Openclaw Godmode Skill Repo

CC_GodMode 核心能力评估

核心用法

CC_GodMode是一个自编排多智能体开发工作流框架，采用"你说WHAT，AI决定HOW"的设计理念。用户通过简洁命令触发完整开发流程：

触发命令：

• New Feature: [X] - 完整工作流：调研→设计→实现→测试→文档
• Bug Fix: [X] - 快速修复：实现→验证→测试
• API Change: [X] - 安全API变更（强制API守护审查）
• Research: [X] - 技术调研与最佳实践
• Process Issue #X / Prepare Release - issue处理与发布准备

8个专业化Agent分工：

| Agent | 职责 | 模型 |

|-------|------|------|

| @researcher | 技术调研与知识发现 | haiku |

| @architect | 系统架构设计 | opus |

| @api-guardian | API生命周期管理（API变更强制 gate） | sonnet |

| @builder | 全栈实现 | sonnet |

| @validator | 代码质量门（TypeScript、测试、安全） | sonnet |

| @tester | UX质量门（E2E、截图、可访问性、性能） | sonnet |

| @scribe | 技术文档与版本管理 | sonnet |

| @github-manager | GitHub操作（issue/PR/发布） | haiku |

显著优点

1. 双质量门并行验证：@validator（代码）与@tester（UX）并行执行，节省40%验证时间
2. 角色化模型选择：根据任务复杂度智能选择模型（opus决策、sonnet实现、haiku辅助）
3. 强制API安全机制：API路径变更必须通过@api-guardian进行消费者影响分析
4. 完善的报告体系：每个版本自动生成结构化报告到 reports/vX.X.X/
5. 10条黄金规则约束：包括"Version-First"、"No Skipping"、"无许可不推送"等安全设计
6. MCP生态整合：原生支持Playwright、GitHub、Lighthouse、A11y等MCP服务

潜在缺点与局限性

1. 外部依赖重：要求运行时二进制文件、网络访问、凭证配置，部署门槛较高
2. MCP服务依赖：@tester 强制依赖 Playwright MCP，@github-manager 强制依赖 GitHub MCP，缺失则无法完整运行
3. 文档仅包声明：声明为"documentation-only package"，无安装时可执行文件，实际能力依赖运行时指令
4. 超时硬限制：@researcher 30秒硬超时可能导致深度调研中断
5. 预设技术栈：针对React/Node.js/TypeScript优化，其他技术栈适配性未明确
6. 版本文件夹膨胀：每个版本生成7份报告，高频迭代可能导致存储开销

适合人群

• 中小型全栈团队：需要标准化开发流程但缺乏专职架构师
• AI辅助开发实践者：希望将AI从"编码助手"升级为"项目协调者"
• 开源项目维护者：需要自动化issue处理、PR管理和发布流程
• 远程协作团队：通过结构化Agent输出实现异步协作透明度

常规风险

• 凭证泄露风险：GitHub MCP、WebSearch等需要API key，需确保环境隔离
• 代码质量过度依赖AI：双质量门虽完善，但仍需人工最终审查关键变更
• 网络中断导致流程中断：WebFetch、WebSearch、GitHub操作均依赖网络
• 版本冲突："Version-First"规则要求严格遵循语义化版本，团队需统一认知
• 截图存储开销：@tester强制3视口截图可能产生大量图像文件

综合评价

CC_GodMode代表了AI辅助开发向"AI原生软件工程"演进的重要尝试，通过明确的Agent边界、强制质量门和结构化输出，将大型语言模型的非确定性输出转化为可预期的工程流程。适合已具备基础DevOps能力、希望进一步提升AI自动化水平的团队。

CC_GodMode: Multi-Agent Development Orchestration

CC_GodMode is a pure documentation-based orchestration skill that enables self-coordinating multi-agent development workflows. Unlike traditional tools that execute code directly, this skill functions as a sophisticated meta-layer—defining how 8 specialized AI agents collaborate to complete complex development tasks.

Core Architecture

The skill introduces an Orchestrator-Subagent pattern where the primary agent analyzes requests and delegates to specialized workers:

| Agent | Role | Model | Purpose |

|-------|------|-------|---------|

| @researcher | Knowledge Discovery | haiku | Tech evaluation, security research |

| @architect | System Design | opus | High-level architecture decisions |

| @api-guardian | API Lifecycle | sonnet | Breaking change detection |

| @builder | Implementation | sonnet | Full-stack code production |

| @validator | Code Quality | sonnet | TypeScript/tests verification |

| @tester | UX Quality | sonnet | E2E, a11y, performance |

| @scribe | Documentation | sonnet | VERSION/CHANGELOG management |

| @github-manager | GitHub Ops | haiku | Issues, PRs, releases |

Key Innovations

Dual Quality Gates: @validator (code) and @tester (UX) run in parallel for 40% faster validation, with a sync-point decision matrix determining next steps.

Golden Rules: Version-first development, mandatory @api-guardian for API changes, and strict "NEVER git push without permission" policy.

Fail-Safe Design: @researcher has 30-second hard timeouts with graceful degradation; all agents report to versioned reports/vX.X.X/ directories.

Workflow Patterns

• New Feature: research → architect → builder → parallel(validator, tester) → scribe
• Bug Fix: builder → parallel(validator, tester) → done
• API Change: Includes mandatory @api-guardian with consumer impact analysis

Significant Advantages

1. Zero Attack Surface: Pure Markdown documentation—no executable code, no dependencies
2. Transparent Open Source: MIT licensed, fully auditable via GitHub
3. Model-Optimized: Assigns lighter tasks to haiku, complex decisions to opus, balancing cost and quality
4. Production-Ready: Enforces semantic versioning, changelog discipline, and pre-push requirements

Limitations & Considerations

• Token Intensive: Multi-agent parallel execution consumes significant API credits
• MCP Dependencies: Requires optional MCP servers (playwright, github) for full functionality
• Overhead for Simple Tasks: Single-file edits may feel heavyweight
• Runtime Permission Heavy: Frequent user confirmations for Bash, Write, network operations

Risk Assessment

Despite requiring requires_binaries: true and requires_network: true in metadata, these permissions apply to the orchestrated workflows (e.g., @builder running npm test), not the skill itself. The skill's S-grade security rating reflects its document-only nature—all runtime behavior is user-controlled through Claude Code's standard tool approval flow.

Ideal Users

Teams building React/Node.js/TypeScript applications who want standardized, repeatable development workflows with automated quality gates and documentation compliance. Particularly valuable for teams lacking established release processes or those seeking to reduce cognitive load in complex multi-step tasks.