Metadata-Version: 2.4
Name: mcp-devtools
Version: 1.2.4
Summary: A short description of your project.
Author-email: daoch4n <daoch4n@gmail.com>
License: MIT
Project-URL: Homepage, https://github.com/daoch4n/mcp-devtools
Project-URL: Bug Tracker, https://github.com/daoch4n/mcp-devtools/issues
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Requires-Python: >=3.12
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: fastapi
Requires-Dist: uvicorn
Requires-Dist: mcp-server-git
Requires-Dist: httpx
Requires-Dist: python-dotenv
Requires-Dist: PyYAML
Dynamic: license-file

# ⚙️ mcp-devtools: multi-functional development tools MCP server over SSE

[![GitHub repository](https://img.shields.io/badge/GitHub-repo-blue?logo=github)](https://github.com/daoch4n/zen-ai-mcp-devtools)
[![GitHub Workflow Status](https://img.shields.io/github/actions/workflow/status/daoch4n/zen-ai-mcp-devtools/python-package.yml?branch=main)](https://github.com/daoch4n/zen-ai-mcp-devtools/actions/workflows/python-package.yml)
[![PyPI](https://img.shields.io/pypi/v/mcp-devtools)](https://pypi.org/project/mcp-devtools)

- 🔧 `mcp-devtools` offers a comprehensive suite of development tools:
- 📄 For detailed overview of all tools, their arguments, and descriptions, please refer to [Available Tools](#%E2%84%B9%EF%B8%8F-available-tools).
  -  🎋 Git management operations (`git_status`, `git_stage_and_commit`, `git_diff`, `git_diff_all`, `git_log`, `git_create_branch`, `git_reset` `git_checkout`, `git_show`)
  -  📁 Git file operations (`git_read_file`, `git_apply_diff`)
  -  📂 Direct file operations (`search_and_replace`, `write_to_file`)
  -  🖥️ Terminal commands execution (`execute_command`), please review [Automation-Related Security](#-automation-related-security-considerations).
  -  🤖 AI-assisted file operations using [Aider](https://github.com/Aider-AI/aider) (`ai_edit` (`aider_status` for systems ready check ))
    <br> ℹ️ When using the `ai_edit` tool, please refer to the [Aider Configuration documentation](docs/aider_config.md).

## ⛎ Use Cases

- 🌐 Use it to extend online chat-based assistants such as ChatGPT, Google Gemini or AI Studio, Perplexity, Grok, OpenRouter Chat, DeepSeek, Kagi, T3 Chat with direct access to local files, git, terminal commands execution and AI-assisted file editing capabilities via [MCP-SuperAssistant](https://github.com/srbhptl39/MCP-SuperAssistant/) or similar projects.
- 👩🏻‍💻 Use it to boost code editors like Cursor, Windsurf or VSCode extensions like Roo Code, Cline, Copilot or Augment with intuitive Git management and AI-assisted file editing capabilities and say goodbye to those pesky diff application failures wasting your time or `Roo having trouble...` breaking your carefully engineered automation workflows. Aider seems to get diffs right! (if it still doesn't work quite well, try to find perfect way for your AI model by explicitly passing different `edit_format` [parameters](#ai_edit) to `ai_edit` tool):
  - `unidiff` seems to work better with GPT
  - `diff-fenced`  performs best with Gemini
  - `diff` provides balanced quick results on all models (default)
  - `whole` is the slowest but most reliable option as it simply overwrites file

## 🦘 [Agentic-Driven Workflow](https://github.com/daoch4n/research/blob/main/agentic-driven-workflows.md) with Roo

https://github.com/user-attachments/assets/4d218e5e-906c-4d24-abc3-09ab0acdc1d0

  - For [Roo Code](https://github.com/RooCodeInc/Roo-Code), place the [.roomodes](https://github.com/daoch4n/zen-ai-mcp-devtools/blob/main/.roomodes) file into your repo root to experience automated two-levels deep [nested agents execution](https://www.perplexity.ai/search/nested-agent-execution-BsD4hcqjTdKlEUKJLnv9.g) flow:
    ### 😺 Basic Flow (Roo 🪃 adapted to summon Aider as level 2 subagent)
    - `🛸 AI Orchestrator` agent is your Basic Flow manager, so talk to it if you prefer to keep your agent management simple. It coordinates agents by breaking complex problems into logical subtasks and delegates each using the Roo Code native `new_task` tool, providing comprehensive, context-rich instructions for each specialized AI agent. Each subtask receives explicit instructions to only perform the outlined work and signal completion using native `attempt_completion` tool. The orchestrator tracks all subtasks, synthesizes results, and suggests workflow improvements to reach near complete automation of your dev workflow ([this pdf](https://github.com/daoch4n/research/blob/main/2025_01_18_pdf_techai_google_whitepaper_prompt_engineering.pdf) might help). Works fine using fast models with reasoning enabled to speed up your workflow or deep reasoning models to add basic subtasks management.
    - `🤖 AI Code` and `👾 AI Debug` agents (fast models like Gemini Flash 2.5 / GPT 4.1 work well here with lower temperature) delegate all code changes to Aider via `ai_edit` tool and enforce strict MCP tool schema compliance and review of Aider agent work with automated agent redelegation on unsatisfactory or missing results by checking diff (or stdout) output of `ai_edit` tool call.
    - `👽 AI Architect` and `🔬 AI Researcher` agents are responsible for architectural directions and geneal R&D needs. Deep reasoning models (like DeepSeek R1 / Gemini Pro 2.5 / OpenAI o1/o3/o4mini) are recommended for `👽 AI Architect`, and internet-connected models (like Perplexity Sonar) are recommended for `🔬 AI Researcher`. Adjust temperatures based on creativity level needed.
    ### 😼 Advanced Flow ([RooRoo](https://github.com/marv1nnnnn/rooroo) 🧭 adapted to summon Aider as level 2 subagent)
    - `🧭 Rooroo Navigator` agent is your Advanced Flow manager, so talk to it if you prefer to keep your agent management in orderly manner. Responsible for overall project coordination and task orchestration, task lifecycles, delegation to Planner, Developer, Analyzer, Idea Sparker, processes your commands, and oversees your workflows. Provides `context.md` files to tasks, either the ones generated by `🗓️ Rooroo Planner`, or generates new one if Planner wasn't deemed neccessary for the task. If used without deep reasoning model, Navi tends to forget generating them on its own, but delegated agents will remind Navi, and task will be redelegated correctly with `context.md` provided. So decide on LLM model yourself for this one: if you prefer speed but can tolerate some inefficencies due to Navi forgetting to write context files, use fast models with reasoning enabled. So if you don't mind Navi thinking for a minute on every step, use deep reasoning model here as well. Still can't decide? This [DeepSeek R1/V3 hybrid](https://chutes.ai/app/chute/aef797d4-f375-5beb-9986-3ad245947469?tab=api) is a good candidate.
    - `👩🏻‍💻 Rooroo Developer` agent gets detailed instructions from `🧭 Rooroo Navigator`  via `context.md` passed to it and outputs task result in strict JSON schema asking Navi for clarification or complaining if context file was not provided. (context file in question needs to be generated either by Navi itself or via invocation of specialized `🗓️ Rooroo Planner` agent). Same model recommendations apply from Basic flow.
    - `📊 Rooroo Analyzer` agent combines the functions of `👽 AI Architect` and `🔬 AI Researcher` from Basic flow, it also gets task context via `context.md` passed and complains back to Navi if it wasn't found. Deep reasoning models remommended. Internet-connected models might provide more relevant analysis results. Adjust temperature based on creativity needed.
    - `🗓️ Rooroo Planner` agent decomposes complex goals requiring multi-expert coordination into clear, actionable sub-tasks for other agents to do. It is also the main supplier of `context.md` files for them. Deep reasoning model also recommended.
    - `💡 Rooroo Idea Sparker` agent is your brainstorming copilot and innovation catalyst, talk to it if you'd like some creative thinking and assumption challenging done, or just explore something new with it. Deep reasoning model with higher temperature set or internet-connected model recommended here.

## 1️⃣ Prerequisites

- Python 3.12, [uv](https://github.com/astral-sh/uv)

### 🐧 Linux/macOS

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```

### 🪟 Windows

```powershell
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
```

## 2️⃣ Usage

### 🐍 Running from PyPi

```bash
uvx mcp-devtools@1.2.0 -p 1337
```

### 🐈‍⬛ Running from GitHub

#### 🐧 Linux/macOS

```bash
git clone "https://github.com/daoch4n/zen-ai-mcp-devtools/"
cd zen-ai-mcp-devtools
./server.sh -p 1337
```

#### 🪟 Windows

```powershell
git clone "https://github.com/daoch4n/zen-ai-mcp-devtools/"
cd zen-ai-mcp-devtools
.\server.ps1 -p 1337
```

## 3️⃣ MCP Server Configuration

To integrate `mcp-devtools` with your AI assistant, add the following configuration to your MCP settings file:

```json
{
  "mcpServers": {
    "devtools": {
      "url": "http://127.0.0.1:1337/sse",
      "disabled": false,
      "alwaysAllow": [],
      "timeout": 900
    }
  }
}
```

## 4️⃣ [AI System Prompt](https://github.com/daoch4n/research/blob/main/prompt-engineering-in-agentic-driven-software-development-research.md) Example

```
You are an AI assistant with access to development tools.
Use the tools provided by the 'devtools' MCP server for git management, file operations, and terminal access.
When using any tool from 'devtools', always provide the absolute path to the current repository's working directory as the 'repo_path' option.
Do not set 'repo_path' to any other folder.
The 'repo_path' must be explicitly requested from the user at the beginning of the conversation.
When using the 'execute_command' tool, the current working directory will be set to the 'repo_path' provided.
When performing file manipulations with 'execute_command', ensure to pass the full path in the terminal command, including the 'repo_path' prefix, for the manipulated file path.
When using the 'ai_edit' tool, never put actual square brackets and other JSON breaking symbols in the tool calls; instead, use verbal descriptions of them.
```

## ⁉️ Known Issues and Workarounds

### 💾 Direct Code Editing vs 🤖 Delegated Editing

**Issue:**

*    🔍 When using the `write_to_file` tool for direct code editing, especially with languages like JavaScript that utilize template literals (strings enclosed by backticks), you may encounter unexpected syntax errors. This issue stems from how the AI assistant generates the `content` string, where backticks and dollar signs within template literals might be incorrectly escaped with extra backslashes (`\`).

**Mitigation:** 

*    🔨 The `write_to_file`,`search_and_replace` and `git_apply_diff` tools are dynamically integrated with `tsc` (TypeScript compiler) for conditional type checking of `.js`, `.mjs`, and `.ts` files on edit. The output of `tsc --noEmit --allowJs` is provided as part of the tool's response. AI assistants should parse this output to detect any compiler errors and *should not proceed with further actions* if errors are reported, indicating a problem with the written code.

**Workarounds:**

*    🤖 (most reliable) Instruct your AI assistant to delegate editing files to MCP-compatible coding agent by using `ai_edit` tool instead, as it is more suitable for direct code manipulation, automatically commits changes and produces resulting diff as tool output, and let AI assistant act as task orchestrator that will write down plans and docs with `write_to_file` tool then delegate actual coding to specialized agent, get its report (diff) as tool call result, use `git_read_file` tool to double check agent's work, and manage commits and branches (`ai_edit` tool basically integrates `Aider` via some logic ported from [its MCP bridge](https://github.com/sengokudaikon/aider-mcp-server)).
*    🖥️ (if you're feeling lucky) Instruct your AI assistant to craft a terminal command to edit problematic file via `execute_command` tool.

### ❔ Aider limitations due to its commit-first nature

**Issue:**

*    🔍 When using `ai_edit` tool in a dirty repo state, e.g. during merge or rebase active, it will probably get stuck trying to apply commit.
  
**Workarounds:**

*    ⚙️ Temporarily disable auto commiting functions in your `.aider.conf.yml` configuration file.

## 🙈 Automation-Related Security Considerations

- 🛡️ For automated workflows, always run MCP Servers in isolated environments (🐧[Firejail](https://github.com/netblue30/firejail) or 🪟[Sandboxie](https://github.com/sandboxie-plus/Sandboxie))
- 🗃️ Filesystem access boundaries are maintained via passing `repo_path` to every tool call, so AI assistant only has read/write access to files in the current workspace (relative to any path AI decides to pass as `repo_path` , make sure system prompt is solid on cwd use).
- ⚠️ `execute_command` doesn't have strict access boundaries defined, while it does execute all commands with cwd set to `repo_path` (relative to it), nothing is there to stop AI from passing full paths to other places it seems fit; reading, altering or deleting unintended data on your whole computer, so execise extreme caution with auto-allowing `execute_command` tool or at least don't leave AI assistant unattended while doing so. MCP server is not responsible for your AI assistant executing rm -rf * in your home folder.

## ℹ️ Available Tools


### `git_status`
- **Description:** Shows the current status of the Git working tree, including untracked, modified, and staged files.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      }
    },
    "required": [
      "repo_path"
    ]
  }
  ```


### `git_diff_all`
- **Description:** Shows all changes in the working directory, including both staged and unstaged modifications, compared to the HEAD commit. This provides a comprehensive view of all local changes.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      }
    },
    "required": [
      "repo_path"
    ]
  }
  ```


### `git_diff`
- **Description:** Shows differences between the current working directory and a specified Git target (e.g., another branch, a specific commit hash, or a tag).
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "target": {
        "type": "string",
        "description": "The target (e.g., branch name, commit hash, tag) to diff against. For example, 'main', 'HEAD~1', or a full commit SHA."
      }
    },
    "required": [
      "repo_path",
      "target"
    ]
  }
  ```

### `git_stage_and_commit`
- **Description:** Stages specified files (or all changes if no files are specified) and then commits them to the repository with a given message. This creates a new commit in the Git history.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "message": {
        "type": "string",
        "description": "The commit message for the changes."
      },
      "files": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "An optional list of specific file paths (relative to the repository root) to stage before committing. If not provided, all changes will be staged."
      }
    },
    "required": [
      "repo_path",
      "message"
    ]
  }
  ```


### `git_reset`
- **Description:** Unstages all currently staged changes in the repository, moving them back to the working directory without discarding modifications. This is equivalent to `git reset` without arguments.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      }
    },
    "required": [
      "repo_path"
    ]
  }
  ```

### `git_log`
- **Description:** Shows the commit history for the repository, listing recent commits with their hash, author, date, and message. The number of commits can be limited.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "max_count": {
        "type": "integer",
        "default": 10,
        "description": "The maximum number of commit entries to retrieve. Defaults to 10."
      }
    },
    "required": [
      "repo_path"
    ]
  }
  ```

### `git_create_branch`
- **Description:** Creates a new Git branch with the specified name. Optionally, you can base the new branch on an existing branch or commit, otherwise it defaults to the current active branch.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "branch_name": {
        "type": "string",
        "description": "The name of the new branch to create."
      },
      "base_branch": {
        "type": "string",
        "nullable": true,
        "description": "Optional. The name of the branch or commit hash to base the new branch on. If not provided, the new branch will be based on the current active branch."
      }
    },
    "required": [
      "repo_path",
      "branch_name"
    ]
  }
  ```

### `git_checkout`
- **Description:** Switches the current active branch to the specified branch name. This updates the working directory to reflect the state of the target branch.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "branch_name": {
        "type": "string",
        "description": "The name of the branch to checkout."
      }
    },
    "required": [
      "repo_path",
      "branch_name"
    ]
  }
  ```

### `git_show`
- **Description:** Shows the metadata (author, date, message) and the diff of a specific commit. This allows inspection of changes introduced by a particular commit.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "revision": {
        "type": "string",
        "description": "The commit hash or reference (e.g., 'HEAD', 'main', 'abc1234') to show details for."
      }
    },
    "required": [
      "repo_path",
      "revision"
    ]
  }
  ```

### `git_apply_diff`
- **Description:** Applies a given diff content (in unified diff format) to the working directory of the repository. This can be used to programmatically apply patches or changes.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "diff_content": {
        "type": "string",
        "description": "The diff content string to apply to the repository. This should be in a unified diff format."
      }
    },
    "required": [
      "repo_path",
      "diff_content"
    ]
  }
  ```

### `git_read_file`
- **Description:** Reads and returns the entire content of a specified file within the Git repository's working directory. The file path must be relative to the repository root.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "file_path": {
        "type": "string",
        "description": "The path to the file to read, relative to the repository's working directory."
      }
    },
    "required": [
      "repo_path",
      "file_path"
    ]
  }
  ```


### `search_and_replace`
- **Description:** Searches for a specified string or regex pattern within a file and replaces all occurrences with a new string. Supports case-insensitive search and line-range restrictions. It attempts to use `sed` for efficiency, falling back to Python logic if `sed` fails or makes no changes.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "file_path": {
        "type": "string",
        "description": "The path to the file to modify, relative to the repository's working directory."
      },
      "search_string": {
        "type": "string",
        "description": "The string or regex pattern to search for within the file."
      },
      "replace_string": {
        "type": "string",
        "description": "The string to replace all matches of the search string with."
      },
      "ignore_case": {
        "type": "boolean",
        "default": false,
        "description": "If true, the search will be case-insensitive. Defaults to false."
      },
      "start_line": {
        "type": "integer",
        "nullable": true,
        "description": "Optional. The 1-based starting line number for the search and replace operation (inclusive). If not provided, search starts from the beginning of the file."
      },
      "end_line": {
        "type": "integer",
        "nullable": true,
        "description": "Optional. The 1-based ending line number for the search and replace operation (inclusive). If not provided, search continues to the end of the file."
      }
    },
    "required": [
      "repo_path",
      "file_path",
      "search_string",
      "replace_string"
    ]
  }
  ```

### `write_to_file`
- **Description:** Writes the provided content to a specified file within the repository. If the file does not exist, it will be created. If it exists, its content will be completely overwritten. Includes a check to ensure content was written correctly and generates a diff.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory."
      },
      "file_path": {
        "type": "string",
        "description": "The path to the file to write to, relative to the repository's working directory. The file will be created if it doesn't exist, or overwritten if it does."
      },
      "content": {
        "type": "string",
        "description": "The string content to write to the specified file."
      }
    },
    "required": [
      "repo_path",
      "file_path",
      "content"
    ]
  }
  ```

### `execute_command`
- **Description:** Executes an arbitrary shell command within the context of the specified repository's working directory. This tool can be used for tasks not covered by other specific Git tools, such as running build scripts, linters, or other system commands.
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the directory where the command should be executed."
      },
      "command": {
        "type": "string",
        "description": "The shell command string to execute (e.g., 'ls -l', 'npm install')."
      }
    },
    "required": [
      "repo_path",
      "command"
    ]
  }
  ```

### `ai_edit`
- **Description:** AI pair programming tool for making targeted code changes using Aider. Use this tool to:
  1. Implement new features or functionality in existing code
  2. Add tests to an existing codebase
  3. Fix bugs in code
  4. Refactor or improve existing code
  5. Make structural changes across multiple files

  The tool requires:
  - A repository path where the code exists
  - A detailed message describing what changes to make. Please only describe one change per message. If you need to make multiple changes, please submit multiple requests.

  **Edit Format Selection:**
  If the `edit_format` option is not explicitly provided, the default is selected based on the model name:
  - If the model includes `gemini`, defaults to `diff-fenced`
  - If the model includes `gpt`, defaults to `udiff`
  - Otherwise, defaults to `diff`

  Best practices for messages:
  - Be specific about what files or components to modify
  - Describe the desired behavior or functionality clearly
  - Provide context about the existing codebase structure
  - Include any constraints or requirements to follow

  Examples of good messages:
  - "Add unit tests for the Customer class in src/models/customer.rb testing the validation logic"
  - "Implement pagination for the user listing API in the controllers/users_controller.js file"
  - "Fix the bug in utils/date_formatter.py where dates before 1970 aren't handled correctly"
  - "Refactor the authentication middleware in middleware/auth.js to use async/await instead of callbacks"
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository's working directory where the AI edit should be performed."
      },
      "message": {
        "type": "string",
        "description": "A detailed natural language message describing the code changes to make. Be specific about files, desired behavior, and any constraints."
      },
      "files": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "A list of file paths (relative to the repository root) that Aider should operate on. This argument is mandatory."
      },
      "options": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "Optional. A list of additional command-line options to pass directly to Aider (e.g., ['--model=gpt-4o', '--dirty-diff']). Each option should be a string."
      },
      "edit_format": {
        "type": "string",
        "enum": [
          "diff",
          "diff-fenced",
          "udiff",
          "whole"
        ],
        "default": "diff",
        "description": "Optional. The format Aider should use for edits. Defaults to 'diff'. Options: 'diff', 'diff-fenced', 'udiff', 'whole'."
      }
    },
    "required": [
      "repo_path",
      "message",
      "files"
    ]
  }
  ```

### `aider_status`
- **Description:** Check the status of Aider and its environment. Use this to:
  1. Verify Aider is correctly installed
  2. Check API keys
  3. View the current configuration
  4. Diagnose connection or setup issues
- **Input Schema:**
  ```json
  {
    "type": "object",
    "properties": {
      "repo_path": {
        "type": "string",
        "description": "The absolute path to the Git repository or working directory to check Aider's status within."
      },
      "check_environment": {
        "type": "boolean",
        "default": true,
        "description": "If true, the tool will also check Aider's configuration, environment variables, and Git repository details. Defaults to true."
      }
    },
    "required": [
      "repo_path"
    ]
  }
