Agentic Flow MCP
An MCP server and CLI coordinate specialized agents, retain working patterns and connect tasks with project hooks.
- 838 ★ on the repository
- Last commit
Broad access
Why this level
- Initialization changes project settings and hooks; workers can run commands and edit code.
- Stored memory and metrics can expose project information to connected components.
Install
- Works in
Manual install
npx agentic-flow initAdds settings and hooks to the current project. This unpinned command uses the current package; the saved manifest is version 3.1.0.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Initialization changes project settings and hooks; workers can run commands and edit code.
- Stored memory and metrics can expose project information to connected components.
This is third-party code. Review the repository files before installing.
What it does
Agentic Flow MCP provides agent-coordination tools. Its CLI routes requests, starts background jobs and reads their results. Work is connected with stored patterns from previous project activity. Initialization adds integration infrastructure to the workspace. Hooks can receive edit and command events to supply context and record outcomes. A separate MCP process exposes these capabilities to Claude Code.
Who it is for. For developers who need coordinated roles and reusable context inside a coding project.
Good fit when
- Choose an agent role for a new task.
- Prepare codebase context for later work in a large project.
- Launch background analysis and retrieve results later.
- Carry successful and failed edit outcomes into subsequent tasks.
Not a fit when
- You need passive reference material without project configuration changes.
- Hooks and background processes are prohibited.
- You expect guaranteed savings for every task from project marketing claims.
Example request
Show what Agentic Flow would add to the test project and propose minimal coordination without launching background work.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
Initialization adds settings and hooks to the current project, and `--force` can overwrite them: save your changes in Git first. Selected external models may require keys and paid calls that consume their account funds. Memory and metrics contain code and task outcomes, so inspect those files before sharing the repository. Background workers execute commands and edit code; connecting MCP does not by itself restrict them to reading files.
How to disable. Stop MCP and background processes, then remove added project hooks. Inspect generated memory and agent directories before cleaning them.
FAQ
Which version was checked?
Version 3.1.0 in the current package manifest. Pin commands to it for repeatable setup.
How do I route a task?
Use `npx agentic-flow@3.1.0 hooks route "implement user authentication"` with your description.
How do I inspect background jobs?
Run `npx agentic-flow@3.1.0 workers status`. Retrieve output with `npx agentic-flow@3.1.0 workers results`.
Can I start with empty directories?
Yes. Initialization supports `--minimal` for a minimal structure without the full library.
Which events can hooks observe?
Before and after edits, before commands and at session start. Inspect the resulting configuration entries after installation.
Discussion
No questions or reviews yet.
Related
MCP serversUseAgent’s MCP server sends tasks to remote agents, collects their results and records the runs in an organization’s ledger.
An MCP server that inserts a human feedback step into agent tasks: a web or desktop window for text, images and prompts
An MCP server that turns a request into a step plan with dependencies and keeps context across agent sessions
MCP Agent Mail
mcp_agent_mail
A mail coordination layer for coding agents: agents have memorable names, inboxes and outboxes, history search and voluntary file reservations
A universal MCP CLI client from Apify: persistent sessions, OAuth, JSON output for scripts and shell access to any MCP server for agents
A bridge between local CLI agents: Codex, Cursor or Claude Code delegate work to Antigravity, Grok, Kimi and seven more.
README in short
The distribution separates routing, working memory and background-worker commands. Documentation covers recording work trajectories and reusing discovered patterns.