coco

<div align="center"> <br> <img src="assets/logo-light.svg" alt="Coco" width="160" height="220"> # Coco: The Universal AI Agent Orchestration Framework ### Transform your AI coding assistant into a high-performance engineering team, backed by an advisory board of 389 world-class minds. Free | MIT License | 90-Second Onboarding <br> [](https://opensource.org/license/mit) [](CHANGELOG.md) [](skills/) [](commands/) [](systems/superintelligence/) [](https://github.com/rkz91/coco/actions) <br> <p> &nbsp;&nbsp;<a href="#install"><kbd> &nbsp; Install &nbsp; </kbd></a>&nbsp;&nbsp; <a href="#core-feature-coco-superintelligence"><kbd> &nbsp; Superintelligence &nbsp; </kbd></a>&nbsp;&nbsp; <a href="#three-pillars"><kbd> &nbsp; Core Architecture &nbsp; </kbd></a>&nbsp;&nbsp; <a href="#skills-catalog-deep-dive"><kbd> &nbsp; Skills &nbsp; </kbd></a>&nbsp;&nbsp; <a href="#system-bundles-deep-dive"><kbd> &nbsp; Systems &nbsp; </kbd></a>&nbsp;&nbsp; <a href="#see-it-run"><kbd> &nbsp; Demos &nbsp; </kbd></a> </p> <br> </div> <p align="center"> <img src="assets/demo.gif" alt="/SI-Decide --debate: a routed expert panel argues in reacting rounds before an attributed verdict" width="720" /> </p> <div align="center">_{<em>A real <code>/SI-Decide --debate</code> run — Coco routes an expert panel, deliberates in reacting rounds, and returns an attributed verdict with named dissent.</em>}</div> --- <div align="center"> > **One developer writing single prompts? That was last year.** > > Coco turns your AI coding session into an entire engineering department. It spawns specialized agents, runs parallel code tasks, coordinates complex review panels, persists phase state across resets, and enforces deterministic verification gates before code is committed. </div> ```bash git clone https://github.com/rkz91/coco.git && cd coco && bash install.sh ``` <div align="center">_{90 seconds. That's all it takes to wire Coco's workspace engine into your IDE.}</div> --- ## Core Feature: Coco Superintelligence ### A Cross-Team Advisory Board of 389 World-Class Minds Coco's signature differentiator is Superintelligence. It allows you to summon custom, real-world expert panels directly inside your coding sessions to guide architectural, engineering, risk, finance, and business decisions: ```bash /SI-Decide "Should we migrate our database to pgvector?" /SI-AI-Decide "Which embedding model gives the best cost/performance?" /SI-Eng-Pre-Mortem "Review our zero-downtime cache migration strategy" /SI-GRC-Review "Check this customer onboarding flow for GDPR compliance" ``` * **389 Expert Personas**: Organized across 9 specialized departments: * **Engineering** (70 experts): Experts in software architecture, cloud systems, systems coding, compilers, and infrastructure. * **AI** (59 experts): Neural network research, LLM optimization, safety, alignment, and vector database engineering. * **Product and Design** (56 experts): Design systems, UX design, product strategy, and growth loops. * **Finance** (47 experts): Valuation, corporate finance, macroeconomic modeling, and fintech infrastructure. * **Trading** (46 experts): Quantitative analysis, market microstructure, derivatives pricing, and crypto liquidity. * **Risk and Compliance / GRC** (30 experts): GDPR, HIPAA, SOC2 compliance, security auditing, and international regulations. * **Strategy** (29 experts): Platform economics, business models, competitive analysis, and strategic growth. * **Data and Analytics** (29 experts): Data engineering, pipeline optimization, predictive modeling, and analytics stacks. * **Sales / GTM / Marketing** (23 experts): Product-market fit, sales operations, growth marketing, and enterprise GTM. * **242 Generated Slash Commands**: Comprises 17 Cross-Team Meta Commands (like `/SI-Decide` or `/SI-Tradeoff` routing across multiple domains) and 225 Per-Team Commands (25 per team, such as `/SI-AI-Decide` or `/SI-Eng-Pre-Mortem`). * **Cross-Team Meta-Orchestration**: For complex queries spanning domains ("build and sell an AI audit tool"), a Stage-A local router uses nomic-embed to score the prompt against team registries and cells, greedily picking a proportional 16-32 person panel. * **Citable and Grounded**: Every stance is verified from public sources and carries a direct evidence URL. Personas are validated against a strict anti-fabrication gate. ### The Selection Algorithm (Two-Stage Routing + Approval) The Meta-Orchestrator uses a staged algorithm to avoid loading massive file systems at runtime — two routing stages (team/cell, then scoped personas) followed by an approval-and-execution step: 1. **Stage A: Team and Cell Routing (Cheap)**: Reads the meta-registry (`superintelligence/registry.json`) and scans only the team descriptions and cell definitions. It scores team relevance using keyword and domain overlap, selecting the top 1-4 teams. If a single team dominates, it delegates to that team's normal orchestrator. 2. **Stage B: Scoped Persona Selection (Scoped)**: Loads compact persona records (slugs, cells, domains, stances, conflict mappings) for only the selected teams. It allocates the 16-32 panel budget proportionally to team relevance, runs the greedy scorer (`0.40·domain + 0.30·cell-coverage + 0.30·conflict-pairing`), and applies a cross-team tension pass to pair opposing viewpoints (e.g., security vs. growth). 3. **Stage C: Approval and Execution**: Presents the roster grouped by team with a one-line selection rationale. The executing action verb consumes the approved roster, and the resulting output includes full line-by-line attribution (e.g., "Aswath Damodaran (Finance): ..."). ### Persona Build and Validation Pipeline The 389-persona roster was compiled using a systematic multi-tier workflow: - **Generation**: Evaluated Local Qwen 3.6 (via LM Studio) -> GPT-5-Nano (via QB Gateway) -> Gemini 3.5 Flash -> Claude Research Agents (web search). Claude research agents proved to be the quality winner for resolving historical data and cited Signal. - **Validation Gate**: Run via `validate_persona.py`, enforcing that every persona must have at least 4 live, non-404 cited evidence URLs, a valid home team, no fabricated quotes, and at least 2 verified signals within the last 12 months (or archetype persistent signals for historical figures). --- ## The Coco Asset Library A standard installation equips your workspace with a lightweight set of tools, while full activation unlocks up to 864 total assets to orchestrate any software engineering workflow: <table align="center"> <tr> <td align="center" width="20%"><h3>146</h3>_Skills<br><small>63 Core + 83 Bundle</small></td> <td align="center" width="20%"><h3>277</h3>_{Slash Commands}<br><small>35 Core + 242 Bundle</small></td> <td align="center" width="20%"><h3>34</h3>_{Specialized Agents}<br><small>10 Core + 24 Bundle</small></td> <td align="center" width="20%"><h3>389</h3>_{Expert Personas}<br><small>Superintelligence Board</small></td> <td align="center" width="20%"><h3>15</h3>_{Cross-IDE Rules}<br><small>Cursor MDC Rules</small></td> </tr> </table> <div align="center"> _{<strong>Core Installation:</strong> 123 active assets (63 Skills, 35 Commands, 10 Agents, 15 Rules)}<br> _{<strong>Orchestration Bundles:</strong> <strong>+68 GSD skills</strong> · <strong>+24 GSD agents</strong> · <strong>+6 Brain skills</strong> · <strong>+9 Superintelligence skills</strong> · <strong>+242 SI commands</strong> · <strong>3 Workflows</strong>} </div> --- ## Three Pillars of Coco <table> <tr> <td width="33%" valign="top"> ### 1. Autonomous Team Execution `/team:ship` initiates a fully coordinated build pipeline consisting of 6 development stages and 7 hard verification gates. Specialized agents act in their specific domains (Research, Architecture, Plan, Review, Build, QA). Crucially, code changes must clear the Test Evidence Protocol (independent re-runs, strict coverage metrics, and skip-prevention) before opening a pull request. </td> <td width="33%" valign="top"> ### 2. Disk-Persistent State AI chat contexts are fragile and wipe on `/clear`. Coco solves this with 68 disk-backed orchestration skills that persist phase state, decisions, and progress logs across resets. Includes atomic git commits for every successful step, allowing you to review history, roll back bad design paths, or resume long-running migrations from any checkpoint. </td> <td width="33%" valign="top"> ### 3. Vendor-Neutral Portability Coco compiles its rules, templates, and agent definitions into pure Markdown and YAML frontmatter. Using native IDE adapters, the same skill library is injected into Claude Code, Cursor, Codex CLI, or any AGENTS.md compatible environment. Keep your workflows, scripts, and commands even if you switch your AI editor in the future. </td> </tr> </table> --- ## Skills Catalog Deep-Dive Coco comes loaded with 146 skills. These are not basic prompt templates, but fully realized agent instructions containing state management, verification logic, and error handlers: ### Visual Design and Styling Systems * **ui-ux-pro-max**: Generates modern visual presets containing 50+ styling tokens, 21 color palettes, font pairings, and responsive charts. * **axiom-liquid-glass**: Implements high-end glassmorphism visual styling specifications using HSL tailored colors, backdrop filters, and subtle border shadows. * **swiftui-liquid-glass**: Tailors the visual glassmorphic design system specifically for Apple SwiftUI layouts. * **web-design-guidelines**: Enforces strict layout grids, typography scales, responsive query setups, and color harmony constraints. * **tailwind-patterns**: A library of production-ready, accessible Tailwind CSS layout patterns (navigation, tables, carousels, cards). ### Enterprise Engineering and Architecture * **c4-architecture**: Guides the creation of system architecture diagrams following the C4 model (Context, Container, Component, Code). * **api-design-principles**: Enforces RESTful and gRPC API standards, focusing on clean resource naming, status code mappings, and error shapes. * **arb-review**: Prepares codebases and designs for Architecture Review Board audits, assessing scalability, security, and portability. * **vercel-react-best-practices**: Enforces Next.js App Router rules, React Server Components boundaries, and optimized hook placement. * **subagent-driven-development**: Instructs the primary AI on how to delegate complex engineering tasks to sandboxed, specialized subagents. ### Document and Spreadsheet Processing * **xlsx**: Creates and formats data-dense Excel spreadsheets locally using Python openpyxl scripts. * **pdf**: Generates structured PDFs (reports, specs, invoices) with custom tables and headers. * **docx**: Assembles word documents with standard paragraph, heading, and table styling. * **doc-sync**: Automates synchronization between code implementations and project documentation (PRDs, readmes, APIs). ### Operations, Safety, and Disaster Recovery * **dr-plan**: Generates disaster recovery specifications containing RTO/RPO targets, failover steps, and data integrity verification. * **recovery-plan**: Formulates active incident response plans to restore services after database corruption or server failure. * **using-git-worktrees**: Sets up parallel, isolated development branches on the local disk using git worktrees. * **finishing-a-development-branch**: Automates the PR prep process: cleans local branch, squashes commits, runs lint, updates changelogs, and opens PR. ### Product Management and Communication * **prd-mastery**: Generates deep Product Requirement Documents containing user stories, scope limits, metrics, and risk mitigations. * **stakeholder-comms**: Generates clear, non-technical progress summaries and release notes tailored to stakeholders. * **nfr-tracker**: Audits and tracks Non-Functional Requirements (latency, cost, throughput, security) against defined SLAs. * **brainstorming**: Conducts structured brainstorming sessions (SCAMPER, First Principles, Mind Mapping) and outputs formatted markdown tables. --- ## System Bundles Deep-Dive System bundles are opt-in packages that extend the workspace engine with specialized databases, pipelines, and persona rosters: ### 1. GSD (Get Shit Done) Bundle An orchestration engine consisting of **68 skills and 24 agents** that manages the lifecycle of complex codebases. * **State Database**: Uses a disk-backed JSON/YAML phase database that records active goals, decisions, milestones, and blockages. * **Workstreams**: Spins up parallel workspaces (isolated git checkouts) to test refactors or debug features in sandbox environments without polluting main. * **Forensics**: Performs post-mortem audits when a phase fails. Scans commit histories, state files, and agent logs to generate structured root-cause reports. * **Autonomous Mode**: Bypasses confirmation steps, allowing parallel waves of subagents to execute plan milestones end-to-end. ### 2. Brain Bundle A local knowledge graph engine consisting of **6 skills** that connects email, chat, code, and documentation. * **Local SQLite Store**: Indexes code entities (classes, components), business decisions, project terms, and stakeholder feedback. * **Wiki Generator**: Automatically builds local, hyperlinked wiki articles for every recorded entity in the project. * **Mail Sync**: Ingests and links project email threads directly to related code files. ### 3. Team Bundle A workspace multi-agent pipeline engine. * **Development Loop**: Spawns specialized subagents (Research, Architect, QA) to execute changes, review diffs, and write unit tests. * **Deterministic Gates**: Intercepts merges and commits to run lints, import audits, reference checks, and verification suites. --- ## YOLO Mode: True Autonomy <table> <tr> <td valign="top"> ```bash /coco yolo ``` </td> <td valign="top"> **Bypass approval gates and let the team run unattended.**<br> _{Activates autonomous multi-stage execution. Pipelines transition from phase to phase without requesting manual user confirmations. Combine with <code>/gsd-autonomous</code> and <code>--systems gsd</code> to run code migrations or test sweeps overnight. Easily revert to safety using <code>/coco careful</code> or <code>/coco normal</code>.} </td> </tr> </table> --- ## Before Coco vs. After Coco <table> <tr> <th width="20%">Scenario</th> <th width="38%">Standard AI Assistant (Claude Code / Cursor / Codex)</th> <th width="42%">Coco-Orchestrated Workspace</th> </tr> <tr> <td><strong>Feature Development</strong></td> <td>Generates a single component, misses side-effects, requires manual guidance.</td> <td>Scans repo, designs interface, plans phases, executes parallel build agents, tests, and verifies.</td> </tr> <tr> <td><strong>Code Auditing</strong></td> <td>Suggests basic syntax fixes; misses architectural constraints.</td> <td>Applies a 7-category audit (TDZ errors, import mismatches, dead code, CSS regressions, and mock leaks).</td> </tr> <tr> <td><strong>System Debugging</strong></td> <td>Repeats similar suggestions; loses track of prior attempts.</td> <td>Systematic isolation loop (Reproduce → Isolate → Hypothesize → Fix → Verify) with saved checkpoints.</td> </tr> <tr> <td><strong>Website Cloning</strong></td> <td>Outputs a generic layout with broken assets.</td> <td>Extracts exact design tokens (colors, radii, spacing, type scale) into a production-ready template.</td> </tr> <tr> <td><strong>Context Resets</strong></td> <td>Forgets active plan; starts over from scratch.</td> <td>Reads workspace state from disk and resumes the active phase immediately.</td> </tr> </table> --- ## High-Impact Command Demos <table> <tr> <td width="50%" valign="top"> ### Website Re-engineering ```bash /clone-website https://stripe.com ``` > Extracts colors, typography, spacing, shadows, and SVG assets directly from a live URL into a single-file, responsive HTML template. </td> <td width="50%" valign="top"> ### Code-Integrity Verification ```bash /code-verification ``` > Audits fresh code against 7 bug vectors before execution, preventing broken references, imports, and mock leakage. </td> </tr> <tr> <td valign="top"> ### Systematic Debugging ```bash /systematic-debugging ``` > Drives a methodical root-cause isolation workflow. Records hypotheses and findings, preserving progress across API limits. </td> <td valign="top"> ### Spec & PRD Generation ```bash /prd-generator ``` > Generates a comprehensive, 13-section Product Requirement Document (User Stories, Metrics, Risks, Architecture). </td> </tr> <tr> <td valign="top"> ### Design Token Extraction ```bash /ui-ux-pro-max ``` > Generates visual presets: 50+ modern styling tokens, 21 color palettes, font pairings, and charts configured for your stack. </td> <td valign="top"> ### Personal Knowledge Base ```bash bash install.sh --systems brain /brain-init && /brain-update ``` > Builds a local, SQLite-backed entity graph of your project. Automatically indexes decisions, terms, and file relations. </td> </tr> </table> --- ## Power Features Overview <table> <tr> <td width="50%" valign="top"> ### CoCo 会话路由器 ```bash /coco ``` > 一个中央控制面板，可以自动将类似“什么阻塞了？”，“总结文件”，或“运行测试”等查询路由到正确的子命令。 </td> <td width="50%" valign="top"> ### 多模态记忆 ```bash /media-memory ``` > 本地嵌入式数据库（Gemini Embedding + ChromaDB），索引、搜索和关联图表、音频和文档与您的代码。 </td> </tr> <tr> <td valign="top"> ### 调度和循环代理 ```bash # 主机原生调度（例如 Claude Code）指向任何 Coco 命令： /schedule run /team:review every Monday /loop 5m /team:verify ``` > CoCo 不自带调度器。如果您的主机 CLI 提供调度或循环（例如 Claude Code 的 `/schedule` 和 `/loop`），请将其指向任何 CoCo 命令 — `/team:review`、`/team:verify`、`/code-verification` — 以运行无人值守的审查或监视。 </td> <td valign="top"> ### 电子邮件分类套件 ```bash /email:summary /email:today ``` > 读取您的活动 Outlook 实例（旧版 AppleScript 或新 Outlook MIME 缓存）以获取、搜索、线程和草拟项目邮件的回复。 </td> </tr> <tr> <td valign="top"> ### 工作区分支 ```bash /gsd-workstreams create feature-a /gsd-new-workspace ``` > 在同一存储库中分支开发工作流。启动具有独立项目状态的隔离、沙盒克隆。 </td> <td valign="top"> ### AGENTS.md 编译器 ```bash bash adapters/codex/install.sh ``` > 将所有活动规则、技能和命令定义编译成 Aider、Cline 和 Continue 识别的标准 `AGENTS.md` 文件。 </td> </tr> </table> --- ## CoCo 工作原理 ```mermaid flowchart LR U["您<br/>(斜杠命令、纯英文)"] --> AI["您的 AI 助手<br/>(Claude Code、Cursor 等)"] AI --> A["IDE 适配器"] A --> S["CoCo 资产<br/>技能 · 代理 · 命令 · 规则"] S -. 读取运行时 .-> AI AI --> 代理["并行子代理<br/>(执行波)"] 代理 --> AI ``` --- ## 安装矩阵 <table> <tr> <td width="50%" valign="top"> **标准 90 秒设置** ```bash git clone https://github.com/rkz91/coco.git cd coco bash install.sh ``` *安装程序将自动检测您活动的 IDE/CLI。* </td> <td width="50%" valign="top"> **通过全局 npm 包** ```bash # 身份验证 GitHub Packages 注册表 echo "@rkz91:registry=https://npm.pkg.github.com" >> ~/.npmrc echo "//npm.pkg.github.com/:_authToken=YOUR_TOKEN" >> ~/.npmrc # 安装并启动 npm install -g @rkz91/coco-cli coco ``` </td> </tr> <tr> <td valign="top"> **激活编排包** ```bash # 启用 GSD、Brain 和 Team 系统 bash install.sh --systems gsd,brain,team # 启用超级智能角色 bash install.sh --systems superintelligence ``` </td> <td valign="top"> **手动适配器覆盖** ```bash bash install.sh --adapter claude-code bash install.sh --adapter cursor bash install.sh --adapter codex bash install.sh --adapter generic ``` </td> </tr> </table> --- ## 保持最新 CoCo 会在有新版本时通知您 — 它最多每天一次检查 GitHub 存储库（源），打印一行横幅，并发送 **无遥测数据**。使用 `COCO_NO_UPDATE_CHECK=1` 禁用。 <table> <tr> <td width="50%" valign="top"> **git 克隆** ```bash # 打印版本 + 更新状态 bash scripts/check-update.sh # 应用更新 git pull --ff-only && bash install.sh ``` </td> <td width="50%" valign="top"> **npm 安装** ```bash # 打印版本 + 检查更新 npx @rkz91/coco-cli version # 应用更新 npx @rkz91/coco-cli update ``` </td> </tr> </table> _{发行版本在 [`CHANGELOG.md`](CHANGELOG.md) 和 <a href="https://github.com/rkz91/coco/releases">GitHub Releases</a> 页面上跟踪。} --- ## 系统兼容性 <table> <tr> <th>AI 工具 / CLI</th> <th>适配器名称</th> <th>状态</th> <th>详细信息</th> </tr> <tr> <td><strong><a href="https://docs.anthropic.com/en/docs/claude-code">Claude Code</a></strong></td> <td><code>claude-code</code></td> <td>稳定</td> <td>完全支持斜杠命令、代理和本地设置。</td> </tr> <tr> <td><strong><a href="https://cursor.com/">Cursor</a></strong></td> <td><code>cursor</code></td> <td>稳定</td> <td>链接 MDC 规则和自定义工作区技能。</td> </tr> <tr> <td><strong><a href="https://github.com/openai/codex">Codex CLI</a></strong></td> <td><code>codex</code></td> <td>稳定</td> <td>将资产编译到根上下文中。</td> </tr> <tr> <td><strong>Aider / Cline / Windsurf</strong></td> <td><code>generic</code></td> <td>稳定</td> <td>生成可移植的根级 <code>AGENTS.md</code> 定义。</td> </tr> <tr> <td><strong>VS Code (Continue)</strong></td> <td><code>vscode-continue</code></td> <td>实验性（存根）</td> <td>仅限脚手架 — 完整的线路跟踪在 <a href="https://github.com/rkz91/coco/issues/4">#4</a> 中，目标是 v0.2。</td> </tr> <tr> <td><strong>Antigravity (Google)</strong></td> <td><code>antigravity</code></td> <td>计划中</td> <td>计划在 v0.2 版本中集成。</td> </tr> </table> --- ## 技术规格 <table> <tr><td><strong>规格版本</strong></td><td>1.0.0</td></tr> <tr><td><strong>许可证</strong></td><td><a href="https://opensource.org/license/mit">MIT</a></td></tr> <tr><td><strong>可用技能总数</strong></td><td>146（安装所有包时）63 核心 + 68 GSD + 6 脑 + 9 超智能</td></tr> <tr><td><strong>斜杠命令</strong></td><td>277（安装所有包时）- 35 核心（已发货）+ 242 超智能（225 每团队 + 17 跨团队，在超级智能包安装时生成）</td></tr> <tr><td><strong>专用代理</strong></td><td>34（10 核心 + 24 GSD 包）</td></tr> <tr><td><strong>系统包</strong></td><td>4（GSD、Brain、Team、Superintelligence）— 使用 <code>--systems &lt;name&gt;</code> 选择加入</td></tr> <tr><td><strong>跨 IDE 规则</strong></td><td>15（.mdc 文件）</td></tr> <tr><td><strong>定义的 工作流</strong></td><td>3（.md 管道）</td></tr> <tr><td><strong>安装时间</strong></td><td>&le; 90 秒</td></tr> <tr><td><strong>遥测 / SaaS</strong></td><td>无 — 100% 本地文件</td></tr> </table> _{核心安装提供 63 技能 + 35 命令 + 10 代理。上述计数反映了启用所有四个系统包（<code>bash install.sh --systems gsd,brain,team,superintelligence</code>）的完整安装。超级智能斜杠命令在安装时从团队注册表中本地生成 — 不会传输或远程存储命令文件。} --- ## 常见问题 <details> <summary><strong>我的代码库或数据是否发送给第三方？</strong></summary> 不。CoCo 完全是本地的。它由 markdown 文件、配置文件和 shell 脚本组成，驻留在您的机器上。您现有的 AI 工具的隐私政策保持不变。 </details> <details> <summary><strong>如何修改或自定义现有技能？</strong></summary> 每个技能都在 `skills/<name>/SKILL.md` 下定义。您可以直接编辑此文件以完善提示指令或添加自定义工具，然后运行 `bash install.sh` 以更新符号链接。 </details> <details> <summary><strong>如何干净地卸载 CoCo？</strong></summary> 由于 CoCo 使用符号链接，因此删除它完全是无损的： ```bash find ~/.claude ~/.cursor -type l -lname "*$(pwd)*" -delete ``` 这仅删除指向您的 CoCo 存储库文件夹的链接。 </details> <details> <summary><strong>如果我以后更换 AI 助手，会发生什么？</strong></summary> 只需使用新的适配器标志（例如 `bash install.sh --adapter cursor`）重新运行安装程序。您的所有自定义技能和工作流将立即随之而来。 </details> --- ## 贡献与社区 CoCo 采用 MIT 许可，欢迎贡献。 - **从这里开始：** [`CONTRIBUTING.md`](CONTRIBUTING.md) — 如何添加技能、命令、代理或适配器。 - **好起步问题：** [帮助 / 好起步问题](https://github.com/rkz91/coco/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) — 新手的小型、范围明确的任务。 - **报告错误 / 请求功能：** [打开问题](https://github.com/rkz91/coco/issues/new/choose)。 - **安全：** 见 [`SECURITY.md`](SECURITY.md) 以进行责任披露。 - **行为准则：** 见 [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md)。 在打开 PR 之前，请在本地验证 — CI 运行了一个前端元数据检查、清单验证和安装脚本语法检查。 --- ## 构建于 CoCo 建立在优秀的开源框架和标准之上。特别感谢： * **[obra/superpowers](https://github.com/obra/superpowers)** （Jesse Vincent）用于基础的系统调试和验证技能。 * **[gsd-build/get-shit-done](https://github.com/gsd-build/get-shit-done)** 用于 68 技能的 GSD 项目编排框架。 * **[JCodesMore/ai-website-cloner-template](https://github.com/JCodesMore/ai-website-cloner-template)** 用于站点克隆结构。 * **[agents.md](https://agents.md/)** 社区用于供应商中立的代理上下文标准。 * **[nickwinder/synthteam](https://github.com/nickwinder/synthteam)** 用于 `--debate` 商议背后的多代理辩论协议。 完整的署名 — 包括功能级灵感和许可证 — 位于 [CREDITS.md](CREDITS.md)。如果某些内容影响了 CoCo 但未被列出，那就是一个错误：打开一个问题。