by eggbrid2
Open Android AI agent runtime for phone control, app automation, VLM screen reading, skill routing, mini apps, and Mihomo VPN workflows.
# Add to your Claude Code skills
git clone https://github.com/eggbrid2/mobileClawLast scanned: 5/30/2026
{
"issues": [],
"status": "PASSED",
"scannedAt": "2026-05-30T16:21:08.497Z",
"npmAuditRan": true,
"pipAuditRan": true
}mobileClaw is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by eggbrid2. Open Android AI agent runtime for phone control, app automation, VLM screen reading, skill routing, mini apps, and Mihomo VPN workflows. It has 406 GitHub stars.
Yes. mobileClaw passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/eggbrid2/mobileClaw" and add it to your Claude Code skills directory (see the Installation section above).
mobileClaw is primarily written in Kotlin. It is open-source under eggbrid2 on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh mobileClaw against similar tools.
No comments yet. Be the first to share your thoughts!
MobileClaw is an Android AI agent lab. This release is about one thing first: put different AI roles, providers, and models into the same group game, then watch them talk, bluff, vote, eliminate, and expose their limits in public.
This is not a normal group chat. It is a phone-sized AI arena.
Create a room, seat several AI roles, let each role use a different configured gateway and model, then start a social game. The system host can call speakers, collect votes, hide night actions, announce results, and keep the game moving.
The fun is watching the models collide:
Use it for traitor-hunt games, debate rooms, model-vs-model tests, custom elimination games, or pure "AI arena watching" when you just want to see different agents fight for the table.
Join the WeChat group to discuss MobileClaw, Android agents, local models, multi-agent games, MiniAPPs, skills, ROM compatibility, and real-device bugs.
This WeChat group QR code is valid until July 16, 2026.
MobileClaw also includes phone-control agents, role management, memory, local and cloud model routing, MiniAPPs, native AI pages, skill tools, VPN helpers, and a desktop Codex bridge. The group-game release sits on top of that larger Android agent runtime.
Useful docs:
git clone https://github.com/eggbrid2/mobileClaw.git
cd mobileClaw
./scripts/assemble_debug.sh
Debug APK:
app/build/outputs/apk/debug/app-debug.apk
python3 scripts/pgyer_release.py build-upload \
--gradle-task assembleDebug \
--notes "MobileClaw AI arena group play release"
Keep Pgyer secrets in local.properties, .pgyer.env, or environment
variables. Do not commit them.
MIT. See LICENSE.