DuoPlus CloudPhone Agent
Paste the following prompt into your AI chat to install this skill:
Please install @user_1df07c13/duoplus-agent according to https://skillhub.cn/install/skillhub.md.
About this skill
Problem addressed
DuoPlus CloudPhone Agent targets stable UI automation on remote cloud phones: launching apps, locating page elements, clicking, typing, swiping, waiting, and verifying state. Like local device automation, it wraps common operations into repeatable commands, reducing the repetitive cost of writing raw ADB commands, assembling Base64 payloads, and manually parsing UI hierarchies.
How it works
The skill primarily sends JSON commands through scripts/send_command.sh, which the DuoPlus side executes. A typical flow includes:
- Element discovery: fetch the UI hierarchy, pull it locally, then use grep to find text, resource_id, or bounds="[x1,y1][x2,y2]".
- Clicking and input: prefer CLICK_ELEMENT by text; fall back to CLICK_COORDINATE when text matching fails. Before typing, tap the input field to focus it, then send INPUT_CONTENT.
- Page operations: supports LONG_COORDINATE, DOUBLE_TAP_COORDINATE, SLIDE_PAGE, GO_TO_HOME, and PAGE_BACK.
- Waiting and verification: many commands are fire-and-forget and do not return execution results, so take a screenshot after each action, compress it, and pull it locally for analysis; WAIT_FOR_SELECTOR can also wait for a target element to appear.
Boundaries and notes
- The coordinate system uses relative values from
0to1000, not pixels. - Swipe coordinates should avoid overly regular integers, and start/end points should vary between consecutive actions to reduce the risk of looking like a fixed script.
- Text input depends on focus: run
CLICK_ELEMENTorCLICK_COORDINATEbeforeINPUT_CONTENT. - Leave a local wait of
1to3seconds between commands, and avoid using device-sideshell sleep.
Use Cases
- Use a remote cloud phone to launch a specified app and verify whether expected login-page elements appear during smoke testing.
- Automate a login flow by clicking the Login element by text, falling back to coordinates, and taking a screenshot to confirm state.
- Enter an account into a focused input field, submit it with Enter, and wait briefly between commands for the UI to update.
- Swipe the mobile page to locate a loading-complete text, then wait for the expected selector before continuing the test run.
Best For
- Mobile automation test engineer who needs reliable clicking, input, swiping, and wait assertions.
- Cloud phone operations engineer who needs app launch and login-page element checks.
- UI hierarchy analyst who needs to dump UI data and grep text or resource IDs for coordinates.
- Automation script author who needs reusable ADB broadcast JSON commands for mobile actions.
Related Skills
Collects daily conversations, task logs, and errors, then runs a nine-module review to surface blind spots, generate actions, and update memory.
Local long-term memory for AI agents that appends key facts and daily logs, supports cross-session context and search recall, and auto-archives old diaries to gzip files.
Run a grilling session to interact with or test AI agents.
A systematic prompt optimization skill that refines prompts using a four-step distillation framework (diagnose, structure, think, compress) and methodologies from four prompting masters.