argent-device-interact

Drive iOS simulators, Android emulators, and Chromium apps via argent MCP interaction tools.

Updated Aug 29, 2026
One-click install
npx skills add https://github.com/Seyamalam/pstu_final --skill argent-device-interact-seyamalam
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: argent-device-interact
Source: https://github.com/Seyamalam/pstu_final/tree/main/.agents/skills/argent-device-interact
Command: npx skills add https://github.com/Seyamalam/pstu_final --skill argent-device-interact-seyamalam

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Automating UI interactions across iOS simulators, Android emulators, and Chromium (CDP) apps normally requires platform-specific tooling and brittle coordinate scripts. This Skill provides a unified tool surface for tapping, swiping, typing, launching apps, and inspecting screen state through one consistent interface. ## Core Features & Use Cases - Unified cross-platform gestures: Tap, swipe, pinch, rotate, drag, and custom touch sequences using normalized 0.0–1.0 coordinates, auto-dispatched to iOS, Android, or Chromium based on the device UDID. - Screen discovery and waiting: Inspect the accessibility tree with describe, block until elements appear with await-ui-element, and batch multi-step flows with run-sequence. - Chromium extras: Manage tabs, cookies, and web storage on CDP targets, and type secrets via {{secret:NAME}} placeholders that never enter the transcript. - Use Case: Automate an end-to-end login flow on an Android emulator — launch the app by bundle ID, tap the email field, type credentials, submit, and wait for the home screen to appear, all in one sequenced call. ## Quick Start Use the argent device interaction tools to list available devices, launch my app on the first ready emulator, and tap the sign-in button.

Frequently Asked Questions about argent-device-interact

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I tap a button in an iOS simulator or Android emulator?

Use the gesture-tap tool with the device UDID and normalized coordinates between 0.0 and 1.0. First run describe to find the element's frame coordinates, then tap that point; never navigate by tapping home-screen icons.

How do I automate typing text and submitting a form on an emulator?

Tap the text field first to focus it, then use the keyboard tool with a text parameter. Since one keyboard call cannot carry both text and a key, batch a text step and an enter key step in a single run-sequence call.

Can the same tools drive an Electron or Chrome browser app?

Yes. Any Chromium runtime exposing a CDP endpoint — an Electron app booted with electronAppPath or a browser launched with --remote-debugging-port — is auto-discovered and driven with the same describe, tap, swipe, keyboard, and screenshot tools.

Why does describe fail or return an empty element list?

Failures occur when the ax-service daemon cannot start, the screen is locked, or content lacks accessibility labels. Check that the simulator is booted, unlock the device, or fall back to screenshot or native-describe-screen with an explicit bundleId.

How do I type passwords without exposing them in the transcript?

Use a {{secret:NAME}} placeholder in the keyboard or paste text parameter. The value is resolved on the tool-server machine from ARGENT_SECRET_ environment variables or .argent/secrets.env files, and the auto-screenshot is skipped.

When should I use run-sequence instead of individual tool calls?

Use run-sequence to batch independent steps like repeated scrolls, form filling, or tap-then-wait-then-tap flows into one call with a single screenshot. Avoid it when a later step depends on observing an earlier step's result.