claudecode-godmode-on

内容来源：README.md（说明文档） · 原始地址 · 查看安装指南

原始内容

<div align="center">

# CC_GodMode

### *"What happens when an AI system is used to improve itself?"*

**You're looking at the answer.**

[![Version](https://img.shields.io/badge/Version-8.6.0-blue)](./CHANGELOG.md)
[![Architecture](https://img.shields.io/badge/Architecture-Modular%20%2B%20Skills-green)](./skills/)
[![Agents](https://img.shields.io/badge/Agents-8%20Core%20%2B%201%20Security%20%2B%206%20Dept-purple)](./docs/AGENTS.md)
[![Plugin](https://img.shields.io/badge/Plugin-Ready-orange)](./CLAUDE.md)
[![Ultracode Ready](https://img.shields.io/badge/Ultracode-Ready-brightgreen)](./docs/ARCHITECTURE.md)
[![Self-Improving](https://img.shields.io/badge/Self--Improving-Yes%2C%20Really-red)](./docs/STORY.md)

</div>

---

## The System That Builds Itself

Welcome to the machine shop. Except the machines are building themselves — and they're getting better at it.

**CC_GodMode v7.0.0, v7.1.0, and v8.0.0 weren't built by hand.** They were planned, architected, implemented, validated, documented, and shipped by the exact agents defined in this repo. v8.0.0 went further: it was produced by a **parallel dynamic workflow** — a dozen edit subagents fanned out at once, each adversarially verified by another. The orchestrator delegated, gated, and opened the PR itself.

That's not marketing copy. That's what happened — [the full story is here](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/STORY.md).

---

## Install in 30 Seconds

CC_GodMode is a Claude Code plugin that turns your workflow into a self-orchestrating multi-agent system.

**macOS / Linux:**
```bash
git clone https://github.com/cubetribe/ClaudeCode_GodMode-On.git
cd ClaudeCode_GodMode-On
./scripts/apply-global-claude-setup.sh
./scripts/apply-global-claude-setup.sh --check   # Verify installation
```

**Windows (PowerShell):**
```powershell
git clone https://github.com/cubetribe/ClaudeCode_GodMode-On.git
cd ClaudeCode_GodMode-On
.\scripts\apply-global-claude-setup.ps1
.\scripts\apply-global-claude-setup.ps1 -Check   # Verify installation
```

Then activate in any project:
```bash
cd your-project
cp ~/.claude/templates/CLAUDE-ORCHESTRATOR.md ./CLAUDE.md
claude
```

Done. The orchestrator is active. → **Full setup, MCP servers, and prompt-based fallback: [Installation Guide](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/INSTALLATION.md).**

---

## Daily Usage — Two Steps

The orchestrator loads from `CLAUDE.md` automatically. But its headline power — **parallel-first dynamic workflows — only runs at Ultracode effort, and Ultracode is session-scoped**. So each new session is two steps:

**Step 1 — Turn on Ultracode.** Set it once per session, in the effort selector at the bottom of Claude Code, or by command:

```
/model best        # resolves to Opus 4.8 — the model this system is optimized for
/effort ultracode  # xhigh reasoning + automatic parallel dynamic workflows
```

**Step 2 — Say what you want, prefixed with `GodMode:`**

```
GodMode: New Feature: user authentication with JWT
GodMode: Bug Fix: cart total miscalculates with discount codes
GodMode: Research: best approach for real-time sync in React 18
```

You say *what* you want — the system figures out *which* agents to call, in *what* order, at *what* cost. Ultracode is the one thing that does **not** persist across sessions, so make Step 1 a habit; skip it and GodMode still orchestrates and gates correctly, it just won't fan out to its full parallel width. The trigger `GodMode:` is case-insensitive (`GODMODE:` works too).

---

## What Is This?

**CC_GodMode** transforms Claude Code into a self-orchestrating multi-agent development team — driven by the `best` / Claude Opus 4.8 orchestrator at **ultracode** effort, fanning work out across parallel Claude Code subagents for implementation, validation, and documentation.

**You say WHAT. The AI figures out HOW.**

```
You: "I need user authentication with JWT"

Orchestrator:
  → Analyzes the request and determines the version bump
  → Creates the report folder
  → Delegates to @architect for design
  → Delegates to @api-guardian for API impact
  → Delegates to @builder for implementation
  → @validator checks code quality  ┐ in parallel
  → @tester checks UX quality        ┘
  → @scribe documents everything
  → @github-manager opens the PR

You: *drinks coffee*
```

| Without CC_GodMode | With CC_GodMode |
|:---|:---|
| You: "Design the feature" | You: "Build Feature X" |
| You: "Now implement it" | ☕ |
| You: "Check the types" | ☕ |
| You: "Update the consumers" | ☕ |
| You: "Write the docs" | ☕ |
| You: "Did I forget something?" | AI: "Done. Here's the report." |

---

## Parallel-First & Ultracode

v8.0.0's headline: **parallelization is the default**, not an afterthought.

- **Orchestrator optimized for Opus 4.8 at ultracode** — xhigh reasoning plus automatic dynamic workflows for substantive tasks. The `best` alias resolves to Opus 4.8; if your org has access to a higher tier it is picked up automatically — strictly optional, no feature depends on it.
- **Fan-out by default** — independent units (multi-file edits, audits, migrations, multi-angle research) spawn parallel subagents in a single message; the orchestrator fans in and synthesizes their verdicts.
- **Dynamic-workflows escalation** — when a job outgrows ~10 concurrent subagents, it escalates to tens-to-hundreds of subagents with **adversarial verification** (agents try to refute each other's findings). See [`skills/dynamic-workflows/`](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/skills/dynamic-workflows/SKILL.md).
- **Smart Routing stays the default** — risk-based, minimal-agent paths; ~30–50% token reduction vs. always-Full-Gates. Parallel is *faster, not cheaper*, so max-parallel is a deliberate opt-in.

→ Deep dive: [Architecture](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/ARCHITECTURE.md) · cost model: [Agent Model Selection](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/AGENT_MODEL_SELECTION.md).

---

## The Agents

**15 specialists**, each with one job, a model assignment, and effort tuning:

- **8 core** — `@researcher` `@architect` `@api-guardian` `@builder` `@validator` `@tester` `@scribe` `@github-manager` (always available).
- **1 security gate** — `@security`, for secrets, injection, auth/authz, crypto, and dependency review.
- **6 department** — `@ci-security-guardian` `@docs-dx` `@quality-operations` `@runtime-platform` `@workflow-design` `@workspace-governance` (activate when their domain is in scope).

After `@builder`, the **dual quality gates** — `@validator` (code) and `@tester` (UX) — run **in parallel** and both must pass before `@scribe` documents and ships.

→ Full roster, quality gates, and workflows: **[The Agents](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/AGENTS.md)**.

---

## The Rules

1. **Plan-First** — Non-trivial work starts with a plan (`plans/vX.Y.Z/`) split into sprints with explicit write-scope ownership; the version is decided at release, never at work start (ADR-004)
2. **Smart Routing default** — Risk-based routing; Full-Gates for high-risk signals
3. **Architecture gate (split)** — Inline brief for small/medium; @architect (Opus) for new modules / breaking changes
4. **@api-guardian is MANDATORY** — For any API/schema/type change (enforced by hook)
5. **Dual Quality Gates** — Both @validator AND @tester must pass (parallel execution)
6. **@tester MUST create screenshots** — Every page tested at 3 viewports
7. **No Skipping within the selected path** — Smart Routing picks the minimal set; that set executes fully
8. **Sprint-scoped reports & single-writer hot files** — parallel agents get disjoint write scopes; `VERSION`/`CHANGELOG.md` have exactly one writer (the release tooling / @scribe)
9. **NEVER push without permission** — Applies to ALL agents
10. **Release invariant, machine-checked** — `VERSION == CHANGELOG == tag == GitHub release`, enforced locally and in CI

---

## Documentation

**Guides** (start here):
- **[Installation Guide](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/INSTALLATION.md)** — script install, MCP servers, prompt-based fallback, recovery
- **[Architecture](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/ARCHITECTURE.md)** — parallel-first orchestration, file structure, dual-location model, the hook
- **[The Agents](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/AGENTS.md)** — the 15-agent roster, quality gates, workflows, and modes
- **[The Story & Design Philosophy](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/STORY.md)** — how (and why) the system builds itself
- **[ROADMAP.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/ROADMAP.md)** — living roadmap · **[plans/](https://github.com/cubetribe/ClaudeCode_GodMode-On/tree/HEAD/plans/)** — active plans & sprint files

**Reference:**
- **[CHANGELOG.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/CHANGELOG.md)** — full version history
- **[AGENT_MODEL_SELECTION.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/AGENT_MODEL_SELECTION.md)** — model aliases, pricing, and cost optimization
- **[AGENT_ARCHITECTURE.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/AGENT_ARCHITECTURE.md)** — dual-location install/update/verify procedures
- **[orchestrator/AGENTS.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/AGENTS.md)** — agent registry & handoff matrix
- **[orchestrator/VERSIONING.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/VERSIONING.md)** — **the release law** (single source of truth, invariant, procedures)
- **[orchestrator/WORKFLOWS.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/WORKFLOWS.md)** · **[MODES.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/MODES.md)** · **[QUALITY-GATES.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/QUALITY-GATES.md)** · **[VERSIONING.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/VERSIONING.md)** · **[META-DECISIONS.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/orchestrator/META-DECISIONS.md)**

**Policies:**
- **[REPORT_TEMPLATES.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/templates/REPORT_TEMPLATES.md)** · **[CONTEXT_SCOPE_POLICY.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/policies/CONTEXT_SCOPE_POLICY.md)** · **[SECURITY_TOOLING_POLICY.md](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/policies/SECURITY_TOOLING_POLICY.md)**

---

## FAQ

**Q: Why 15 agents?**
A: 8 core cover the standard workflow, 1 optional `@security` gate activates for security-sensitive changes, and 6 optional department agents activate only when their domain is in scope. Separation of concerns — each agent has ONE job.

**Q: What's the difference between @validator and @tester?**
A: `@validator` = code quality (TypeScript, tests, security). `@tester` = UX quality (E2E, visual, a11y, perf). They run in parallel.

**Q: Can agents push without my permission?**
A: No. "NEVER git push without permission" is enforced across all agents.

→ More, plus the origin story: [The Story & Design Philosophy](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/docs/STORY.md).

---

## Version

**CC_GodMode v8.6.0**

What's in the box:
- **15 agents** (8 core + 1 security gate + 6 department) with effort-field budget tuning
- **13 skills** for workflows, quality gates, release, research, API changes, modes, teams, bootstrap, and dynamic workflows
- **Parallel-first orchestration** — fan-out by default, dynamic-workflows escalation with adversarial verification
- **Dual quality gates** (parallel execution for speed)
- **Smart Routing by default** (~30–50% token savings vs. old always-Full-Gates)
- **Version-first workflow** with automated checks

See the [CHANGELOG](https://github.com/cubetribe/ClaudeCode_GodMode-On/blob/HEAD/CHANGELOG.md) for the full history.

---

## Credits

**Dennis Westermann** ([www.dennis-westermann.de](https://www.dennis-westermann.de))
*Years of suffering, distilled into this repo. Now the repo improves itself. Was it worth it?*

---

## License

**Proprietary License** — Private use permitted. Commercial use requires permission.

Copyright (c) 2025 Dennis Westermann

---

<div align="center">

**Made with mass sleep deprivation**

*The experiment continues.*

⭐ Star if you're not too unsettled ⭐

</div>

CC_GodMode

"What happens when an AI system is used to improve itself?"

You're looking at the answer.

The System That Builds Itself

Welcome to the machine shop. Except the machines are building themselves — and they're getting better at it.

CC_GodMode v7.0.0, v7.1.0, and v8.0.0 weren't built by hand. They were planned, architected, implemented, validated, documented, and shipped by the exact agents defined in this repo. v8.0.0 went further: it was produced by a parallel dynamic workflow — a dozen edit subagents fanned out at once, each adversarially verified by another. The orchestrator delegated, gated, and opened the PR itself.

That's not marketing copy. That's what happened — the full story is here.

Install in 30 Seconds

CC_GodMode is a Claude Code plugin that turns your workflow into a self-orchestrating multi-agent system.

macOS / Linux:

git clone https://github.com/cubetribe/ClaudeCode_GodMode-On.git
cd ClaudeCode_GodMode-On
./scripts/apply-global-claude-setup.sh
./scripts/apply-global-claude-setup.sh --check   # Verify installation

Windows (PowerShell):

git clone https://github.com/cubetribe/ClaudeCode_GodMode-On.git
cd ClaudeCode_GodMode-On
.\scripts\apply-global-claude-setup.ps1
.\scripts\apply-global-claude-setup.ps1 -Check   # Verify installation

Then activate in any project:

cd your-project
cp ~/.claude/templates/CLAUDE-ORCHESTRATOR.md ./CLAUDE.md
claude

Done. The orchestrator is active. → Full setup, MCP servers, and prompt-based fallback: Installation Guide.

Daily Usage — Two Steps

The orchestrator loads from CLAUDE.md automatically. But its headline power — parallel-first dynamic workflows — only runs at Ultracode effort, and Ultracode is session-scoped. So each new session is two steps:

Step 1 — Turn on Ultracode. Set it once per session, in the effort selector at the bottom of Claude Code, or by command:

/model best        # resolves to Opus 4.8 — the model this system is optimized for
/effort ultracode  # xhigh reasoning + automatic parallel dynamic workflows

Step 2 — Say what you want, prefixed with GodMode:

GodMode: New Feature: user authentication with JWT
GodMode: Bug Fix: cart total miscalculates with discount codes
GodMode: Research: best approach for real-time sync in React 18

You say what you want — the system figures out which agents to call, in what order, at what cost. Ultracode is the one thing that does not persist across sessions, so make Step 1 a habit; skip it and GodMode still orchestrates and gates correctly, it just won't fan out to its full parallel width. The trigger GodMode: is case-insensitive (GODMODE: works too).

What Is This?

CC_GodMode transforms Claude Code into a self-orchestrating multi-agent development team — driven by the best / Claude Opus 4.8 orchestrator at ultracode effort, fanning work out across parallel Claude Code subagents for implementation, validation, and documentation.

You say WHAT. The AI figures out HOW.

You: "I need user authentication with JWT"

Orchestrator:
  → Analyzes the request and determines the version bump
  → Creates the report folder
  → Delegates to @architect for design
  → Delegates to @api-guardian for API impact
  → Delegates to @builder for implementation
  → @validator checks code quality  ┐ in parallel
  → @tester checks UX quality        ┘
  → @scribe documents everything
  → @github-manager opens the PR

You: *drinks coffee*

Without CC_GodMode	With CC_GodMode
You: "Design the feature"	You: "Build Feature X"
You: "Now implement it"	☕
You: "Check the types"	☕
You: "Update the consumers"	☕
You: "Write the docs"	☕
You: "Did I forget something?"	AI: "Done. Here's the report."

Parallel-First & Ultracode

v8.0.0's headline: parallelization is the default, not an afterthought.

Orchestrator optimized for Opus 4.8 at ultracode — xhigh reasoning plus automatic dynamic workflows for substantive tasks. The best alias resolves to Opus 4.8; if your org has access to a higher tier it is picked up automatically — strictly optional, no feature depends on it.
Fan-out by default — independent units (multi-file edits, audits, migrations, multi-angle research) spawn parallel subagents in a single message; the orchestrator fans in and synthesizes their verdicts.
Dynamic-workflows escalation — when a job outgrows ~10 concurrent subagents, it escalates to tens-to-hundreds of subagents with adversarial verification (agents try to refute each other's findings). See skills/dynamic-workflows/.
Smart Routing stays the default — risk-based, minimal-agent paths; ~30–50% token reduction vs. always-Full-Gates. Parallel is faster, not cheaper, so max-parallel is a deliberate opt-in.

→ Deep dive: Architecture · cost model: Agent Model Selection.

The Agents

15 specialists, each with one job, a model assignment, and effort tuning:

8 core — @researcher @architect @api-guardian @builder @validator @tester @scribe @github-manager (always available).
1 security gate — @security, for secrets, injection, auth/authz, crypto, and dependency review.
6 department — @ci-security-guardian @docs-dx @quality-operations @runtime-platform @workflow-design @workspace-governance (activate when their domain is in scope).

After @builder, the dual quality gates — @validator (code) and @tester (UX) — run in parallel and both must pass before @scribe documents and ships.

→ Full roster, quality gates, and workflows: The Agents.

The Rules

Plan-First — Non-trivial work starts with a plan (plans/vX.Y.Z/) split into sprints with explicit write-scope ownership; the version is decided at release, never at work start (ADR-004)
Smart Routing default — Risk-based routing; Full-Gates for high-risk signals
Architecture gate (split) — Inline brief for small/medium; @architect (Opus) for new modules / breaking changes
@api-guardian is MANDATORY — For any API/schema/type change (enforced by hook)
Dual Quality Gates — Both @validator AND @tester must pass (parallel execution)
@tester MUST create screenshots — Every page tested at 3 viewports
No Skipping within the selected path — Smart Routing picks the minimal set; that set executes fully
Sprint-scoped reports & single-writer hot files — parallel agents get disjoint write scopes; VERSION/CHANGELOG.md have exactly one writer (the release tooling / @scribe)
NEVER push without permission — Applies to ALL agents
Release invariant, machine-checked — VERSION == CHANGELOG == tag == GitHub release, enforced locally and in CI

Documentation

Guides (start here):

Installation Guide — script install, MCP servers, prompt-based fallback, recovery
Architecture — parallel-first orchestration, file structure, dual-location model, the hook
The Agents — the 15-agent roster, quality gates, workflows, and modes
The Story & Design Philosophy — how (and why) the system builds itself
ROADMAP.md — living roadmap · plans/ — active plans & sprint files

Reference:

CHANGELOG.md — full version history
AGENT_MODEL_SELECTION.md — model aliases, pricing, and cost optimization
AGENT_ARCHITECTURE.md — dual-location install/update/verify procedures
orchestrator/AGENTS.md — agent registry & handoff matrix
orchestrator/VERSIONING.md — the release law (single source of truth, invariant, procedures)
orchestrator/WORKFLOWS.md · MODES.md · QUALITY-GATES.md · VERSIONING.md · META-DECISIONS.md

Policies:

REPORT_TEMPLATES.md · CONTEXT_SCOPE_POLICY.md · SECURITY_TOOLING_POLICY.md

FAQ

Q: Why 15 agents? A: 8 core cover the standard workflow, 1 optional @security gate activates for security-sensitive changes, and 6 optional department agents activate only when their domain is in scope. Separation of concerns — each agent has ONE job.

Q: What's the difference between @validator and @tester? A: @validator = code quality (TypeScript, tests, security). @tester = UX quality (E2E, visual, a11y, perf). They run in parallel.

Q: Can agents push without my permission? A: No. "NEVER git push without permission" is enforced across all agents.

→ More, plus the origin story: The Story & Design Philosophy.

Version

CC_GodMode v8.6.0

What's in the box:

15 agents (8 core + 1 security gate + 6 department) with effort-field budget tuning
13 skills for workflows, quality gates, release, research, API changes, modes, teams, bootstrap, and dynamic workflows
Parallel-first orchestration — fan-out by default, dynamic-workflows escalation with adversarial verification
Dual quality gates (parallel execution for speed)
Smart Routing by default (~30–50% token savings vs. old always-Full-Gates)
Version-first workflow with automated checks

See the CHANGELOG for the full history.

Credits

Dennis Westermann (www.dennis-westermann.de) Years of suffering, distilled into this repo. Now the repo improves itself. Was it worth it?

License

Proprietary License — Private use permitted. Commercial use requires permission.

Made with mass sleep deprivation

The experiment continues.

⭐ Star if you're not too unsettled ⭐