Tier 1 · Claude Code internals Claude Code internals · fetched 2026-07-02

cc-learn

An interactive, fully static, bilingual (EN/ZH) documentation site covering Claude Code CLI internals, built from source analysis of v2.1.88 (live at cc.clawlabz.xyz). Turns the raw analyses into navigable technical documentation pages per subsystem.

View source on GitHub

Key takeaways

01
Navigable per-subsystem docs beat raw source dumps for learning
02
English pages make it the most accessible Tier 1 entry point
03
Good model for documenting your own harness

Flows built on this research

Harness Engineering

Version Your Harness and Track Behavior

Treat prompts, tools, and loop logic as versioned, changelogged artifacts - so behavior changes are decisions, not surprises.

4 steps · 45-75 minutes

Agent Architecture

From CLI Harness to Product

Grow a working agent loop into a product: sessions, config, resumability, and the packaging polish the successful terminal agents share.

5 steps · 90-150 minutes

From the archive

Verbatim excerpts mined from our local archive of this repository — the prompts, schemas, and patterns worth stealing.

Insight

Claude Code by the numbers, from restored source

Interactive documentation built from source analysis of @anthropic-ai/claude-code v2.1.88:

| Architecture | 4-layer system design, bootstrap flow, key files |
| Tools | All 39 built-in tools with descriptions |
| Commands | 101 slash commands, searchable |
| Services | 36 services - API, MCP, OAuth, LSP, analytics |
| Security | 3-layer permission model, bash validation, sandbox modes |
| Multi-Agent | Coordinator pattern, concurrency, skills system |

A browsable inventory of one production coding agent's surface area - useful as a checklist of the subsystems a serious harness eventually grows.

README.md