get_suggestionsLABSCO SUMMARY
Suggestions that carry a source.
The returned suggestions name the file they came from, which is the practical difference between a grounded answer and a fluent one — you can check it. It is aimed squarely at Cline, and the project_path.txt mechanism means switching projects is a manual edit, not automatic.
What it is
An MCP server for the Cline coding agent. It vectorises your project's documentation, detects which languages, frameworks and libraries the codebase actually uses, and pulls in official documentation links for them — so suggestions come from a knowledge base rather than from recall.
What you get
get_suggestions— code suggestions for the snippet you pass ascodeContext, withcodeandlanguage, answered from the indexed knowledge base and returning the source each suggestion came from- Documentation integration — loads and vectorises files from the
docsdirectory or from URLs you supply, and accepts multiple documents and links - Technology detection — identifies the languages, frameworks and libraries in the project
- Automatic documentation retrieval — finds and adds official documentation links for the technologies it detected
- Project path automation — reads
project_path.txtso it works against the project currently open in Cline
Requirements
Node.js v14 or higher, npm v6 or higher, and an OpenAI API key. Clone the repo, npm install, put OPENAI_API_KEY in a .env, then npm run build. Register node build/index.js in cline_mcp_settings.json with the key in the env block, write the absolute path of your current project into project_path.txt, and restart Cline. Package version 0.1.0.
Setup effort
Build from source
USE CASES
Getting suggestions that match your own docs
Index the project's documentation, then ask for code suggestions and read where each one came from.
Working with a library the model half-remembers
Let it detect the technologies in use and pull the official documentation into the knowledge base first.
1 tool
Listed in project docs. Your client may show a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool named in the project's documentation.
Requires
- Node.js v14 or later
- OPENAI_API_KEY in your environment
INSTALLATION
Connect your client
MCP Clients
Cline
Clone and build it
Build the server from the repository (Node.js v14 or later).
git clone https://github.com/AviOfLagos/MCP-coding-assistant.git
cd MCP-coding-assistant
npm install
npm run buildAdd to cline_mcp_settings.json
<YOUR_PATH_TO_CODING_ASSISTANT_SERVER> is the absolute path of the MCP-coding-assistant folder you just cloned. Then create project_path.txt in that folder holding the absolute path of the project you are working on, and restart Cline.
{
"mcpServers": {
"coding-assistant": {
"command": "node",
"args": [
"<YOUR_PATH_TO_CODING_ASSISTANT_SERVER>/build/index.js"
],
"env": {
"OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
}
}
}
}