diag-struct-mcp

# Diagnostic Structurizer (diag-struct-mcp) A Model Context Protocol (MCP) server that runs heavy local CLI checkers (e.g., `cargo clippy`, `svelte-check`, `tsc`) and converts their extensive terminal logs into a clean, structured JSON format dedicated for AI agents. ## Notes for AI Agents The optimal prompts and usage for this tool are described in [SKILL.md](./SKILL.md), please follow them. ## Features - **Token Consumption Optimization**: Prevents raw terminal tracebacks from being fed into the context window, optimizing token usage. - **Unified Interface**: Handles multiple language/framework linters as a single error extraction pipeline. ## Setup ### 0. Prerequisites - **Rust Build Toolchain** ([install with rustup](https://rustup.rs/)) - **Node.js & npm** (for parsing JS/TS/Svelte) If the target project does not have the tools installed, run the following command: ```powershell npm install -g typescript svelte-check ``` ### 1. Installation You can install and build the tool with a single command from GitHub. ```powershell cargo install --git https://github.com/adasima/diag-struct-mcp --force ``` *(The above command will automatically place `diag-struct-mcp.exe` in `$HOME/.cargo/bin/`)* ### 2. Agent Configuration Add the following configuration to your agent's configuration file (e.g., `mcp_config.json`) to enable the tool: ```json "mcpServers": { "diag-struct-mcp": { "command": "diag-struct-mcp" } } ``` ### 3. Automated Installation and Configuration To install and configure all tools at once, run the following one-liner: ```powershell Invoke-RestMethod "https://raw.githubusercontent.com/adasima/ai-workshop/main/mcp_installer.ps1" | Invoke-Expression ```