Mcacp

io.github.OortonautPublished by GitHub account Oortonautv0.1.3Updated Oct 24, 2026

MCP-to-ACP bridge — let any MCP client drive ACP coding agents

VerifiedSTDIODesktop onlyDeveloper ToolsAI & ML

Overview

AI-generated overview

An MCP-to-ACP bridge that lets any MCP client spawn, manage and prompt ACP coding agents such as Claude Code, Gemini CLI or Codex.

What you can do
  • Discover, install, uninstall and upgrade ACP agents from the registry, and scan editor configs for existing ones
  • Spawn agents, perform the ACP handshake, shut them down and inspect their status
  • Create, resume, list and close sessions with a working directory and permission policy
  • Send prompts, poll or block for events, cancel prompts and switch agent modes
  • Resolve permission requests interactively through the host or in operator mode
When to use it
Useful when you want to drive several ACP coding agents from one MCP client, compare agents on the same task, or chain them into pipelines where one writes code, another reviews and a third runs tests.
Requirements
Runs locally over stdio as an npm package (mcacp), installed globally or via npx. Needs Node.js and the ACP agents it launches. Optional configuration through mcacp.json, located via the --config flag or the MCACP_CONFIG environment variable, and network access to the ACP agent registry.
Before you install
Installing and uninstalling agents from the registry downloads and removes software on your machine. Permission policies such as allow_all auto-approve agent actions, and in operator mode the host agent decides, so agents may edit files or run commands. Sessions are persisted to disk.

Installation

In SourceWeft

  1. Open Mcacp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

MCACP

Bridge any MCP client to any ACP coding agent.

MCACP is an MCP server that speaks the Agent Client Protocol on behalf of your MCP host. It lets Claude Code, Zed, VS Code, or any MCP-compatible client spawn, manage, and interact with ACP coding agents — turning a single chat window into a multi-agent control plane.

MCP Client (Claude Code, Zed, VS Code, ...)    |    |  MCP (stdio)    v  MCACP    |    |  ACP (stdio, per agent)    v Claude Code  ·  Gemini CLI  ·  Codex  ·  Auggie  ·  Copilot  ·  ...

What can you do with this?

Run any ACP agent from any MCP client. You don't need a dedicated IDE plugin for each agent. Install Gemini from Claude Code. Run Codex from Zed. Mix and match.

Compare agents side by side. Give the same task to Claude Code ACP and Gemini CLI in parallel sessions and compare the results.

Build agent-driven pipelines. One agent writes code, another reviews it, a third runs the tests. MCACP handles the lifecycle and permission flow for all of them.

Control permissions interactively. On hosts that support MCP elicitation (Zed), permission requests pop up as interactive forms. On hosts that don't (Claude Code), they fall back to operator mode where the host agent decides.

Resume where you left off. Sessions persist to disk. Shut down, restart, load_session — the agent picks up where it stopped.

Discover and install agents from chat. Browse the ACP registry, install agents, check for upgrades — without leaving your conversation.

Quick start

Install

bash
npm install -g mcacp

Or run directly:

bash
npx mcacp

Add to your MCP client

Claude Code (~/.claude/settings.json):

json
{  "mcpServers": {    "mcacp": {      "command": "npx",      "args": ["-y", "mcacp"]    }  }}

Zed (settings.json):

json
{  "context_servers": {    "mcacp": {      "command": { "path": "npx", "args": ["-y", "mcacp"] }    }  }}

Then from chat:

"Search for available agents"        →  registry_search"Install the Gemini agent"           →  agent_install"Initialize it and start a session"  →  initialize + new_session"Ask it to refactor auth"            →  prompt_start + prompt

MCP tools

MCACP exposes 22 tools organized into five groups.

Registry

ToolDescription
list_installed_agentsList locally installed agents
registry_searchSearch the ACP agent registry
agent_installInstall an agent from the registry
agent_uninstallRemove an installed agent
agent_check_upgradesCheck for available upgrades
discover_agentsScan editor configs (Zed, JetBrains) for agents
reload_configReload mcacp.json from disk

Lifecycle

ToolDescription
initializeSpawn an agent process and perform the ACP handshake
shutdownGracefully shut down an agent

Sessions

ToolDescription
new_sessionCreate a session with working directory and permission policy
load_sessionResume a persisted session
list_sessionsList stored sessions for an agent
close_sessionClose a session (preserved for later resume)

Interaction

ToolDescription
prompt_startSend a prompt (returns immediately)
promptBlock until events are available
prompt_eventsNon-blocking poll for events
grant_permissionResolve a pending permission request (operator mode)
request_permissionSend an MCP elicitation to the host for a permission decision
cancelCancel an in-progress prompt
set_modeSwitch agent operating mode

Status

ToolDescription
list_running_agentsList all agents with status and sessions
get_agent_statusDetailed status of a running agent
set_agent_statusSet status text for an agent

Permission policies

Each session can use a different policy:

PolicyBehavior
allow_allAuto-approve all agent actions
deny_allBlock all tool use
elicitForward as MCP elicitations to the host; falls back to operator if unsupported
operatorQueue as events — the host agent decides via grant_permission

Operator flow

Orchestrator                   MCACP                      ACP Agent    |                            |                            |    |-- prompt(sessionId, ...) ->|                            |    |                            |-- session/prompt --------->|    |                            |<- request_permission ------|    |<- {pendingPermission: ...} |                            |    |                            |                            |    |-- grant_permission ------->|                            |    |   {toolCallId, optionId}   |-- permission outcome ----->|    |                            |<- prompt result -----------|    |<- {stopReason, updates}    |                            |

Configuration

MCACP looks for mcacp.json in these locations (first found wins):

  1. --config flag or MCACP_CONFIG env var
  2. .mcacp/mcacp.json in the current directory
  3. ~/.config/mcacp/mcacp.json
jsonc
{  "registries": ["https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json"],  "defaultPermissionPolicy": "elicit",  "agent_servers": {    "my-custom-agent": {      "command": "node",      "args": ["my-agent.js"]    }  }}

See docs/configuration.md for the full schema.

Building from source

bash
git clone https://github.com/Oortonaut/mcacp.gitcd mcacpnpm installnpm run buildnode dist/index.js

License

Apache-2.0

Source: README.md at commit cf3f452

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.1.3LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts