maestro
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The file
MAESTRO.mdprovides an installation commandcurl -fsSL "https://get.maestro.mobile.dev" | bashto fetch and execute the Maestro CLI setup script from the tool's official domain. - [COMMAND_EXECUTION]: The instructions throughout
MAESTRO.mdandFLUTTER.mddescribe executing various system and development commands includingadbfor Android interaction,xcodebuildandplutilfor iOS simulator configuration, andbrewfor package management. - [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by processing untrusted data from external services:
- Ingestion points: Test flows ingest content from external mail service APIs (e.g., Mailpit via
GET /api/v1/messages) to retrieve OTP codes, as described inMAESTRO.md. - Boundary markers: The skill does not recommend the use of delimiters or specific instructions to ignore embedded commands within the fetched data.
- Capability inventory: The environment supports execution of shell commands (
maestro test), mobile device manipulation (adb,xcodebuild), and network operations via JavaScript. - Sanitization: No sanitization or validation logic is defined for the data retrieved from external API responses before interpolation into test steps.
- [DYNAMIC_EXECUTION]: In
MAESTRO.md, the skill recommends executing a local mock API server usingnpx tsx scripts/mock-api-server.ts, which involves running locally generated or provided scripts at runtime. - [EXTERNAL_DOWNLOADS]: The skill fetches configuration rules and citations from external repositories including
mobile-dev-inc/maestro-docsandtovimx/maestro-mobile-testing-skill, and provides direct links to official Maestro Studio binaries hosted onstudio.maestro.dev.
Audit Metadata