By greenpolo
Delegate code writing and editing tasks to the Qwen3-Coder AI agent directly from Claude Code CLI via ACP. Run tasks in background or wait for results, resume previous sessions or start fresh, and select from available models for flexible AI-assisted coding workflows.
npx claudepluginhub greenpolo/cc-multi-cli-plugin --plugin qwen
If you have access to multiple AI coding CLIs (Codex, Gemini, Cursor, Copilot, etc.), this plugin lets Claude Code delegate to whichever one is best for the task — without you having to switch tools or run them yourself.
Each CLI is wired up through its native protocol (ACP, ASP, JSON-RPC). This allows you to pick and choose all the best features from each CLIs; like /debug mode from Cursor, /research from Copilot, etc... Sessions, streaming, tool calls, and background jobs all work normally.
Paste into Claude Code:
/plugin marketplace add https://github.com/greenpolo/cc-multi-cli-plugin
/plugin install multi@cc-multi-cli-plugin
/multi:setup
/multi:setup detects which CLIs you have, installs the matching sub-plugins, and wires Exa + Context7 MCPs into each.
Two skills ship with the plugin:
multi-cli-anything — adds ANY CLI (Qwen, Aider, OpenCode, anything that speaks ACP) as a subagent that Claude can invoke at will. Claude scaffolds the new plugin in the marketplace.
customize — change which CLI handles what. "Make Gemini the writer instead of Cursor." Claude does the file edits, reinstalls, and tells you what restarts are needed.
Just ask Claude in plain English. The skills activate automatically.
/gemini:research | Deep research with Gemini's 1M-token context |
/gemini:explore | Fast codebase exploration (Gemini 3 Flash) |
/codex:execute | Hand a plan step to Codex |
/cursor:write | Bulk code writing |
/cursor:plan | Design an approach before coding |
/cursor:debug | Hypothesis-driven debugging |
/copilot:research | GitHub + web investigation |
/copilot:review | GitHub-context code review |
/copilot:plan | Copilot's plan mode |
Claude can also auto-dispatch to these without you typing the command.
All of them are interchangeable, and can be altered to whatever you want using the /customize skill.
These are upstream CLI bugs the plugin can't fully fix, but works around where possible. If you hit something not listed, set ACP_TRACE=1 and check stderr — that reveals which JSON-RPC traffic is or isn't crossing the wire.
Cursor agent acp 2026.04.17 — Terminal/MCP regression. The Terminal (execute) tool can stick at in_progress forever, and MCP tools silently fail. Cursor staff has acknowledged it. The plugin auto-injects a permissive allowlist into ~/.cursor/cli-config.json which keeps simple shell exec working; complex multi-tool runs may still hang. To pin an older Cursor build: export CURSOR_AGENT_PATH=<path-to-old-build>. (forum)
Cursor agent acp — mcpServers ignored. MCP servers passed via ACP session/new are silently dropped in agent acp mode (per Cursor staff). The plugin still wires them — they're consumed correctly by Gemini, Copilot, and Qwen — and falls back to Cursor's own ~/.cursor/mcp.json for Cursor. (forum)
Cursor agent acp — session/request_permission not sent. Cursor's permission gate runs out-of-band against cli-config.json instead of the ACP-standard request flow. The plugin handles both paths (auto-approve and allowlist injection). (forum)
When upstream CLIs fix any of the above, the plugin needs no changes — the workarounds become benign.
Apache 2.0. See NOTICE for upstream credits.
Delegate coding, planning, and debugging to Cursor CLI via ACP. Part of cc-multi-cli-plugin. Requires the `multi` plugin.
Share bugs, ideas, or general feedback.
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex and gemini CLIs when installed) to get diverse perspectives on coding problems
Expert agents for specific programming languages (Python, Go, Rust, etc.)
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Meta-prompting and spec-driven development system for Claude Code. Productivity framework for structured AI-assisted development.