affaan-m/ECC
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
JavaScript · ★ 305,751 · Fork 44,999 · 在 GitHub 查看
在手机上浏览源码、订阅每日 AI 趋势日报与项目架构深度解读:获取 CopoHub 客户端(iOS & HarmonyOS)→
GitHub Trending 上榜记录
affaan-m/ECC 已在 GitHub Trending 每日榜单上榜 57 天,最高排名第 1,首次上榜 2026-05-25,最近一次上榜 2026-11-18。
| 日期 | 排名 | Star | 当日增长 |
|---|---|---|---|
| 2026-11-18 | #13 | 305,751 | +883 |
| 2026-11-17 | #13 | 305,452 | +902 |
| 2026-11-16 | #46 | 304,591 | +722 |
| 2026-11-14 | #26 | 302,662 | +559 |
| 2026-11-13 | #26 | 302,439 | +608 |
| 2026-10-28 | #8 | 291,255 | +1,165 |
| 2026-10-27 | #5 | 290,839 | +1,183 |
| 2026-10-26 | #5 | 289,785 | +987 |
| 2026-10-18 | #3 | 282,918 | +1,145 |
| 2026-10-17 | #3 | 282,443 | +1,145 |
| 2026-10-16 | #3 | 281,442 | +1,057 |
| 2026-10-15 | #10 | 280,309 | +1,015 |
| 2026-10-04 | #3 | 272,521 | +897 |
| 2026-10-03 | #3 | 272,174 | +954 |
| 2026-09-21 | #1 | 264,243 | +826 |
| 2026-09-20 | #1 | 263,652 | +1,012 |
| 2026-09-19 | #4 | 262,416 | +958 |
| 2026-09-18 | #4 | 261,980 | +1,171 |
| 2026-09-17 | #13 | 261,100 | +1,173 |
| 2026-09-16 | #21 | 260,174 | +1,046 |
| 2026-09-10 | #13 | 255,484 | +1,133 |
| 2026-09-09 | #4 | 255,112 | +1,427 |
| 2026-09-08 | #4 | 254,206 | +1,897 |
| 2026-09-07 | #1 | 252,733 | +1,905 |
| 2026-09-06 | #1 | 251,176 | +1,486 |
| 2026-09-05 | #2 | 249,777 | +1,325 |
| 2026-09-04 | #4 | 248,408 | +1,139 |
| 2026-09-03 | #6 | 247,086 | +749 |
| 2026-09-02 | #11 | 246,251 | +623 |
| 2026-09-01 | #10 | 245,720 | +621 |
| 2026-08-31 | #10 | 245,188 | +548 |
| 2026-08-24 | #10 | 242,738 | +427 |
| 2026-08-23 | #3 | 242,521 | +427 |
| 2026-08-22 | #3 | 242,142 | +428 |
| 2026-08-21 | #11 | 241,753 | +348 |
| 2026-07-31 | #14 | 236,457 | +804 |
| 2026-07-30 | #3 | 236,172 | +857 |
| 2026-07-29 | #3 | 235,507 | +860 |
| 2026-07-28 | #5 | 234,735 | +692 |
| 2026-07-26 | #18 | 233,458 | +377 |
| 2026-07-25 | #18 | 233,262 | +364 |
| 2026-07-03 | #10 | 225,444 | +486 |
| 2026-07-02 | #10 | 225,132 | +721 |
| 2026-06-24 | #16 | 220,852 | +593 |
| 2026-06-23 | #16 | 220,456 | +582 |
| 2026-06-06 | #5 | 208,735 | +1,361 |
| 2026-06-05 | #3 | 208,279 | +1,750 |
| 2026-06-04 | #2 | 207,127 | +2,141 |
| 2026-06-03 | #3 | 205,063 | +2,147 |
| 2026-06-02 | #3 | 203,756 | +1,842 |
| 2026-05-31 | #7 | 199,676 | +908 |
| 2026-05-30 | #7 | 199,228 | +1,406 |
| 2026-05-29 | #2 | 198,519 | +1,413 |
| 2026-05-28 | #2 | 197,164 | +2,062 |
| 2026-05-27 | #2 | 195,937 | +2,062 |
| 2026-05-26 | #2 | 194,205 | +2,025 |
| 2026-05-25 | #4 | 192,226 | +2,052 |
同期上榜的项目
- elementalsouls/Claude-BugHunter – A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
- mvanhorn/last30days-skill – AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
- alibaba/open-code-review – Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
- Tencent/LoopForge – Self-hosted workflows for coding-agent delivery. Orchestrate CodeBuddy, Codex and Claude Code with scripts and human approval gates. Includes Local workflows for coding assistants.
- dream-num/univer – Build your own AI office with Univer Office SDK. Bring editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.
- getarcaneapp/arcane – Modern Docker Management, Designed for Everyone
ECC 项目深度剖析与商业评估报告
ECC 项目深度剖析与商业评估报告
1. 核心定位与商业落地价值
1.1 根本痛点解决 ECC 解决的是 AI Agent 开发工具链碎片化与效率黑洞 问题。当前 AI 辅助编程领域,Claude Code、Codex、Cursor 等工具各自为政,开发者面临: 配置碎片化:每个 harness 需要不同的规则、提示词、工具配置 上下文遗忘:Agent 会话之间无法持久化学习到的模式 安全盲区:Agent 自动执行的代码修改缺乏安全审计 性能不可控:Token 消耗、模型选择、并发控制缺乏系统化管理
ECC 通过提供统一的“harness-native 操作员系统”,将上述问题抽象为可复用的组件层。
1.2 企业级商业落地场景
| 场景 | 价值描述 | 目标客户 | |------|----------|----------| | 企业级 AI 开发标准化 | 统一多团队、多项目的 Agent 行为规范,确保代码质量一致性 | 中大型软件企业(50+ 开发者) | | 安全合规审计 | AgentShield 提供 102 条静态分析规则,可集成到 CI/CD 管道作为安全门禁 | 金融、医疗等受监管行业 | | MLOps 流水线优化 | MLE 工作流 + 成本感知 LLM 管道,可显著降低模型调用成本 | AI/ML 团队 | | 多语言技术栈管理 | 12 种语言规则体系,支持 Java、Kotlin、Swift 等企业主流语言 | 多语言技术栈企业 | | 远程/分布式团队协作 | 持续学习系统可跨会话共享“本能”(instincts),团队知识不流失 | 远程优先的工程团队 |
1.3 竞品差异化与护城河
vs. GitHub Copilot:ECC 不是代码补全工具,而是完整的 Agent 编排系统,具有记忆、安全、学习三大核心能力 vs. Cursor 内置规则:ECC 跨 7 个 harness 统一行为,而 Cursor 规则仅限自身生态 vs. 自定义 Claude Code 配置:ECC 提供了 60 个 Agent、232 个 Skill 的现成组件库,免去从零构建的工程成本 护城河:182K+ Stars、28K+ Forks 的网络效应 + 10 个月持续迭代的组件库深度 + Hermes 操作员系统的架构创新
在 README 中展示上榜徽章
[](https://github.fq6825.top/repo/affaan-m/ECC)