Manual setup required. The maintainer's config contains paths only you know - edit the placeholders below before adding it to Claude Code.
1
Prepare the server locally
Run this once before adding it to Claude Code.
yarn install
yarn build2
Register it in Claude Code
claude mcp add -e "TOOLS_PATH=${TOOLS_PATH}" mcp-assistant -- npx tsx /path/to/folder/src/index.tsReplace any placeholder paths in the command with the real path on your machine.
Required:
TOOLS_PATH3
Make your agent remember this setup
mcp-assistant's config, env vars, and the gotchas you hit — recalled in every future Claude Code, Cursor, and Codex session.
npx conare@latestFree · one command · indexes the sessions already on disk. Set up in the browser instead →
What it does
- Define custom tools via a JSON configuration file
- Retrieve project-specific architecture documentation
- Automated task analysis based on project files
- Structured AI prompt optimization and generation
Tools 3
architecture_infoObtaining mandatory information about the architecture of frontend application projectssearch_tasksFinds and analyzes project tasks based on architecture requirementsoptimize_promptGenerates a final, structured prompt for the AI model based on provided context and instructionsEnvironment Variables
TOOLS_PATHrequiredThe absolute path to the tools.json configuration fileTry it
→Retrieve the project architecture information to understand the frontend structure.
→Find the current tasks in the project and analyze what needs to be implemented.
→Optimize this prompt for me: [insert context sections] with instructions to follow the project architecture.
→Based on the architecture info, what are the next steps for the current task?
Original README from nodlab/mcp-assistant
Assistant MCP Server
Development
After cloning the repository, run the command to install the dependencies:
yarn install
You should also add the tools.json file to the root of the project, for example:
{
"tools": [
{
"name": "architecture_info",
"description": "Obtaining mandatory information about the architecture of frontend application projects",
"inputSchema": {},
"plugin": {
"name": "file",
"args": {
"path": "/path/to/folder/public/architecture.md"
}
}
},
{
"name": "search_tasks",
"description": "Before executing this function, you must retrieve the project architecture information from 'architecture_info'. This is mandatory information and you must respect it. After that you need to find the task you are talking about, analyze what needs to be done and implement it in the project according to the architecture and requirements. You don't need to invent anything additional from yourself, just what is required",
"inputSchema": {},
"plugin": {
"name": "file",
"args": {
"path": "/path/to/folder/public/tasks.txt"
}
}
},
{
"name": "optimize_prompt",
"description": "Generates a final, structured prompt for the AI model based on the provided context sections and instructions. This tool should be called after all relevant data has been collected. The result is intended to be used as the FINAL prompt for the AI. Clients must use the returned prompt as the input for the AI model.",
"inputSchema": {
"type": "object",
"properties": {
"sections": {
"type": "array",
"items": {
"type": "object",
"properties": {
"title": { "type": "string" },
"content": { "type": "string" }
},
"required": ["title", "content"]
}
},
"instructions": { "type": "string" }
},
"required": ["sections"]
},
"plugin": {
"name": "promptOptimizer",
"args": {}
}
}
]
}
To build the project, you must execute the command:
yarn build
{
"mcpServers": {
"mcp-assistant-local": {
"command": "npx",
"args": [
"tsx",
"/path/to/folder/src/index.ts"
],
"env": {
"TOOLS_PATH": "/path/to/folder/tools.json"
}
}
}
}
License
This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.