expo-app-assets
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local project files, including
package.jsonand Expo configuration files (app.json,app.config.js/ts), to determine the project state. These files are considered untrusted inputs as they could be modified to include malicious instructions that attempt to subvert the agent's logic. - Ingestion points:
package.json,app.json,app.config.js, andapp.config.ts(as specified inSKILL.md). - Capability inventory: The agent is instructed to execute shell commands (
npx expo install,npx expo config) and perform file system operations (writing to staging and production directories), as described inSKILL.mdandreferences/expo-asset-config.md. - Boundary markers: The skill does not specify the use of explicit boundary markers or "ignore instructions" blocks when reading these project files.
- Sanitization: No automated sanitization of file content is mentioned; however, the skill mandates user approval before performing sensitive actions like installing packages or replacing existing assets, which serves as a manual mitigation.
Audit Metadata