Step 1
Source analysis
Map entrypoints, runtime boundaries, tools, permissions, MCP, remote capabilities, and standout implementations.
Execution Plan
This site is built around real Claude Code source analysis, not generic product copy. The current study uses a local app-source snapshot for internals and the public GitHub repository for plugin and ecosystem context.
Step 1
Map entrypoints, runtime boundaries, tools, permissions, MCP, remote capabilities, and standout implementations.
Step 2
Translate the analysis into architecture maps, module explainers, execution flows, reading tracks, and concept pages.
Step 3
Export a static Next.js site and deploy it automatically to GitHub Pages from the repository workflow.
Why this project matters
Claude Code behaves like a platformized terminal product with a query runtime, concurrency-aware tool execution, policy layers, plugins, skills, MCP, and remote modes.
Source basis
Snapshot date: 2026-03-31. Public ecosystem repo commit: 4411cba.
Recommended entry
That route gives the fastest path from product shape to concrete execution logic.
Draw.io-style overview
Analysis Focus
The main learning payoff is not one file. It is the relationship between runtime routing, the query loop, the tool runtime, the permission system, and the extensibility model.
Start with how many personalities the product has: interactive CLI, headless engine, bridge process, remote path, and MCP server.
Follow one turn from prompt input to tool execution and transcript persistence.
Pay attention to permission modes, settings layers, and auth paths. They explain how autonomy stays governable.
Architecture at a glance
This map mirrors the actual code structure and becomes the backbone of the rest of the learning site.
Fast-path entrypoints decide whether Claude Code behaves like a full interactive CLI, a headless engine, a bridge process, or an MCP server.
The REPL and supporting UI state render conversation history, task status, tool output, permissions, and remote controls.
User input is normalized, streamed through query orchestration, and converted into ordered, concurrency-aware tool execution.
Commands, skills, plugins, MCP servers, and LSP integrations are all loaded into the same runtime surface and filtered by policy.
A deep settings and permission system governs what the runtime is allowed to do, how it authenticates, and how sessions resume safely.
Claude Code can act as a remote-control target, manage remote sessions, sync team memory, and expose itself over MCP.
Read this next
Get the boundaries right before diving into subsystems.
Then explore
Study role, key files, inputs, outputs, and learning value.
Then simulate
Trace runtime chains step by step and connect them to source files.
Finish with
Choose a route by experience level or technical interest.
30-minute route
Read `cli.tsx` → `main.tsx` → `commands.ts` → `tools.ts` → Architecture page diagrams.
2-hour route
Follow `processUserInput.ts` → `query.ts` → `QueryEngine.ts` → tool orchestration files → session storage.
Platform route
Focus on permissions, auth, config, plugin loading, and MCP client logic.