codebase-memory-mcp: wire the knowledge graph, stop re-reading files

The Stack

Steps

1. 1

   Write the MCP server config and validate it

   Add codebase-memory-mcp to your MCP client config (Claude Code uses .claude/settings.json; Cursor uses .cursor/mcp.json). The server block needs a name, the path to the binary as the command, and no extra flags. CI checks the config parses and has the required fields; the binary install and indexing are fenced.

   cat > mcp-config.json <<'JSON'
   {
     "mcpServers": {
       "codebase-memory": {
         "command": "/usr/local/bin/codebase-memory-mcp",
         "args": []
       }
     }
   }
   JSON
   node -e '
   const fs = require("fs");
   const c = JSON.parse(fs.readFileSync("mcp-config.json", "utf8"));
   function bad(m) { console.error("BAD: " + m); process.exit(1); }
   const servers = c.mcpServers || c.mcp_servers || {};
   const keys = Object.keys(servers);
   if (keys.length === 0) bad("mcpServers must have at least one entry");
   const srv = servers[keys[0]];
   if (!srv.command || typeof srv.command !== "string") bad("server entry must have a command (path to binary)");
   if (!Array.isArray(srv.args)) bad("args must be an array");
   console.log("config OK: MCP server " + JSON.stringify(keys[0]) + " points at " + srv.command);
   '

2. 2

   Install the binary and index your repo (fenced)

   Download the binary for your platform (one-line curl from the releases page), then run index_repository with the absolute path to your repo. The first index on a large codebase takes a few minutes; subsequent queries run sub-millisecond. Watch your token counts before and after on a representative task. Binary install and indexing are fenced.