Zero-LLM structural code retrieval for AI coding agents — tree-sitter index, BM25+dense+structural fusion, served over MCP.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add skeletongraph -- uvx skeletongraph --path '<PATH>'~/.codex/config.toml
[mcp_servers.skeletongraph]
command = "uvx"
args = ["skeletongraph", "--path", "<PATH>"]
# codex mcp add skeletongraph -- uvx skeletongraph --path '<PATH>'~/.cursor/mcp.json
{
"mcpServers": {
"skeletongraph": {
"command": "uvx",
"args": [
"skeletongraph",
"--path",
"<PATH>"
]
}
}
}.vscode/mcp.json
{
"servers": {
"skeletongraph": {
"type": "stdio",
"command": "uvx",
"args": [
"skeletongraph",
"--path",
"<PATH>"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"skeletongraph": {
"command": "uvx",
"args": [
"skeletongraph",
"--path",
"<PATH>"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"skeletongraph": {
"command": "uvx",
"args": [
"skeletongraph",
"--path",
"<PATH>"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"skeletongraph": {
"command": "uvx",
"args": [
"skeletongraph",
"--path",
"<PATH>"
]
}
}
}