atcode

Build Tree-sitter AST knowledge graphs for cross-file code exploration.

29|5|Updated Mar 26, 2026
One-click install
npx skills add https://github.com/siorigin/atcode --skill atcode
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: atcode
Source: https://github.com/siorigin/atcode/tree/main/backend/atlas_mcp
Command: npx skills add https://github.com/siorigin/atcode --skill atcode

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the pain of understanding complex codebases by turning repositories into a queryable AST knowledge graph that supports cross-file structural reasoning instead of brittle text search.

Core Features & Use Cases

  • Qualified-name code discovery: Find functions, classes, and methods by name and receive stable qualified_name identifiers for further exploration.
  • Call tracing and impact analysis: Trace incoming/outgoing calls to answer “who calls this?” and “what does this depend on?”
  • One-call deep dives: Use graph-backed exploration to retrieve source, callers, and dependencies for a specific code element.
  • Project graph operations: Build, refresh, clean, and sync graphs for medium-to-large repositories, including real-time incremental updates.

Quick Start

Ask your AI to run: list_repos(), then set_project() with the exact project name you want to explore, and then call find_nodes("main") to get qualified_name targets for deeper tracing.

Frequently Asked Questions about atcode

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I trace function call chains across multiple files in a large codebase?

Call tracing across files is done by building a Tree-sitter AST knowledge graph. This graph exposes incoming and outgoing dependency queries to map structural relationships that text search cannot track.

What is an AST knowledge graph and how does it help with code exploration?

An AST knowledge graph is a structural representation of repository code parsed by Tree-sitter and stored in Memgraph. It enables cross-file reasoning and dependency discovery instead of brittle text search.

How do I find and retrieve source code for specific functions using an AST graph?

You find code by using qualified-name discovery to locate functions and classes, which returns stable identifiers. These identifiers enable graph-backed deep dives to retrieve source, callers, and dependencies.

Can I perform dependency analysis on medium-to-large repositories in real-time?

Yes, dependency analysis scales to medium-to-large repositories through project graph operations. The system supports building, refreshing, and syncing graphs with real-time incremental updates.

Does code exploration via an AST knowledge graph require an MCP or HTTP server?

Yes, the graph queries and lifecycle operations require an MCP or HTTP server backed by Memgraph. This server architecture manages the qualified_name-driven tool chaining needed for repository exploration.

What is the best way to start exploring a codebase using a knowledge graph?

Start by calling list_repos to view available projects, then set_project with the exact name. Finally, call find_nodes with a search term to get qualified_name targets for deeper tracing.