skills/expo/skills/eas-app-stores/Gen Agent Trust Hub

eas-app-stores

Pass

Audited by Gen Agent Trust Hub on Oct 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the eas-cli package globally via npm. It also uses npx to execute @expo/agent-cli for feedback and testflight for iOS beta submissions. These tools are associated with the vendor Expo and represent standard operational procedures for the service.
  • [COMMAND_EXECUTION]: The instructions rely heavily on shell commands to manage the build and submission lifecycle, including eas build, eas submit, and diagnostic tools like plutil and sips for inspecting iOS archives and icons. These are appropriate for the skill's primary purpose of app deployment.
  • [DYNAMIC_EXECUTION]: The skill describes the use of a JavaScript-based configuration file (store.config.js) to manage App Store metadata. This file is executed by the EAS environment and can perform dynamic operations, such as fetching localization data from remote APIs (e.g., api.example.com).
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes various project configuration files including eas.json, app.json, and store.config.json.
  • Ingestion points: Reads eas.json, app.json, package.json, store.config.json, Xcode project files, and Podfile from the local project directory.
  • Boundary markers: None identified; the skill assumes the project files are trusted and doesn't explicitly warn against instructions embedded within configuration values.
  • Capability inventory: The skill has the capability to execute shell commands (eas build/submit, npm install), perform network operations (via eas-cli and dynamic metadata config), and write to the file system (via eas init and metadata pulling).
  • Sanitization: No specific sanitization or validation of the content within these configuration files is mentioned before the agent uses them to construct commands or logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 28, 2026, 09:57 PM