SkillFoxx

flow CLI

Workflow manager CLI: scripts, secrets and automation of many projects behind one interface, plus an MCP server for agents.

  • 139 ★ on the repository
  • Last commit
  • Apache-2.0
CLI

Broad access

Why this level

  • Through MCP the agent runs commands on your machine with user permissions and gets vault secrets.
  • Workflows can make network requests and ship code: review what the `.flow` files contain.
All reasons and checks

flowexec/flowAuthor:flowexec

Install

Manual install

curl -sSL https://install.flowexec.io | bash

The first Quick Start command in the README.

Access and checks

How we review

Broad accessThe tool can change data, run commands or handle secrets.

  • Through MCP the agent runs commands on your machine with user permissions and gets vault secrets.
  • Workflows can make network requests and ship code: review what the `.flow` files contain.

This is third-party code. Review the repository files before installing.

What it does

flow is a workflow manager for people who work across many projects. Each repository is registered as a workspace, and scripts, builds and automation are described in `.flow` files and run from any folder with commands like `flow deploy staging`. `flow browse` opens a terminal interface listing every workflow across all projects. Secrets live in encrypted local vaults and are injected at run time instead of being written into code. Workflows run serially, in parallel or conditionally, make HTTP requests, ask questions and start background jobs. The built-in MCP server `flow mcp` lets Claude Code, Cursor and other agents run those workflows.

Who it is for. Developers and DevOps engineers with a dozen repositories where each has its own Makefile, npm scripts and undocumented flags.

Good fit when

  • Each project builds and ships its own way and the command order has to be recalled.
  • An agent should run deploy and backup through described workflows instead of inventing commands.
  • Script secrets sit in plain files and should move into a vault.
  • A new project starts from a template with a ready workflow layout.

Not a fit when

  • A single project with a couple of scripts: a Makefile or npm scripts are enough.
  • An agent must not run commands on your machine, even predefined ones.
  • Commands run once and nobody plans to reuse them.

Example request

List all workflows of the api project and run `flow deploy staging`, but stop before the health check step and ask me.

Limitations

Workflows an agent starts through MCP run with your permissions and receive vault secrets, so do not register workspaces with destructive commands without review. The recommended install pipes a script through `curl`; read it beforehand. The `.flow` format, templates and vault backends are covered on the documentation site, while the README gives only an overview. The MCP server tool list is not given in the README, look in the documentation.

How to disable. Remove the MCP entry from agent settings, delete workspaces with flow commands and remove the binary. Secret vaults stay on disk, clean them separately.

FAQ

How do I connect flow to Claude Code?

Add an MCP server over stdio with command `flow` and argument `mcp`. The same entry is in the repository `.mcp.json`.

Where are secrets stored?

In encrypted local vaults. A workflow receives them at run time through `secretRef` and `envKey`, and values never sit in the workflow file.

What counts as a workspace?

Any repository you registered with `flow workspace add`. Its workflows run from any folder.

How do I try it quickly?

Run `flow workspace add examples https://github.com/flowexec/examples.git --set`, then `flow browse` to open a ready set of examples.

How does flow differ from a Makefile?

It sits above projects: one interface to all scripts instead of a separate Makefile and npm scripts in every repository.

Discussion

No questions or reviews yet.

CLI tools

An MCP server and Python SDK that build, run and refine AI agent workflows from a natural-language description.

MCP serverBroad access202

Flow-Like MCP

Flow-Like

Flow-Like MCP nodes expose visual workflow functions to an agent over HTTP, with owner-selected tools and access

MCP serverExtended access980

Open Dynamic Workflow turns repeatable agent tasks into validated scripts for Codex, Gemini, Cursor and other agent CLIs.

CLIExtended access84
Official

The MCP server of the Loonflow platform: an agent reads the ticket and approval queue and acts on tickets

MCP serverBroad access2.1K
Official

apra-fleet runs a fleet of AI agents across machines and models through MCP and drives multi-hour workflows

MCP serverBroad access106

README in short

The README gives a three-command quick start, a sample `api.flow` file with deploy and backup workflows and links to the documentation. A separate examples repository and an `llms.txt` documentation index for AI tools sit alongside.

README badge

Report a problem