Skip to content

README.md

Latest commit

 

History

History
78 lines (59 loc) · 4.42 KB

README.md

File metadata and controls

78 lines (59 loc) · 4.42 KB

CoreGraph Documentation

CoreGraph is a Rust CLI that builds an in-memory code symbol graph for multi-language and monorepo codebases. It combines tree-sitter (symbol extraction) with stack-graphs (cross-file name resolution), serves the graph from a background daemon, and answers questions about callers, impact, dead code, and cross-language inconsistencies — with every edge tagged by a confidence/trust model so an LLM (or a human) knows how much to trust each relationship.

These pages are the reference for the coregraph CLI, its graph model, and its integrations (MCP, LSP, HTTP). Start with the Getting started group, then dip into Concepts when you want to know why a result looks the way it does.

Getting started

Page What it covers
overview.md What CoreGraph is, the problem it solves, and core terminology
use-cases.md Task-oriented recipes: find callers, run impact, hunt dead code, check inconsistencies
cli.md Every subcommand, global options, output formats (human / llm / json), and the token budget

Concepts

Page What it covers
graph-model.md Symbol kinds, edge kinds, analysis origins, the four trust models, and the documentation layer
confidence.md How edge confidence is computed and decayed
architecture.md Daemon + thin-client design, in-memory layout, and concurrency model
data-flow.md How a file moves from disk through indexing into a query response
change-tracking.md File watching, cold start, epochs, healing, and stale-edge handling

Integrations and operations

Page What it covers
integrations.md MCP stdio bridge, LSP stdio bridge, and the HTTP API
security.md Path-traversal guards, secret handling, and network exposure
operations.md Error handling, graceful degradation, and logging

Configuration

Page What it covers
appendix/config-example.md Annotated config.toml example — every key and its default

Roadmap

Page What it covers
roadmap.md What ships today and what is still planned (not yet implemented)

Contributing and internals

These pages are for people building or modifying CoreGraph, not for using the CLI. They live under contributing/ (how to build, test, and evolve the project) and internals/ (deep implementation reference).

Page What it covers
contributing/development.md Building from source, local install, and the build / lint / test workflow
contributing/testing.md Test levels, fixtures, the golden + tier-2 e2e suites, and CI
contributing/schema-versioning.md Snapshot schema versions, serde compatibility rules, and migration
internals/tech-stack.md Rust crates, key dependencies, and the per-language parsing pipeline
internals/bloom-filter.md Bloom filter used in indexing
internals/manifest-parser.md Per-ecosystem manifest parsing (Cargo, npm, Maven, Gradle, Go, Python)

Suggested reading order

If you are new, read in this order:

  1. overview.md — get the one-paragraph picture and the vocabulary.
  2. cli.md — install, run coregraph index, and learn the core commands.
  3. use-cases.md — work through real tasks against your own repo.
  4. graph-model.md and confidence.md — understand what each edge means and how much to trust it.
  5. integrations.md — wire CoreGraph into your editor (LSP), an MCP client, or an HTTP service.
  6. architecture.md and change-tracking.md — go deeper on the daemon, the in-memory graph, and how it stays fresh.

Contributing to CoreGraph itself? Start with contributing/development.md to build, install, and run the checks, then contributing/testing.md and architecture.md.