mobilerun-control

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The code intentionally installs/enables a device-side "Portal" accessibility service, programmatically writes secure settings to enable it, exposes content-provider and TCP endpoints (including an auth_token and state_full/a11y tree), and provides clipboard/screenshot and keyboard input endpoints — all of which enable covert remote control or data extraction from the phone if misused.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). SKILL.md’s required runtime workflow repeatedly ingests outsider-authored free text from the OUTSIDE phone UI (e.g., app content like Reddit posts/comments, web pages) via mobilerun device ui / mobilerun device screenshot outputs that are copied into the LLM context for “DECIDE (you): read the tree/screenshot; pick the target element by its TEXT…”