Agent skill

AI App Bridge Use

by mobileAiDev in mobileAiDev/ai-app-bridge

使用 AI App Bridge 观察、操作和验证 Android、iOS、Flutter、WebView 或 Web App;适用于真实交互、流程自动化和设备诊断。

Apache-2.0Auto-check passedMobile

Install AI App Bridge Use

skills CLI
$ npx skills add mobileAiDev/ai-app-bridge --skill ai-app-bridge-use -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install mobileAiDev/ai-app-bridge ai-app-bridge-use --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/mobileAiDev/ai-app-bridge.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/ai-app-bridge-use .claude/skills/ai-app-bridge-use && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
ai-app-bridge-use
GitHub stars
227
Token cost
~1.4k tokens
SKILL.md length
422 words
Files
2
Skills in repo
1
Repo updated
First seen
Licence
Apache-2.0

At a glance

使用 AI App Bridge 观察、操作和验证 Android、iOS、Flutter、WebView 或 Web App;适用于真实交互、流程自动化和设备诊断。

  • Tasks that involve Cross-platform mobile apps
  • SKILL.md covers 选择入口, 共享调用合同, 响应与提取 and 目标与动作, plus 4 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

AI App Bridge Use is an agent skill from mobileAiDev/ai-app-bridge. 使用 AI App Bridge 观察、操作和验证 Android、iOS、Flutter、WebView 或 Web App;适用于真实交互、流程自动化和设备诊断。

Its SKILL.md is about 1.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).

It sits in Mobile, covering Cross-platform mobile apps. It works with Android, Flutter, iOS and Model Context Protocol. The repository describes itself as: AI App Bridge is a runtime bridge for AI agents to close the loop on mobile app iteration: inspect the running app, operate UI and WebView surfaces, collect structured runtime… The licence is Apache-2.0.

When your agent uses it

  • Tasks that involve Cross-platform mobile apps

Example prompts

  • “/ai-app-bridge-use”

Requirements

  • Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit a76302e. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

AI App Bridge Use loads about 1.4k tokens when it runs. Until then it costs about 25 tokens; SKILL.md has 422 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~25
When it runs · the whole SKILL.md, loaded when a task matches
~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from mobileAiDev/ai-app-bridge at commit a76302e, republished under its Apache-2.0 licence (© mobileAiDev). 422 words, ~1,394 tokens.

Download SKILL.mdSave it as .claude/skills/ai-app-bridge-use/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
ai-app-bridge-use
description
使用 AI App Bridge 观察、操作和验证 Android、iOS、Flutter、WebView 或 Web App;适用于真实交互、流程自动化和设备诊断。

AI App Bridge Use

选择入口

  • Intent:日常操作和未知流程,观察与 Agent 决策绑定;适合页面探索、系统窗口交互。
  • Script:固定流程和重复运行,自由编写 JavaScript/Python(trusted-local-code),用 ctx.call 调用设备能力;权限声明不是 OS 沙箱。
  • 单次命令:观察、单步动作、安装、权限夹具和诊断可独立调用,无须包装成完整流程。

共享调用合同

MCP 入口是 capabilities 和 run;capabilities 是独立工具,不是 run 的命令。命令参数全部放在 run.arguments,包括目标和 operation,使用当前命令名与 JSON 类型。默认 capabilities 或 domain 查询只取目录;domain 取值为 execution、evidence、core、app、action、flutter、webview、ios、web、diagnostics、advanced;android 是平台不是 domain,平台筛选只用于 Intent decide 的 platform。用 command 查合同,Intent/Script/evidence 加 operation 只取当前操作。Intent decide 可再加实际 platform、provider、action,例如 {"command":"intent","operation":"decide","platform":"android","provider":"native","action":"tap"}。CLI --help COMMAND 接受相同筛选;发现正文上限 96 KiB,宽查询超限时按提示用 command/operation 收窄,不反复读取整个 schema。

每次 run 顶层必填 extract:只需整个小结果时显式 null;树、网络、日志等大结果优先用 regex 或 JS/Python 提取需要的字段。业务参数仍在 arguments;extract/output 不传入设备命令。CLI 使用 --extract null 或 JSON 对象。

CLI 与 MCP 共用独立执行 Runtime、命令合同和 operationId。CLI 响应是一行紧凑 JSON,业务值在 value;MCP 工具正文同样是紧凑 JSON。Script 的调用返回值另见下文。客户端退出不会取消任务,用原 operationId 显式 cancel。取消不撤销已派发效果;版本错误会列出两侧身份,只在确定 Runtime 是待更新一侧时安排显式 stop。

旧 MCP 实例可能与已安装 CLI 不同。缺少 Intent/Script 或参数不匹配时,核对实际入口版本,选用支持当前合同的入口;不要套用旧 batch、工具别名或外层参数。

卡死 / 换 agent: Android 遇到 target_busy、device_ownership_unresolved 或旧动作一直没有终态时,按返回的 recoveryHint 调 device-ownership {operation:"force-stop",serial:"原设备"};CLI 为 ai-app-bridge device-ownership --operation force-stop --serial 原设备 --extract null。这是独立重置入口:无需旧回执,强制结束旧 Host 及共享 Runtime 的任务,重置手机执行器并解除占用。共享这些 Host 的其他任务也会结束;已派发效果记为未知。读取 ownershipReleased/remoteResetPending;手机已重置后重新观察并开始新操作,手机断开时重连后再执行一次。

响应与提取

先看 execution 的原执行事实及 failureStage,再消费 value;Script 内部 ctx.call 仍按 ok/result 处理,不加 extract。control 保留续跑字段和采集覆盖;当前 Script 问题从 control.pendingQuestion 读取,即使 events 被游标过滤也可回答。

默认最终正文预算 96 KiB,可用 output.maxBytes 设为 16–256 KiB。提取失败或超预算不等于动作失败;存在 control.source.persisted:true 时,用 response 的 operation:read、原 ref 和新的 extract 重读,不重发原动作。未保存或留存已过期时没有可恢复的大结果,不捏造 ref。提取会减少交付内容,不承诺减少采集耗时。

提取脚本只使用 ctx.inputs = {kind, response, execution, control},返回严格 JSON;没有 ctx.call。保留实际断言需要的来源身份、时间和状态,不能将选出的几条成功记录当作完整覆盖。完整示例见 docs/RESPONSE_EXTRACTION.md。

目标与动作

  • Android:明确 serial 和 packageName。iOS:deviceId/bundleId;Native Intent 还需原 WDA Runner/session 绑定。Web:从当前连接取得 sessionId/runtimeEpoch/targetId。
  • Intent/Script 的目标带 platform。执行目标保持明确指定的 App/provider;Android 前台匹配、不匹配、未知是独立观测事实,不代表动作成功或失败。更多绑定按平台合同补齐。
  • selector、nodeRef、pageRef 和坐标来自当前观察。多重匹配、过期引用或 reobserve_required 需要重新观察。Android 前台 warning 由 agent 决定继续、观察或调整目标;无需 force。Intent 切 provider 通过 observe,切 Android App 显式传 observationTarget.packageName。
  • 已派发但结果未知(ambiguous)时先观察,避免换 provider 或端点重放动作。保留原 error、message、dispatched 和操作状态。
  • 若使用 freeze,后续操作和结束交付前先 thaw。

Intent

start 提供 goal、显式 target 和所需 provider;默认 supervised。 保留返回的 operationId,读取当前观察后以 decide 提交决策。 decision 包含唯一 decisionId、当前 basedOnRevision 和 agentDecision;act 的 action 遵循该观察的 provider 合同,控件动作使用唯一 selector。 需要刷新或切换 provider 时用 observe,随后使用新 revision。 complete/fail/inconclusive 也需要当前 revision,且不带 action;完成决策不能代替实际结果证据。 supervised 不会仅凭 goal 自动执行:没有决策时停在 waiting_for_decision 直到 timeoutMs。 status 显式给 limit(限制条数,不限制字节);续读历史用上一页的 history.lastSequence 作为 afterSequence,不能用 Script 的 eventSequence。完整 start → decide → status → complete 范例见 COMMAND_CONTRACT.md Execution operation contracts。 安装与权限弹窗命令会返回受监督 Intent,须继续观察和决策;具体收尾条件见对应合同章节。

Show full SKILL.md (148 more words)Show less

Script

start 的 script 内提供 target、language(javascript 或 python),以及 source/sourcePath 二选一。源码入口、权限和 API 按需查 SCRIPT_AUTHORING.md。 ctx.call 返回 envelope:先检查 ok,设备数据在 result;调用失败和 ctx.assert 的 verdict 由源码处理。 用原 operationId 查询 status/wait;每次 waitMs 最多 60000,running 或 finishing 时用上一响应的 eventSequence 作为 afterSequence 继续等待。 waiting_for_agent 是源码的 ctx.askAgent:用 control.pendingQuestion(或本页 agent_question_created 事件)里的 requestId/revision 调 decide,或 cancel;不要等到超时。完整 start → wait → result 范例见 SCRIPT_AUTHORING.md Lifecycle: start, wait, result。 completed 仅说明源码返回并持久化;status/wait 的 resultRef 不是最终值。 完成后调用 script 的 operation:"result" 读取 result、resultRef 和 persisted,检查 representation 及实际断言结果。读取失败保留错误,不从进度事件拼出返回值。

验证与证据

按用户要求的结果选取本轮证据;动作回执、UI 变化、业务结果和证据覆盖分别判断,证据不足保留为 inconclusive。 要保留可移交的过程文件,在 Intent/Script start 时设置新的 recordingDir;已有操作的保留记录可通过 evidence 导出。 evidence export 使用原 operationId 和对应 namespace(intent/script);包含已记录文件需 includeRecordedPayloads:true。保存返回的 manifestSha256,交给离线 evidence verify。 归档校验只证明保留内容的完整性和覆盖范围;缺页、缺引用或已淘汰记录仍需如实报告,不能据此推断业务通过。

按需文档

从 command -v ai-app-bridge 取得入口并解析符号链接;其 bin/.. 是 CLI 发布包根目录。源码仓库中为 desktop/ai-app-bridge-cli。以下路径均相对此包根目录,不相对本技能;先查标题或关键词,只读相关章节。

  • docs/COMMAND_CONTRACT.md:入口与 Runtime 看 Discovery and entrypoints;Intent 看 Execution operation contracts;目标看 Target and dispatch 及对应 iOS/H5/Web 章节;安装/权限看 Installation is an Intent operation / Runtime permission requests use Intent。

  • docs/SCRIPT_AUTHORING.md:首次写脚本看 Start and observe、Lifecycle: start, wait, result(含可运行的 JS/Python 回归范例)、Calls and assertions;命令准入看 Capability selection;采集或暂停需求再读对应章节。

  • docs/EVIDENCE_ARCHIVE.md:需要记录、导出或离线校验时读取,包含文件范围与 coverage 的具体边界。

  • docs/RESPONSE_EXTRACTION.md:单次提取、两语言源码、正文预算、失败后原 ref 重读及退出码。

© mobileAiDev, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in skills/ai-app-bridge-use of mobileAiDev/ai-app-bridge.

  • SKILL.md
  • agents/openai.yaml

Open the folder on GitHubat commit a76302e

Compare with similar skills

AI App Bridge Use next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

AI App Bridge Use compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
AI App Bridge Use this skillmobileAiDev/ai-app-bridge227—~1.4kAutomated safety check: PassApache-2.0
Engine Whats Newflutter/flutter179k—~978Automated safety check: PassBSD-3-Clause
SimdeckNativeScript/SimDeck153—~3.6kAutomated safety check: PassMIT
Detour Onboardingsoftware-mansion-labs/skills291—~2.8kAutomated safety check: PassNone
Aaalice Dev SessionsAaalice233/Aaalice_NAI_Launcher188—~902Automated safety check: PassMIT
flutter_soloud Setupalnitak/flutter_soloud425—~3.6kAutomated safety check: NotesMIT

Similar skills

  • Engine Whats New

    flutter/flutter

    Generates the "what's new" release summary and diff file for changes in the Flutter engine (//engine/src/flutter) between two releases (e.g., 3.47 vs 3.44).

    179k GitHub stars~978 tokensUpdated today
    MobileAuto-check passed
  • Simdeck

    NativeScript/SimDeck

    A skill your agent uses for simulator lifecycle, app install/launch, live viewing, UI inspection, touch/keyboard automation, screenshots, recordings, logs, pasteboard, hardware controls, and…

    153 GitHub stars~3.6k tokensUpdated 28 days ago
    MobileAuto-check passed
  • Detour Onboarding

    software-mansion-labs/skills

    Complete onboarding guide for developers who are new to Detour, the open-source deferred deep linking SDK by Software Mansion.

    291 GitHub stars~2.8k tokensUpdated 12 days ago
    MobileAuto-check passed
  • Aaalice Dev Sessions

    Aaalice233/Aaalice_NAI_Launcher

    管理 Aaalice NAI Launcher 项目专用的 Windows 与 Android Flutter 开发会话。用户提到 PC热重载、安卓热重载、启动/关闭/重建独立命令行窗口、模拟器启动异常或检查双端会话时使用。

    188 GitHub stars~902 tokensUpdated 4 days ago
    MobileAuto-check passed
  • flutter_soloud Setup

    alnitak/flutter_soloud

    Walks through adding the flutter_soloud audio engine to a Flutter app, with per-platform setup, initialization, binary-size and logging options, and output device switching.

    425 GitHub stars~3.6k tokensUpdated today
    MobileAuto-check: notes
  • Otakit

    OtaKit/otakit

    Set up, release, inspect, troubleshoot, or self-host OtaKit for Capacitor OTA updates.

    51 GitHub stars~857 tokensUpdated 2 days ago
    MobileAuto-check passed

Categories

Questions about AI App Bridge Use

What does AI App Bridge Use do?

使用 AI App Bridge 观察、操作和验证 Android、iOS、Flutter、WebView 或 Web App;适用于真实交互、流程自动化和设备诊断。. AI App Bridge Use is an agent skill from mobileAiDev/ai-app-bridge.

When should I use AI App Bridge Use?

AI App Bridge Use fits situations like: tasks that involve Cross-platform mobile apps.

How do I install AI App Bridge Use in Claude Code?

Run `npx skills add mobileAiDev/ai-app-bridge --skill ai-app-bridge-use -a claude-code`. Or copy the skill folder (skills/ai-app-bridge-use in mobileAiDev/ai-app-bridge) into .claude/skills/ai-app-bridge-use in your project. Claude Code loads it when a task matches its description.

How do I install AI App Bridge Use in Codex?

Run `npx skills add mobileAiDev/ai-app-bridge --skill ai-app-bridge-use -a codex`. Or copy the skill folder (skills/ai-app-bridge-use in mobileAiDev/ai-app-bridge) into .agents/skills/ai-app-bridge-use in your project. Codex loads it when a task matches its description.

Can I use AI App Bridge Use in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add mobileAiDev/ai-app-bridge --skill ai-app-bridge-use -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ai-app-bridge-use, .gemini/skills/ai-app-bridge-use, .github/skills/ai-app-bridge-use and .opencode/skills/ai-app-bridge-use in your project.

What does AI App Bridge Use need to run?

SKILL.md names no scripts, command-line tools or credentials: AI App Bridge Use is instructions for the agent only. Our summary lists: Python 3.

Does AI App Bridge Use access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is AI App Bridge Use safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does AI App Bridge Use use?

AI App Bridge Use is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does AI App Bridge Use use?

About 1.4k tokens (SKILL.md is roughly 5.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to AI App Bridge Use?

Skills that share tags, products or a category with AI App Bridge Use: Engine Whats New (flutter/flutter, 179k stars), Simdeck (NativeScript/SimDeck, 153 stars), Detour Onboarding (software-mansion-labs/skills, 291 stars) and Aaalice Dev Sessions (Aaalice233/Aaalice_NAI_Launcher, 188 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains AI App Bridge Use?

mobileAiDev (a GitHub user) maintains it in mobileAiDev/ai-app-bridge, which has 227 GitHub stars. The repository was last updated on October 8, 2026.

Source: mobileAiDev/ai-app-bridge on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.