Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add TiagoDanin/Android-Debug-Bridge-MCP --skill adb-mcpgit clone --depth 1 https://github.com/TiagoDanin/Android-Debug-Bridge-MCPWrote 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/tiagodanin/android-debug-bridge-mcp/adb-mcp)<a href="https://agentmods.dev/skills/tiagodanin/android-debug-bridge-mcp/adb-mcp"><img src="https://agentmods.dev/badge/skills/tiagodanin/android-debug-bridge-mcp/adb-mcp/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tiagodanin/android-debug-bridge-mcp/adb-mcp"><img src="https://agentmods.dev/badge/skills/tiagodanin/android-debug-bridge-mcp/adb-mcp.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high System Prompt Leakage · line 110 Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
- medium MCP Rug Pull · line 175 npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.Fix: Pin the version: npx @scope/[email protected]
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 | $0.00100 | $0.02182 |
| Opus 5 | $0.00050 | $0.01091 |
| Sonnet 5 | $0.00020 | $0.00436 |
| Haiku 4.5 | $0.00010 | $0.00218 |
Grade B, and why
adb-mcp scanned grade B with 1 finding 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 5d 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.
Asks the agent to reveal its instructionsmediumSystem prompt leakage
Directions to print, repeat or translate the system prompt extract configuration the operator did not intend to expose.
On an emulator, `emu_finger_touch` answers a fingerprint prompt (`finger_id` How it starts
The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ADB MCP Tools
The android-debug-bridge MCP server exposes an Android device to the agent as
tools. Every device-facing tool takes an optional device argument — a full
serial, a serial prefix, a transport id or a model name — and defaults to
ADB_SERIAL or the only connected device.
The same engine is available as the adb-agent CLI — see the adb-cli skill.
Reach for the CLI when a flow is a straight sequence of steps you want to chain
in one shell call; reach for these tools when you want images in the transcript
and per-step reasoning.
Before the first action
Call list_devices. If it comes back empty, tell the user to start an emulator
or connect a device with USB debugging enabled — do not keep trying other tools,
they will all fail the same way. If it lists more than one device, the result
says whether there is a default target; when there is none, pick one and pass
device on every later call rather than letting each tool fail in turn.
Then call device_info once. Screen size, Android version and emulator status
shape everything after it: which coordinates are valid, whether a permission
model applies, whether a hardware key exists.
The core loop
Read, act, verify:
capture_ui_dumporui_findto see what is on screen.ui_tapwith a label, notinput_tapwith a guess.ui_wait_forbefore touching the next screen, then repeat.
ui_tap re-reads the tree, ranks matches on text, content-description and
resource-id, prefers clickable and enabled elements, and taps the best one.
Fixed coordinates break on a different device; labels usually do not.
ui_tap also refuses to tap through an overlay. The tree reports full logical
bounds with no notion of what is painted on top, so a banner under a bottom bar
still claims those pixels — tapping its center fires the bar and reports
success for the wrong element. The tool looks for an uncovered point inside the
element first, and errors with the name of the blocker when there is none. Pass
force: true to override once you know what is on top.
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.
- 5d ago Changed · +38 lines scan A → B b89e56e7ad5e
- 9d ago First seen · 144 lines · 100 tokens per session scan A 4b86877585d6
adb-mcp is a skill published in the GitHub repository TiagoDanin/Android-Debug-Bridge-MCP (33 stars, last pushed 5d ago), licensed MIT. It adds 100 tokens to every session and 2,182 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it B with 1 finding (asks the agent to reveal its instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
ios-fix
Autonomous iOS bug fixer. (gstack).
google-mobile-ads-validate
Validates a project's Google Mobile Ads (GMA) SDK integration for iOS, Android, or Unity projects. Use when conducting a full pre-launch audit of an app that integrates GMA SDK or when validating any individual GMA SDK integration checks, such as when validating ad unit IDs and ad formats, SKAdNetwork IDs, mediation…
competition-ios-runtime
Internal downstream skill for ctf-sandbox-orchestrator. CTF-sandbox workflow for IPA runtime analysis, Frida hooks, Objective-C or Swift method tracing, Keychain inspection, SSL pinning bypass, URL scheme handling, and iOS request-signing recovery. Use when the user asks to hook an IPA, trace Objective-C or Swift…
swiftui-performance-audit
SwiftUI performance: render, scroll, CPU/memory, updates, layout, Instruments.
android-tombstone-symbolication
Symbolicate the .NET runtime frames in an Android tombstone file. Extracts BuildIds and PC offsets from the native backtrace, downloads debug symbols from the Microsoft symbol server, and runs llvm-symbolizer to produce function names with source file and line numbers. USE FOR triaging a .NET MAUI or Mono Android app…
node-connect
Diagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps. Use when QR/setup code/manual connect fails, local Wi-Fi works but VPS/tailnet does not, or errors mention pairing required, unauthorized, bootstrap token invalid or expired, gateway.bind, gateway.remote.url, Tailscale…