
Devilge
io.github.YerckoPublished by GitHub account Yerckov0.2.3Updated Oct 24, 2026
Android/KMM MCP server: logcat, ADB device control, Compose previews, network, Maestro flows.
Overview
Android/KMM development MCP server that lets an assistant build, install, drive and observe an app on a dev device.
- What you can do
- Inspect project structure and Compose previews, run any Gradle task and parse compile and test errors. List devices and read logcat, app errors and network calls. Drive the running app: tap, type text, send keys, swipe, take screenshots and dump the UI. Locate elements by text or resource ID and wait for the screen to settle. Launch, force-stop, clear data, install APKs and run instrumented tests. Replay navigation as Maestro YAML flows, and batch several tool calls into one request.
- When to use it
- For developers building an Android or KMM app who want an assistant to build, install, launch and drive the app on a dev emulator or test device. Useful when iterating on logcat, app errors and network calls, or replaying recurring UI navigation.
- Requirements
- Runs locally as an npm package; needs Node.js, the Android SDK adb binary and an attached device or emulator. DEVILGE_ANDROID_PROJECT_ROOT must be set to the project's absolute path; optional variables include DEVILGE_DEFAULT_DEVICE_SERIAL and DEVILGE_MAESTRO_BIN_PATH. Maestro flow tools need Maestro installed separately or they report it missing.
Installation
In SourceWeft
- Open Devilge in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
devilge
Model Context Protocol (MCP) server that lets an AI assistant — typically Claude — develop, drive, and observe an Android / KMM app end-to-end.
devilge exposes 33 tools that cover the full inner loop: read the project, build it, install it, launch it, drive its UI, capture errors and network traffic, run tests. Anything an LLM coding agent would otherwise have to ask the user to do manually.
What's inside
215 unit tests in Vitest, all green. Strict TypeScript (strict, noUncheckedIndexedAccess).
Batching for fewer permission prompts
Hosts that confirm every tool call (Claude Desktop, Cowork) can become noisy when the agent walks through multi-step UI flows. devilge_batch collapses a sequence into a single MCP call so the user approves once for the whole sequence.
Rules: capped at 20 actions per call; cannot be nested; cannot include destructive tools (devilge_clear_app_data, devilge_install_apk) — those always require their own dedicated prompt; stops on the first error and reports which step failed.
All tools also expose MCP annotations (readOnlyHint, idempotentHint, destructiveHint, openWorldHint). Hosts that respect annotations can auto-approve safe reads and prompt only on state-changing tools.
Architecture
Clean / hexagonal layering, every concern replaceable in isolation:
src/server.ts is the composition root: it constructs every concrete dependency and wires them into an McpServer. Nothing else in the codebase performs construction.
Security model
devilge runs on a developer's machine, exposes mutating tools to an LLM, and shells out to ADB and (optionally) Gradle / Maestro. It assumes the operator connects only to a dev emulator or wiped test device, never a personal device with logged-in apps.
- Project sandbox. All filesystem reads/writes resolve, after symlink resolution, to paths inside
DEVILGE_ANDROID_PROJECT_ROOTor the configured outputs / flows roots. Any escape throwsSecurityError. - No shell. Every external process (
adb,gradlew,maestro) is spawned withshell: falseand an argv array. Arguments are never concatenated into strings. - Strict argument allowlists. Device serials, logcat tags, package names, activity names, deep links, Gradle tasks, Maestro flow names and env-var keys all pass through
CommandSanitizerregex allowlists before reaching argv. - Resource caps. Logcat capped (5000 line ceiling), ADB stdout cap 8 MiB, Gradle output cap 256 KiB ring buffer, file scans bounded, screenshot timeout 15 s, instrumented-test timeout 30 min default cap.
- Symlinks ignored. Walkers never follow symlinks.
- Logs go to stderr only. Stdout is reserved for the MCP JSON-RPC transport.
- Errors are sanitized. Only
DevilgeErrorsubclasses surface their messages. Unexpected exceptions become an opaqueINTERNAL_ERROR. - Header redaction.
Authorization,Cookie,Set-Cookie,X-API-Keyand other well-known sensitive headers are redacted fromget_network_callsoutput. - Maestro
runScript:denied by default. YAML flows containingrunScript:blocks (which execute JS) are rejected unless the operator explicitly setsDEVILGE_ALLOW_FLOW_SCRIPTS=true. - No frontmost-app check. devilge does NOT verify that the targeted package matches the device's foreground app. That check is a deployment concern — keep your dev device clean.
The MVP does include write tools (input automation, app install, data wipe, force-stop). These are gated to a dev-only device by operator policy, not by server logic.
Setup
Run the tests:
Verify the build:
Connecting to Claude Desktop
Add an entry to your claude_desktop_config.json (path varies by OS):
Restart Claude Desktop. The 33 tools should appear in the tool picker.
Local smoke test (MCP Inspector)
Configuration reference
Optional integrations
Maestro (flows)
Maestro is optional. Without it installed, every other devilge tool keeps working. The three flow tools (run_maestro_flow, list_maestro_flows, validate_maestro_flow) register unconditionally and return MAESTRO_NOT_INSTALLED when the binary isn't found.
To enable:
Restart the inspector. Flows go in <project>/devilge-flows/<name>.yaml. Example:
Then:
MAESTRO_DISABLE_ANALYTICS=true is injected automatically. runScript: blocks are denied unless you opt in.
Headless Compose preview rendering (recipe, no devilge tool)
You can render @Preview Composables to PNG without launching the full app. devilge does NOT add a dedicated tool for this — the existing run_gradle_task plus the official Google plugin cover it cleanly, and adding a wrapper would make us depend on Gradle conventions that vary by project.
This is completely optional: if you don't add the plugin, devilge runs unchanged. You only lose this specific workflow.
To enable in your project, add to your Compose module's build.gradle.kts:
And in gradle/libs.versions.toml:
Once the plugin is in place, render previews from the LLM via the existing tool:
The PNGs land under composeApp/build/outputs/screenshotTest/.... Claude can then read them via its Read tool to verify visual output without installing the app.
Recommended inner-loop workflow
For recurring navigation paths (login, search, etc.), capture once as a Maestro flow and replay with one tool call.
Backlog (not committed, lowest priority)
- Compose Live Edit MCP — true HMR for Android Compose. Major project (~2-3 months MVP, JVMTI agent + bytecode transformation + protocol). Waiting for JetBrains' Compose Hot Reload to land for Android first; the wrapper would be ~1-2 weeks.
pull_room_database— pull Room SQLite from device, expose readonly queries. Useful for inspecting cached state.pull_anr_traces+deobfuscate_stacktrace— diagnose runtime hangs and ProGuard-mapped release crashes.dumpsys_meminfo/dumpsys_gfxinfo/measure_cold_start— runtime performance metrics.describe_compose_codebase— structural map of the project (data classes, XML resources, color literal frequencies, composables) so the LLM doesn't have to grep at session start. Considered, deferred until proven necessary in real use.
License
MIT
Source: README.md at commit 17d8b50
Tools
0Version history
1- v0.2.3LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts