CodeAgentsMobile

作者 eugenepyvovarov已验证

Mobile client for Claude Code.

133
Stars
13
Forks
Swift
语言
2026/8/23
添加时间

⚠️ 第三方软件声明

本 Skill 为第三方开源软件,独立托管于 GitHub。SkillTip 仅为信息目录,不控制或维护底层仓库。所显示的安全检查为自动化且范围有限,安装前请自行审查源码。

阅读服务条款

安装

添加到你的 Claude Code skills 目录:

# Add to your Claude Code skills
git clone https://github.com/eugenepyvovarov/CodeAgentsMobile

快速入门

使用 CodeAgentsMobile 等 Skills 的指南。

安全报告

已验证

上次扫描:—

{
  "status": "PASSED",
  "issues": []
}

README.md

CodeAgents Mobile

An mobile client for OpenCode on iOS. Connect to any Linux SSH server and chat with your coding agent on the go.

Legacy Claude Code / Claude Proxy chat is retired: existing projects auto-migrate to OpenCode (runtime, MCP servers, optional rules/keys). The agent daemon (:8787) remains for background tasks and push.

TestFlight Beta

https://testflight.apple.com/join/eUpweBZV

AppStore

https://apps.apple.com/app/codeagents-mobile/id6748273716

Screenshots

Screenshot 1 Screenshot 2 Screenshot 3 Screenshot 4 Screenshot 5 Screenshot 6 Screenshot 7 Screenshot 8

Features

  • A rebuilt chat experience: faster, clearer messages, and a smoother overall feel
  • Background tasks: start longer actions and keep using the app while they run
  • Skills support: add, manage, and use skills to expand what your agent can do
  • Better file tools: browse, preview, create/edit, upload, and share files
  • Agent controls: rules, permissions, scheduled tasks, and custom environment variables
  • More provider choices: Anthropic, Z.ai, Minimax, or Moonshot
  • Provision servers with Digital Ocean / Hetzner (auto SSH key setup + Claude Code install)
  • MCP servers support
  • Connect to remote servers via SSH

Requirements

  • iOS 17.0+
  • Xcode 15+
  • Swift 5.9+

Getting Started

# Clone the repository
git clone [repository-url]

# Open in Xcode
open CodeAgentsMobile.xcodeproj

# Build and run
# Select your target device/simulator and press Cmd+R

Automation

OpenCode automation uses the repository-owned scripts in scripts/ while the workflow YAML is generated by the automation controller.

  • scripts/ci.sh validates the app with XcodeBuildMCP 2.5.0+ structured JSON diagnostics by running deterministic CodeAgentsMobileTests simulator unit tests; live SSH integration suites are skipped unless run explicitly.
  • scripts/ios-simulator-evidence.sh captures simulator screenshots and demo video for native visual validation and demo evidence.
  • scripts/artifact.sh builds a review simulator .app.zip by default and emits the controller published_artifacts JSON contract; set OPENCODE_ARTIFACT_MODE=testflight to publish a production build through asc publish testflight. Production artifact automation runs this mode with Phase app 6abec477-1930-4e2e-bd18-554059958ccd and emits release metadata so the controller can attach the IPA and checksum to the matching Gitea Release.
  • VERSIONS.TXT stores the release marketing version. Production builds use the Gitea run number as the release build number when available. scripts/artifact.sh injects both values into the app bundle, and Settings shows them together so a reported installation can be identified precisely.

The runner must provide Xcode, XcodeBuildMCP 2.5.0 or newer, and the asc CLI for TestFlight publication. XcodeBuildMCP failures print the detected version plus returned build/runtime/OS log paths when available; scripts do not upgrade the tool automatically. If DEVELOPER_DIR is not set, the scripts default to /Applications/Xcode.app/Contents/Developer when present. TestFlight mode also expects App Store Connect credentials through ASC_APP_ID, TESTFLIGHT_GROUP, ASC_KEY_ID, ASC_ISSUER_ID, and either ASC_PRIVATE_KEY_P8 or ASC_PRIVATE_KEY_FILE. Firebase-enabled builds also require MobileCode/GoogleService-Info.plist; CI can provide it through GOOGLE_SERVICE_INFO_PLIST_BASE64.

Developer Diagnostics

Release and TestFlight builds use Firebase Crashlytics for pseudonymous crash diagnostics. Firebase Analytics is not linked, and the app does not attach user IDs, custom logs, prompts, URLs, paths, or credentials. Debug builds disable Crashlytics collection. See PRIVACY_POLICY.md for the data and retention disclosure.

Debug builds emit lightweight chat recovery timing lines with the [ChatRecoveryTiming] prefix when chats are reopened or resumed. These logs are intended to separate local SwiftData load, Claude→OpenCode migration, OpenCode hydration, persistence saves, media prefetch, MCP fetch, and resume-streaming costs before making recovery behavior changes.

Chat is OpenCode-only. Reopen is local-first: persisted SwiftData messages render before remote recovery. Legacy Claude Proxy projects are promoted by ClaudeToOpenCodeMigrationService (idempotent; does not block showing local messages). The initial OpenCode recovery fetch is bounded and diffed against stored message/part hydration anchors; full-session refresh can run later in the background when more history may exist.

Chat open also defers unrelated remote startup work until after local messages and required active-session recovery are complete. MCP server refresh, managed rules setup, and broad CodeAgents UI media prefetch should run from explicit MCP/tool UI actions, just-in-time send/action paths, or the project-scoped post-ready background queue; project switches cancel that deferred work before it can update stale chat state.

Focused tests: ClaudeToOpenCodeMigrationTests, ChatDeferredStartupTests, ChatMessageLifetimeTests, and OpenCode hydration/session suites under MobileCodeTests/.

Timing metadata is limited to runtime names, project identifiers, operation labels, elapsed milliseconds, statuses, booleans, and counts. Do not add prompts, message text, raw payloads, credentials, URLs, project paths, attachment paths, or file contents to these log fields.

Architecture

Built with SwiftUI and SwiftData following MVVM pattern. Icon - https://tabler.io/icons/icon/brain, edited here https://icon.kitchen

License

This project is licensed under the Apache License, Version 2.0 - see the LICENSE file for details.

常见问题

What is CodeAgentsMobile?

CodeAgentsMobile is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by eugenepyvovarov. Mobile client for Claude Code. It has 133 GitHub stars.

Is CodeAgentsMobile safe to use?

Yes. CodeAgentsMobile 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.

How do I install CodeAgentsMobile?

Clone the repository with "git clone https://github.com/eugenepyvovarov/CodeAgentsMobile" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is CodeAgentsMobile written in?

CodeAgentsMobile is primarily written in Swift. It is open-source under eugenepyvovarov on GitHub, so you can review or fork the full source.

Are there alternatives to CodeAgentsMobile?

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 CodeAgentsMobile against similar tools.

评论 (0)

暂无评论,成为第一个分享想法的人!

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情

claude-code

by anthropics

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

120,03119,897Shell
AI 智能体
查看详情

开发者还喜欢

基于喜欢此 Skill 的开发者投票和收藏

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

n8n

by n8n-io

12

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

201,88160,308TypeScript
MCP 服务器apisai-tools
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情