shirenchuang/agentbro — A macOS Dynamic Island for AI coding agents: Claude Code, Codex, Gemini CLI, Cursor, hooks, skills,…

Description and use

**Brief Summary (2‑3 sentences)**
AgentBro (shirenchuang/agentbro) is an open‑source macOS Dynamic Island that lets AI coding agents such as Claude Code, Codex, Gemini CLI, and Cursor interact with real tools, data, and local agents through a unified Model Context Protocol. Written in Rust, it provides a standard API/SDK/CLI layer for plugging in custom hooks, skills, and agent management, making it easy to expose and consume AI‑driven functionality across development workflows.

**Value**

• **Standardized integration** – By implementing the Model Context Protocol, AgentBro removes the need for bespoke glue code, letting multiple AI assistants share a common interface to tools, files, and services.
• **Rich macOS UX** – The Dynamic Island UI gives developers immediate visual feedback and control over agent actions, improving usability and debugging.
• **Extensible ecosystem** – Hooks, skills, and local agent management are modular, so teams can add new capabilities (e.g., CI pipelines, secret stores) without rewriting the core.

**Practical Adoption Path**

1. **Evaluate the API/CLI** – Clone the repo, run the provided Docker or Homebrew install, and test the sample CLI against a familiar model (e.g., Claude or Gemini).
2. **Create a connector** – Implement a small wrapper (Rust, Python, or Bash) that registers a tool or data source with AgentBro’s protocol; the repository includes templates for common patterns.
3. **Pilot in a dev team** – Deploy the Dynamic Island on developers’ Macs, configure the desired agents, and monitor usage via the built‑in logs.
4. **Scale** – Add custom hooks or a Model Context Protocol server for internal services, and integrate with CI/CD pipelines to automate code generation or review tasks.

**Production Readiness**

• **Activity & Adoption**: 128 stars, 21 forks, recent commits (as of 2026‑06‑24), and growing community usage indicate healthy momentum.
• **Technical maturity**: Core is written in Rust, offering performance and safety; the API surface is stable and documented, with clear SDK/CLI entry points.
• **Risk considerations**: No major metadata or licensing red flags yet, but a final security audit and confirmation of an active maintainer are recommended before mission‑critical deployment. Overall, AgentBro is a strong OSS candidate for a pilot and can be promoted to production once the minor due‑diligence items are cleared.

Problem

shirenchuang/agentbro helps connect AI assistants to real tools and data through a standard protocol.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• connect AI agents to tools
• ship Model Context Protocol servers
• standardize integrations

Evidence

A macOS Dynamic Island for AI coding agents: Claude Code, Codex, Gemini CLI, Cursor, hooks, skills, and local agent management.

Source