app-store-connect
Warn
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent and user to install several third-party tools from non-trusted GitHub repositories. Examples include brew install revylai/tap/greenlight, go install github.com/RevylAI/greenlight/cmd/greenlight@latest, and npx skills add truongduy2611/app-store-preflight-skills.
- [REMOTE_CODE_EXECUTION]: By encouraging the installation and execution of unverified third-party software (such as the asc CLI by rorkai/rudrankriyam and greenlight by RevylAI), the skill creates a path for remote code execution if these external repositories are compromised.
- [COMMAND_EXECUTION]: The skill relies heavily on shell command execution for its primary functions, including build signing, metadata synchronization, and analytics processing. It interpolates numerous variables into shell scripts (e.g., PLAN_HASH, RUN_ID, APP_ID), which requires strict validation of inputs to prevent command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and scans external data such as app metadata, descriptions, and keywords from the local project to detect rejection patterns. This untrusted content is processed by the agent's logic, creating a vulnerability surface for indirect prompt injection attacks where malicious instructions hidden in app metadata could influence agent behavior.
- [DYNAMIC_EXECUTION]: The skill utilizes browser automation (Playwright/Cursor browser MCP) to navigate the App Store Connect web UI for tasks not supported by the public API (e.g., creating app records). While intended for functionality, this dynamic interaction with a web environment using agent privileges presents an increased attack surface.
Audit Metadata