[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-1590":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":13,"stars7d":14,"stars30d":15,"stars90d":13,"forks30d":13,"starsTrendScore":13,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":17,"fork":17,"defaultBranch":18,"hasWiki":19,"hasPages":17,"topics":20,"createdAt":9,"pushedAt":9,"updatedAt":21,"readmeContent":22,"aiSummary":23,"trendingCount":13,"starSnapshotCount":13,"syncStatus":14,"lastSyncTime":24,"discoverSource":25},1590,"claudecode-harness","anothervibecoder-s\u002Fclaudecode-harness","anothervibecoder-s","Production-grade orchestration harness for Claude Code — Context discipline, subagents, and multi-model consensus.",null,212,9,1,0,2,11,3,false,"main",true,[],"2026-06-12 02:00:30","# Claude Code Engineering Harness 🦾\n\nStop chatting with AI. Start engineering with it.\n\nThis repository provides the blueprint for the **\"Project Harness\"** I use to run Claude Code (Opus 4.7) for 8+ hours a day on high-stakes SaaS projects without hitting quotas or losing model accuracy.\n\n🚀 Key Features\n- **Context Discipline:** Drastically cuts token usage by keeping the main session lean.\n- **Hub & Spoke Model:** Uses Sonnet\u002FHaiku subagents for \"dirty work\" (grep, research, unit tests).\n- **Multi-Model Consensus:** Automated triggers for parallel reviews by Gemini Pro + Flash + Sonnet.\n- **Persistent Memory:** A structured hierarchy for session continuity and decision tracking.\n- **Strict Guardrails:** Hard limits on lines-per-fix and mandatory verification steps.\n\n## 🛠️ How to use\n1. Copy the `CLAUDE_EXAMPLE.md` file to your project root.\n2. Replace all `[Placeholders]` with your specific tech stack, ownership matrix, and verification commands.\n3. Configure your `.claude\u002Fsettings.json` to include the `Stop` hooks for auto-retros.\n4. (Optional) Install `claude-flow` and `graphify` to augment the harness.\n   \n## 📈 Results\n- **Zero Quota Hits:** Even on 5x Max plan with heavy tool use.\n- **Zero Hallucinated Success:** Mandatory verification blocks \"fake\" fixes.\n- **Senior-Level Autonomy:** The model stops asking repetitive questions and follows your architectural standards.\n  \n*Inspired by the viral discussion on r\u002FClaudeAI, r\u002Fclaude, and r\u002Fvibecoding.*\n","Claude Code Engineering Harness 是一个面向生产环境的编排工具，旨在通过Claude Code（Opus 4.7）实现高效软件开发。该项目的核心功能包括上下文管理以减少token使用、子代理模型执行辅助任务、多模型共识机制确保代码质量、持久化记忆支持会话连续性和决策追踪以及严格的防护措施保证代码修改的安全性。适合需要长时间依赖AI进行高风险SaaS项目开发或希望提高团队与AI协作效率的场景下使用。","2026-06-11 02:44:51","CREATED_QUERY"]