Agent Herder

MCP control center for coding agents.

Monitor, inspect, and coordinate AI coding sessions from one MCP server: OpenCode, Claude Code, Codex CLI, Qoder, and ZCode.

Русский · 简体中文

Start in 30 seconds

Run it without cloning a repository:

npx -y agent-herder

Add the same command to any MCP client:

{
  "mcpServers": {
    "agent-herder": {
      "command": "npx",
      "args": ["-y", "agent-herder"]
    }
  }
}

Start the harness you want to observe first. For OpenCode, that means:

opencode serve

Related MCP server: all-agents-mcp

What it does

Monitor, inspect, and coordinate. Agent Herder gives your MCP client one control plane for coding-agent sessions that normally live in separate tools.

• See running, idle, stopped, and waiting sessions in one list.

• Find a session's parent and children without guessing IDs.

• Send, resume, stop, steer, or recover sessions where the harness supports it.

• Review permissions, switch models, and inspect worktrees.

Typical request:

Find the parent of this session, list its children, then export the raw transcript of the child that is currently working on the bug.

Why people use it

Use Agent Herder when you need an MCP server for OpenCode, a Claude Code session manager, Codex CLI transcript search, ZCode app-server control, or one dashboard for several AI coding agents. It is especially useful for parent/child agent workflows, parallel coding tasks and session recovery.

Supported harnesses

Core MCP tools

export_transcript copies the adapter-owned raw transcript for one session and its in-workspace parent/child lineage, then always returns a short navigation card. It deliberately does not rank or compress the conversation: use the normal filesystem tools available to the agent to inspect exactly the slice it needs.

Claude Code autopilot is packaged under .claude-plugin/: /autopilot toggles the exact current session, the native Stop hook asks the shared AI judge to continue or finish, and ambiguous decisions appear as NoticePlace/web buttons. See the autopilot guide.

new_or_resume identifies a session by the exact tuple (harness, canonical CWD, name). It reuses one matching native session or creates it, then delivers one message. Multiple exact matches fail closed before delivery. queue reports native acceptance; sync waits for the adapter response. Event idempotency belongs to the webhook/control plane that calls Agent Herder.

Requirements

• Node.js 22+ and npm.

• At least one supported harness installed and available in PATH.

• OPENAI_API_KEY for Codex when the Codex app-server requires it.

Configuration

The common switches are:

Example for Qoder:

export ENABLE_QODER=true
export QODER_CWD=/path/to/project
npx -y agent-herder

Develop locally

npm ci
npm test
npm run build
npm run inspect

The local stdio entrypoint is dist/index.js.

The optional web UI runs on loopback:

export AGENT_HERDER_WEB_PORT=8787
npm start

Open http://127.0.0.1:8787/. For a persistent ACP profile, set ACP_AGENT_COMMAND, ACP_AGENT_ARGS as a JSON array, and ACP_AGENT_PROFILE before starting the server. Set AGENT_HERDER_WEB_PORT to the loopback upstream expected by your reverse proxy (the managed agent.bezrabotnyi.com deployment uses 18787). The matching user-service template is deploy/systemd/agent-herder.service.

FAQ

Does Agent Herder replace my coding agent? No. It connects your MCP client to the sessions owned by OpenCode, Claude Code, Codex, Qoder, or ZCode.

Does export_transcript load everything into the model? No. It writes the raw source to a CWD-scoped archive and returns only the permanent navigation card. The card shows sed for the first lines, tail for the last lines, and rg examples for literal text, regex, or a timestamp.

Can I use only one harness? Yes. Disable adapters you do not run with the ENABLE_* variables.

Transcript archive

export_transcript atomically copies the adapter-owned raw source under the MCP process CWD and writes a lineage manifest beside it. The archive never falls back to display-only text. Its manifest names the source, format, timestamp coverage, and whether the adapter can prove completeness. Parent or child sessions outside the MCP CWD are recorded as excluded, not copied.

The navigation card is returned for every export, regardless of transcript size. It shows how to view the first or last lines and how to use literal, regular-expression, or timestamp searches. OpenCode and ACP archives are explicitly partial until their upstream APIs offer a verified complete-history source.

License

MIT