# yanmxa/scriptflow-mcp [Health: Active]

**Category:** 👨‍💻 Code Execution  
**Repository:** https://github.com/yanmxa/scriptflow-mcp  
**GitHub Stars:** 8  
**npm Downloads (last month):** 35  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/yanmxa-scriptflow-mcp

## Description
A script workflow management system that transforms complex, repetitive AI interactions into persistent, executable scripts. Supports multi-language execution (Bash, Python, Node.js, TypeScript) with guaranteed consistency and reduced hallucination risks.

## Tools
Capabilities this server exposes over MCP:

- **script_add** — Add a new script to the repository
- **script_list** — List and discover available scripts
- **script_run** — Execute a script by name
- **script_edit** — Edit an existing script
- **script_rm** — Remove a script from the repository
- **script_get** — Get detailed information about a specific script

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "scriptflow-mcp": {
    "command": "npx",
    "args": ["scriptflow-mcp"],
    "env": {
      "SCRIPTFLOW_SCRIPTS_DIR": "",
      "SCRIPTFLOW_TIMEOUT": ""
    }
  }
}
```

**Requires environment variables:** `SCRIPTFLOW_SCRIPTS_DIR`, `SCRIPTFLOW_TIMEOUT` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What yanmxa/scriptflow-mcp MCP server does

yanmxa/scriptflow-mcp MCP server provides an MCP interface for managing a local collection of executable scripts. It is designed for workflows that recur across sessions and can be captured as named scripts rather than reconstructed through repeated interactions.

Each script has content, a language, a name, a description, and optional tags. The available operations cover the script lifecycle: create a script, inspect its details, search or list scripts, modify an existing script, execute one by name, and remove it. Scripts can be written in Bash, Python, JavaScript, or TypeScript.

The repository stores script content and metadata as files. A script uses a filename based on its name and language extension, while its metadata is kept in a matching JSON file. This file-based layout can also be placed under version control, although the provided tools do not expose separate version-control operations.

## How it works

An MCP client starts the server, commonly through `npx scriptflow-mcp`. The client can then call the exposed tools to maintain the script repository. `script_add` accepts the script name, description, content, language, and tags. `script_list` supports discovery and filtering, including searches by name, description, or tags.

To inspect one entry, call `script_get` with the script name. `script_edit` updates an existing script, and `script_rm` deletes one. `script_run` executes a named script and can receive arguments. Execution is subject to the configured timeout.

This model fits repeatable tasks where keeping an exact script is preferable to asking an AI model to reproduce the same multi-step interaction each time. It does not turn an interaction into an autonomous workflow by itself; the workflow must be represented by script content that the supported runtime can execute.

## Setup and configuration

Configure yanmxa/scriptflow-mcp MCP server as an MCP process in the client configuration. A typical entry uses the `npx` command with `scriptflow-mcp` as its argument.

`SCRIPTFLOW_SCRIPTS_DIR` selects the repository location. If it is not set, the default is `/tmp/scriptflow-mcp/scripts`. `SCRIPTFLOW_TIMEOUT` sets the execution timeout in milliseconds, with a default of `30000`.

After installation, the repository can be tested with the MCP Inspector using `npx @modelcontextprotocol/inspector npx scriptflow-mcp`. The project also documents an `npm test` command for its test suite.

## Tools and capabilities

- Add scripts with metadata using `script_add`.
- Find scripts using `script_list`, including name, description, or tag searches.
- Read stored details with `script_get`.
- Change existing scripts with `script_edit`.
- Run scripts by name and pass arguments with `script_run`.
- Delete scripts with `script_rm`.
- Store and execute Bash, Python, JavaScript, and TypeScript files.

## Limitations and notes

yanmxa/scriptflow-mcp MCP server executes local script files, so the selected scripts directory and runtime environment determine what can run successfully. The material specifies a configurable timeout but does not describe sandboxing, permission controls, remote storage, or dependency management.

The server manages scripts and their metadata through MCP tools; it does not document tools for directly editing conversation history or generating scripts automatically from prior AI interactions. Any reusable workflow must be added as script content before it can be run.

_Full upstream README: https://allmcps.com/mcp/yanmxa-scriptflow-mcp/readme_