Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/web-infra-dev/midscene-skillsnpx agentmods add skills/web-infra-dev/midscene-skills/android-automationWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/web-infra-dev/midscene-skills/android-automation)<a href="https://agentmods.dev/skills/web-infra-dev/midscene-skills/android-automation"><img src="https://agentmods.dev/badge/skills/web-infra-dev/midscene-skills/android-automation.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00159 | $0.04625 |
| Opus 5 | $0.00079 | $0.02312 |
| Sonnet 5 | $0.00032 | $0.00925 |
| Haiku 4.5 | $0.00016 | $0.00462 |
Grade A, and why
android-device-automation scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 308 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Android Device Automation
CRITICAL RULES — VIOLATIONS WILL BREAK THE WORKFLOW:
- Never run midscene commands in the background. Each command must run synchronously so you can read its output (especially screenshots) before deciding the next action. Background execution breaks the screenshot-analyze-act loop.
- Run only one midscene command at a time. Wait for the previous command to finish, read the screenshot, then decide the next action. Never chain multiple commands together.
- Allow enough time for each command to complete. Midscene commands involve AI inference and screen interaction, which can take longer than typical shell commands. A typical command needs about 1 minute; complex
actcommands may need even longer.- Always report task results before finishing. After completing the automation task, you MUST proactively summarize the results to the user — including key data found, actions completed, screenshots taken, and any relevant findings. Never silently end after the last automation step; the user expects a complete response in a single interaction.
Automate Android devices using npx -y @midscene/android@1. Each CLI command maps directly to an MCP tool — you (the AI agent) act as the brain, deciding which actions to take based on screenshots.
What act Can Do
Inside a single act call on Android, Midscene can tap, double-tap, long-press, type, clear text, scroll or swipe in any direction, pull to refresh, drag items, zoom with two fingers, press keys, and use system navigation such as Back, Home, or recent apps while working from the current visible screen.
Prerequisites
Midscene requires models with strong visual grounding capabilities. The following environment variables must be configured — either as system environment variables or in a .env file in the current working directory (Midscene loads .env automatically):
MIDSCENE_MODEL_API_KEY="your-api-key"
MIDSCENE_MODEL_NAME="model-name"
MIDSCENE_MODEL_BASE_URL="https://..."
MIDSCENE_MODEL_FAMILY="family-identifier"
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 308 lines · 159 tokens per session scan A d083eb987751
android-device-automation is a skill published in the GitHub repository web-infra-dev/midscene-skills (307 stars, last pushed 21d ago), licensed MIT. It adds 159 tokens to every session and 4,625 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
android-unit-test
Guide for writing and converting Chromium Android unit tests across host Robolectric (chromejunittests) and on-device (chromepublicunittestapk) suites.
android-test-batching
Guide for auditing and applying batching annotations (@Batch) and AutoReset rules (AutoResetCtaTransitTestRule) to Android javatests in Chromium. Use this skill when optimizing Java instrumentation test runtimes, resolving test state leaks, or migrating tests to batched execution.
java-memory-leaks
How to identify and fix Java memory leaks in Android Chrome using LeakCanary traces.
loadline
Running and analyzing LoadLine 1 and 2 benchmarks on Android using Crossbench. Use when you need to measure page loading performance, evaluate performance-related changes in Chrome, or collect Perfetto traces with realistic Chrome workload.
accessibility-cleanup
Finds common violations of the Android accessibility API in the Clank App Java code and attempts to address them (e.g. hardcoded state change announcements, missing events, focus moving, assertive live regions, etc).
remove-unused-object-overrides
Identify and safely remove unreferenced Java Object overrides (equals, hashCode, toString) in Clank to eliminate dead virtual methods from release DEX bytecode while strictly preserving equals/hashCode symmetry.