droidrun/mobile-harness

Adoption judgment: treat this as an exploratory candidate that requires validation, not as a proven production-grade mobile automation foundation. The supplied material only shows that one X user shared the project; it does not establish compatibility, reliability, maintenance quality, or production readiness.

The GitHub title and description position it as a set of mobile-control skills for AI agents, covering Android, iOS, and cloud phones. A more accurate analogy is a capability or control layer that gives an agent ways to operate mobile devices, rather than an application that delivers a finished business outcome. It should not be assumed to be a hosted cloud-phone service, a one-click consumer automation app, or a complete mobile testing platform. The evidence contains no API details, setup guide, device matrix, task examples, or observed execution results, so the practical scope and reliability remain unverified.

On prerequisites and cost, the supplied evidence contains only a GitHub repository link. It provides no official pricing, API-fee schedule, device cost, or community cost demonstration.