
understand-anything-knowledge-graph
Turn any codebase into an interactive knowledge graph using Claude Code skills — explore, search, and ask questions about any project visually.
- Installs
- 2.4K
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/reason-machines/trending-skills --skill understand-anything-knowledge-graph- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads understand-anything-knowledge-graph automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Understand Anything — Codebase Knowledge Graph
Skill by ara.so — Daily 2026 Skills collection.
Understand Anything is a Claude Code plugin that runs a multi-agent pipeline over your project, builds a knowledge graph of every file, function, class, and dependency, and opens an interactive React dashboard for visual exploration. It produces plain-English summaries of every node so anyone — developer, PM, or designer — can understand the codebase.
Installation
Via Claude Code plugin marketplace
/plugin marketplace add Lum1104/Understand-Anything
/plugin install understand-anything
From source (development)
git clone https://github.com/Lum1104/Understand-Anything
cd Understand-Anything
pnpm install
pnpm --filter @understand-anything/core build
pnpm --filter @understand-anything/skill build
pnpm --filter @understand-anything/dashboard build
Core Skills / Commands
| Command | What it does |
|---|---|
/understand | Run the full multi-agent analysis pipeline on the current project |
/understand-dashboard | Open the interactive knowledge graph dashboard |
/understand-chat <question> | Ask anything about the codebase in natural language |
/understand-diff | Analyze impact of current uncommitted changes |
/understand-explain <path> | Deep-dive explanation of a specific file or function |
/understand-onboard | Generate an onboarding guide for new team members |
Typical Workflow
1. Analyze a project
# Inside any project directory, in Claude Code:
/understand
This orchestrates 5 agents in sequence (with file-analyzers running up to 3 concurrent):
- project-scanner — discovers files, detects languages/frameworks
- file-analyzer — extracts functions, classes, imports; builds graph nodes and edges
- architecture-analyzer — groups nodes into architectural layers (API, Service, Data, UI, Utility)
- tour-builder — generates ordered learning tours
- graph-reviewer — validates referential integrity
Output is saved to .understand-anything/knowledge-graph.json in your project root.
2. Open the dashboard
/understand-dashboard
The React + Vite dashboard opens in your browser. Features:
- Graph view — React Flow canvas, color-coded by layer, zoom/pan
- Node inspector — click any node for code, relationships, LLM summary
- Search — fuzzy + semantic search across all nodes
- Tours — guided walkthroughs ordered by dependency
- Persona mode — toggle detail level (Junior Dev / PM / Power User)
3. Ask questions
/understand-chat How does authentication work in this project?
/understand-chat What calls the payment service?
/understand-chat Which files are most depended on?
4. Review diff impact before committing
# After making changes:
/understand-diff
Returns a list of affected nodes in the knowledge graph — shows ripple effects before you push.
5. Explain a specific file
/understand-explain src/auth/login.ts
/understand-explain src/services/PaymentService.ts
Knowledge Graph Schema
The graph is stored at .understand-anything/knowledge-graph.json. Key types (from packages/core):
// packages/core/src/types.ts
interface GraphNode {
id: string; // unique: "file:src/auth/login.ts"
type: "file" | "function" | "class" | "module";
name: string;
filePath: string;
layer: ArchitectureLayer; // "api" | "service" | "data" | "ui" | "utility"
summary: string; // LLM-generated plain-English description
code?: string; // raw source snippet
language?: string;
concepts?: LanguageConcept[]; // e.g. "generics", "closures", "decorators"
metadata?: Record<string, unknown>;
}
interface GraphEdge {
id: string;
source: string; // node id
target: string; // node id
type: "imports" | "calls" | "extends" | "implements" | "uses";
label?: string;
}
interface KnowledgeGraph {
version: string;
generatedAt: string;
projectRoot: string;
nodes: GraphNode[];
edges: GraphEdge[];
tours: GuidedTour[];
}
type ArchitectureLayer = "api" | "service" | "data" | "ui" | "utility" | "unknown";
type LanguageConcept =
| "generics"
| "closures"
| "decorators"
| "async-await"
| "interfaces"
| "higher-order-functions"
| "dependency-injection"
| "observers"
| "iterators"
| "pattern-matching"
| "monads"
| "currying";
Working with the Core Package Programmatically
import { loadKnowledgeGraph, searchGraph, buildTour } from "@understand-anything/core";
// Load the persisted graph
const graph = await loadKnowledgeGraph(".understand-anything/knowledge-graph.json");
// Fuzzy search across all nodes
const results = searchGraph(graph, "payment processing");
console.log(results.map(r => `${r.type}:${r.name} (${r.filePath})`));
// Find all callers of a function
const paymentNode = graph.nodes.find(n => n.name === "processPayment");
const callers = graph.edges
.filter(e => e.target === paymentNode?.id && e.type === "calls")
.map(e => graph.nodes.find(n => n.id === e.source));
// Get all nodes in the service layer
const serviceNodes = graph.nodes.filter(n => n.layer === "service");
// Build a guided tour starting from a specific node
const tour = buildTour(graph, { startNodeId: "file:src/index.ts" });
tour.steps.forEach((step, i) => {
console.log(`Step ${i + 1}: ${step.node.name} — ${step.node.summary}`);
});
Dashboard Development
# Start the dashboard dev server (hot reload)
pnpm dev:dashboard
# Build for production
pnpm --filter @understand-anything/dashboard build
The dashboard is a Vite + React 18 app using:
- React Flow — graph canvas rendering
- Zustand — graph state management
- TailwindCSS v4 — styling
- Fuse.js — fuzzy search
- web-tree-sitter — in-browser AST parsing
- Dagre — automatic graph layout
Project Structure
understand-anything-plugin/
├── .claude-plugin/ # Plugin manifest (read by Claude Code)
├── agents/ # Agent definitions (project-scanner, file-analyzer, etc.)
├── skills/ # Skill definitions (/understand, /understand-chat, etc.)
├── src/ # Plugin TypeScript source
│ ├── context-builder.ts # Builds LLM context from the graph
│ └── diff-analyzer.ts # Git diff → affected nodes
├── packages/
│ ├── core/ # Analysis engine
│ │ ├── src/
│ │ │ ├── types.ts # GraphNode, GraphEdge, KnowledgeGraph
│ │ │ ├── persistence.ts
│ │ │ ├── search.ts # Fuzzy + semantic search
│ │ │ ├── tours.ts # Tour generation
│ │ │ ├── schema.ts # Zod validation schemas
│ │ │ └── tree-sitter.ts
│ │ └── tests/
│ └── dashboard/ # React dashboard app
│ └── src/
Incremental Updates
Re-running /understand only re-analyzes files that changed since the last run (based on mtime and content hash stored in the graph metadata). For large monorepos this makes subsequent runs fast.
To force a full re-analysis:
rm -rf .understand-anything/
/understand
Development Commands
pnpm install # Install all dependencies
pnpm --filter @understand-anything/core build # Build core package
pnpm --filter @understand-anything/core test # Run core tests
pnpm --filter @understand-anything/skill build # Build plugin package
pnpm --filter @understand-anything/skill test # Run plugin tests
pnpm --filter @understand-anything/dashboard build # Build dashboard
pnpm dev:dashboard # Dashboard dev server with HMR
Common Patterns
Before a code review
# See what your diff actually touches in the architecture
/understand-diff
Onboarding a new engineer
# Generate a structured onboarding doc grounded in the real code
/understand-onboard
Researching a feature area
/understand-chat What are all the entry points for the GraphQL API?
/understand-explain src/graphql/resolvers/
Understanding an unfamiliar module
/understand-explain src/workers/queue-processor.ts
# Returns: summary, key functions, what calls it, what it calls, concepts used
Troubleshooting
/understand times out on large repos
- The file-analyzer runs up to 3 workers concurrently. Very large repos (>50k files) may need patience. Delete
.understand-anything/and re-run if a previous run was interrupted.
Dashboard doesn't open
- Run
pnpm --filter @understand-anything/dashboard buildfirst if working from source, then retry/understand-dashboard.
Stale graph after major refactor
- Delete
.understand-anything/knowledge-graph.jsonto force full re-analysis:rm .understand-anything/knowledge-graph.json && /understand
pnpm install fails with workspace errors
- Ensure you are using pnpm v8+:
pnpm --version. The project uses pnpm workspaces defined inpnpm-workspace.yaml.
Search returns no results
- Confirm the graph was generated:
cat .understand-anything/knowledge-graph.json | head -5. If empty or missing, run/understandfirst.
Contributing
# Fork, then:
git checkout -b feature/my-feature
pnpm --filter @understand-anything/core test # must pass
# open a PR — file an issue first for major changes
License: MIT © Lum1104
FAQ
What does the understand-anything-knowledge-graph skill do?
Turn any codebase into an interactive knowledge graph using Claude Code skills — explore, search, and ask questions about any project visually.
How do I install understand-anything-knowledge-graph?
Run npx skills add https://github.com/reason-machines/trending-skills --skill understand-anything-knowledge-graph in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/understand-anything-knowledge-graph/ (personal) or .claude/skills/understand-anything-knowledge-graph/ (project).
Which AI agents can use understand-anything-knowledge-graph?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is understand-anything-knowledge-graph safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains understand-anything-knowledge-graph?
It is published by reason-machines in reason-machines/trending-skills. It has 2,442 recorded installs.
More from reason-machines/trending-skills
openclaw-control-center
reason-machines/trending-skills
Local-first, security-first control center for OpenClaw agents — visibility dashboard with readonly defaults, token attribution, collaboration tracing, and safe write operations.
inkos-multi-agent-novel-writing
reason-machines/trending-skills
Multi-agent CLI system for autonomous novel writing, auditing, and revision with human review gates
claude-hud-statusline
reason-machines/trending-skills
A Claude Code plugin that displays a real-time HUD showing context usage, active tools, running agents, and todo progress in your terminal statusline.
agency-agents-ai-specialists
reason-machines/trending-skills
A collection of specialized AI agent personalities for Claude Code, Cursor, Aider, Windsurf, and other AI coding tools — covering engineering, design, marketing, sales, and more.
agent-browser-automation
reason-machines/trending-skills
Headless browser automation CLI for AI agents using native Rust binary with Chrome DevTools Protocol
ui-ux-pro-max-skill
reason-machines/trending-skills
AI design intelligence skill for building professional UI/UX across multiple platforms with 161 reasoning rules, 67 styles, and automated design system generation
Popular code quality & workflow skills
grill-me
mattpocock/skills
A relentless interview to sharpen a plan or design.
improve-codebase-architecture
mattpocock/skills
Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.
handoff
mattpocock/skills
Compact the current conversation into a handoff document for another agent to pick up.
domain-modeling
mattpocock/skills
Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a GLOSSARY.md, or recording or editing an ADR.
codebase-design
mattpocock/skills
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
diagnosing-bugs
mattpocock/skills
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.