions direct the agent to run various CLI tools, including
npx,eas, andcreate-expo-app, for authentication, project initialization, and package installation. - Evidence: Commands such as
npx create-expo-app@latest,npx expo install <pkg>,eas whoami,eas login, andeas initare listed as standard setup steps inSKILL.md. - [EXTERNAL_DOWNLOADS]: The skill triggers the download and execution of packages from the NPM registry at runtime via the
npxcommand. - Evidence:
npx create-expo-app@latestandnpx --yes submit-expo-feedback@latestare used for project scaffolding and feedback submission inSKILL.md. - [DATA_EXFILTRATION]: A feedback mechanism is provided which allows for the transmission of text data from the agent's context to an external destination.
- Evidence: The
submit-expo-feedbackcommand inSKILL.mdtransmits user-supplied feedback strings to an external service using a CLI utility.
Audit Metadata