[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AGD0VHSYPDNG4NRT4EKY2G":3,"skill-same:01M7AGD0VHSYPDNG4NRT4EKY2G":104},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":30,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7AGD0VHSYPDNG4NRT4EKY2G","ck:context-engineering",null,"Check context usage limits, monitor time remaining, optimize token consumption, debug context failures. Use when asking about context percentage, rate limits, usage warnings, context optimization, agent architectures, memory systems.","manhvann\u002Fcodexkit","manhvann","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F108862213?v=4","01M7AGBGXSB8Z8HXXR27QKX035","Manh Nguyen Van","01M7AGBGYDJPBR2EYEWJ0G1K8J",88,"MIT","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ck:context-engineering\ndescription: >-\n  Check context usage limits, monitor time remaining, optimize token consumption, debug context failures.\n  Use when asking about context percentage, rate limits, usage warnings, context optimization, agent architectures, memory systems.\nversion: 1.0.0\nargument-hint: \"[topic or question]\"\n---\n\n# Context Engineering\n\nContext engineering curates the smallest high-signal token set for LLM tasks. The goal: maximize reasoning quality while minimizing token usage.\n\n## When to Activate\n\n- Designing\u002Fdebugging agent systems\n- Context limits constrain performance\n- Optimizing cost\u002Flatency\n- Building multi-agent coordination\n- Implementing memory systems\n- Evaluating agent performance\n- Developing LLM-powered pipelines\n\n## Core Principles\n\n1. **Context quality > quantity** - High-signal tokens beat exhaustive content\n2. **Attention is finite** - U-shaped curve favors beginning\u002Fend positions\n3. **Progressive disclosure** - Load information just-in-time\n4. **Isolation prevents degradation** - Partition work across sub-agents\n5. **Measure before optimizing** - Know your baseline\n\n**IMPORTANT:**\n- Sacrifice grammar for the sake of concision.\n- Ensure token efficiency while maintaining high quality.\n- Pass these rules to subagents.\n\n## Quick Reference\n\n| Topic | When to Use | Reference |\n|-------|-------------|-----------|\n| **Fundamentals** | Understanding context anatomy, attention mechanics | [context-fundamentals.md](.\u002Freferences\u002Fcontext-fundamentals.md) |\n| **Degradation** | Debugging failures, lost-in-middle, poisoning | [context-degradation.md](.\u002Freferences\u002Fcontext-degradation.md) |\n| **Optimization** | Compaction, masking, caching, partitioning | [context-optimization.md](.\u002Freferences\u002Fcontext-optimization.md) |\n| **Compression** | Long sessions, summarization strategies | [context-compression.md](.\u002Freferences\u002Fcontext-compression.md) |\n| **Memory** | Cross-session persistence, knowledge graphs | [memory-systems.md](.\u002Freferences\u002Fmemory-systems.md) |\n| **Multi-Agent** | Coordination patterns, context isolation | [multi-agent-patterns.md](.\u002Freferences\u002Fmulti-agent-patterns.md) |\n| **Evaluation** | Testing agents, LLM-as-Judge, metrics | [evaluation.md](.\u002Freferences\u002Fevaluation.md) |\n| **Tool Design** | Tool consolidation, description engineering | [tool-design.md](.\u002Freferences\u002Ftool-design.md) |\n| **Pipelines** | Project development, batch processing | [project-development.md](.\u002Freferences\u002Fproject-development.md) |\n| **Runtime Awareness** | Usage limits, context window monitoring | [runtime-awareness.md](.\u002Freferences\u002Fruntime-awareness.md) |\n\n## Key Metrics\n\n- **Token utilization**: Warning at 70%, trigger optimization at 80%\n- **Token variance**: Explains 80% of agent performance variance\n- **Multi-agent cost**: ~15x single agent baseline\n- **Compaction target**: 50-70% reduction, \u003C5% quality loss\n- **Cache hit target**: 70%+ for stable workloads\n\n## Four-Bucket Strategy\n\n1. **Write**: Save context externally (scratchpads, files)\n2. **Select**: Pull only relevant context (retrieval, filtering)\n3. **Compress**: Reduce tokens while preserving info (summarization)\n4. **Isolate**: Split across sub-agents (partitioning)\n\n## Anti-Patterns\n\n- Exhaustive context over curated context\n- Critical info in middle positions\n- No compaction triggers before limits\n- Single agent for parallelizable tasks\n- Tools without clear descriptions\n\n## Guidelines\n\n1. Place critical info at beginning\u002Fend of context\n2. Implement compaction at 70-80% utilization\n3. Use sub-agents for context isolation, not role-play\n4. Design tools with 4-question framework (what, when, inputs, returns)\n5. Optimize for tokens-per-task, not tokens-per-request\n6. Validate with probe-based evaluation\n7. Monitor KV-cache hit rates in production\n8. Start minimal, add complexity only when proven necessary\n\n## Runtime Awareness\n\nThe system automatically injects usage awareness via PostToolUse hook:\n\n```xml\n\u003Cusage-awareness>\nCodex Usage Limits: 5h=45%, 7d=32%\nContext Window Usage: 67%\n\u003C\u002Fusage-awareness>\n```\n\n**Thresholds:**\n- 70%: WARNING - consider optimization\u002Fcompaction\n- 90%: CRITICAL - immediate action needed\n\n**Data Sources:**\n- Usage limits: OpenAI OAuth API (`https:\u002F\u002Fapi.openai.com\u002Fapi\u002Foauth\u002Fusage`)\n- Context window: Statusline temp file (`\u002Ftmp\u002Fck-context-{session_id}.json`)\n\n## Scripts\n\n- [context_analyzer.py](.\u002Fscripts\u002Fcontext_analyzer.py) - Context health analysis, degradation detection\n- [compression_evaluator.py](.\u002Fscripts\u002Fcompression_evaluator.py) - Compression quality evaluation\n","# Context Engineering\n\nContext engineering curates the smallest high-signal token set for LLM tasks. The goal: maximize reasoning quality while minimizing token usage.\n\n## When to Activate\n\n- Designing\u002Fdebugging agent systems\n- Context limits constrain performance\n- Optimizing cost\u002Flatency\n- Building multi-agent coordination\n- Implementing memory systems\n- Evaluating agent performance\n- Developing LLM-powered pipelines\n\n## Core Principles\n\n1. **Context quality > quantity** - High-signal tokens beat exhaustive content\n2. **Attention is finite** - U-shaped curve favors beginning\u002Fend positions\n3. **Progressive disclosure** - Load information just-in-time\n4. **Isolation prevents degradation** - Partition work across sub-agents\n5. **Measure before optimizing** - Know your baseline\n\n**IMPORTANT:**\n- Sacrifice grammar for the sake of concision.\n- Ensure token efficiency while maintaining high quality.\n- Pass these rules to subagents.\n\n## Quick Reference\n\n| Topic | When to Use | Reference |\n|-------|-------------|-----------|\n| **Fundamentals** | Understanding context anatomy, attention mechanics | [context-fundamentals.md](.\u002Freferences\u002Fcontext-fundamentals.md) |\n| **Degradation** | Debugging failures, lost-in-middle, poisoning | [context-degradation.md](.\u002Freferences\u002Fcontext-degradation.md) |\n| **Optimization** | Compaction, masking, caching, partitioning | [context-optimization.md](.\u002Freferences\u002Fcontext-optimization.md) |\n| **Compression** | Long sessions, summarization strategies | [context-compression.md](.\u002Freferences\u002Fcontext-compression.md) |\n| **Memory** | Cross-session persistence, knowledge graphs | [memory-systems.md](.\u002Freferences\u002Fmemory-systems.md) |\n| **Multi-Agent** | Coordination patterns, context isolation | [multi-agent-patterns.md](.\u002Freferences\u002Fmulti-agent-patterns.md) |\n| **Evaluation** | Testing agents, LLM-as-Judge, metrics | [evaluation.md](.\u002Freferences\u002Fevaluation.md) |\n| **Tool Design** | Tool consolidation, description engineering | [tool-design.md](.\u002Freferences\u002Ftool-design.md) |\n| **Pipelines** | Project development, batch processing | [project-development.md](.\u002Freferences\u002Fproject-development.md) |\n| **Runtime Awareness** | Usage limits, context window monitoring | [runtime-awareness.md](.\u002Freferences\u002Fruntime-awareness.md) |\n\n## Key Metrics\n\n- **Token utilization**: Warning at 70%, trigger optimization at 80%\n- **Token variance**: Explains 80% of agent performance variance\n- **Multi-agent cost**: ~15x single agent baseline\n- **Compaction target**: 50-70% reduction, \u003C5% quality loss\n- **Cache hit target**: 70%+ for stable workloads\n\n## Four-Bucket Strategy\n\n1. **Write**: Save context externally (scratchpads, files)\n2. **Select**: Pull only relevant context (retrieval, filtering)\n3. **Compress**: Reduce tokens while preserving info (summarization)\n4. **Isolate**: Split across sub-agents (partitioning)\n\n## Anti-Patterns\n\n- Exhaustive context over curated context\n- Critical info in middle positions\n- No compaction triggers before limits\n- Single agent for parallelizable tasks\n- Tools without clear descriptions\n\n## Guidelines\n\n1. Place critical info at beginning\u002Fend of context\n2. Implement compaction at 70-80% utilization\n3. Use sub-agents for context isolation, not role-play\n4. Design tools with 4-question framework (what, when, inputs, returns)\n5. Optimize for tokens-per-task, not tokens-per-request\n6. Validate with probe-based evaluation\n7. Monitor KV-cache hit rates in production\n8. Start minimal, add complexity only when proven necessary\n\n## Runtime Awareness\n\nThe system automatically injects usage awareness via PostToolUse hook:\n\n```xml\n\u003Cusage-awareness>\nCodex Usage Limits: 5h=45%, 7d=32%\nContext Window Usage: 67%\n\u003C\u002Fusage-awareness>\n```\n\n**Thresholds:**\n- 70%: WARNING - consider optimization\u002Fcompaction\n- 90%: CRITICAL - immediate action needed\n\n**Data Sources:**\n- Usage limits: OpenAI OAuth API (`https:\u002F\u002Fapi.openai.com\u002Fapi\u002Foauth\u002Fusage`)\n- Context window: Statusline temp file (`\u002Ftmp\u002Fck-context-{session_id}.json`)\n\n## Scripts\n\n- [context_analyzer.py](.\u002Fscripts\u002Fcontext_analyzer.py) - Context health analysis, degradation detection\n- [compression_evaluator.py](.\u002Fscripts\u002Fcompression_evaluator.py) - Compression quality evaluation\n",{"name":5,"version":24,"description":7,"argument-hint":25},"1.0.0","[topic or question]","main","f4f392e3e6ac09daca9d7c44c3afb20945435e36",".agents\u002Fskills\u002Fcontext-engineering",79,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit","https:\u002F\u002Fgithub.com\u002Fmanhvann","Backend Engineer passionate about building reliable systems and contributing to open source.","2026-08-29 12:03:10",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add manhvann\u002Fcodexkit --skill \"ck:context-engineering\"",{"tool":40,"command":41},"bunx","bunx skills add manhvann\u002Fcodexkit --skill \"ck:context-engineering\"",{"tool":43,"command":44},"pnpm","pnpm dlx skills add manhvann\u002Fcodexkit --skill \"ck:context-engineering\"","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002FSKILL.md",true,[48,54,58,62,66,70,74,78,82,86,90,96,100],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"references\u002Fcontext-compression.md","REFERENCE","参考文档",2335,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fcontext-compression.md",{"path":55,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":56,"sourceHtmlUrl":57},"references\u002Fcontext-degradation.md",3092,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fcontext-degradation.md",{"path":59,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Fcontext-fundamentals.md",2732,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fcontext-fundamentals.md",{"path":63,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":64,"sourceHtmlUrl":65},"references\u002Fcontext-optimization.md",2333,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fcontext-optimization.md",{"path":67,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":68,"sourceHtmlUrl":69},"references\u002Fevaluation.md",2253,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fevaluation.md",{"path":71,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":72,"sourceHtmlUrl":73},"references\u002Fmemory-systems.md",2570,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fmemory-systems.md",{"path":75,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":76,"sourceHtmlUrl":77},"references\u002Fmulti-agent-patterns.md",2319,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fmulti-agent-patterns.md",{"path":79,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":80,"sourceHtmlUrl":81},"references\u002Fproject-development.md",2142,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fproject-development.md",{"path":83,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":84,"sourceHtmlUrl":85},"references\u002Fruntime-awareness.md",4895,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Fruntime-awareness.md",{"path":87,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":88,"sourceHtmlUrl":89},"references\u002Ftool-design.md",2057,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Freferences\u002Ftool-design.md",{"path":91,"resourceType":92,"resourceTypeDesc":93,"sizeBytes":94,"sourceHtmlUrl":95},"scripts\u002Fcompression_evaluator.py","SCRIPT","脚本",11753,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Fscripts\u002Fcompression_evaluator.py",{"path":97,"resourceType":92,"resourceTypeDesc":93,"sizeBytes":98,"sourceHtmlUrl":99},"scripts\u002Fcontext_analyzer.py",11131,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Fscripts\u002Fcontext_analyzer.py",{"path":101,"resourceType":92,"resourceTypeDesc":93,"sizeBytes":102,"sourceHtmlUrl":103},"scripts\u002Ftests\u002Ftest_edge_cases.py",9651,"https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcontext-engineering\u002Fscripts\u002Ftests\u002Ftest_edge_cases.py",[105,112,120,127,134,141],{"skillNo":106,"name":107,"nameZh":6,"description":108,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":109,"categories":110,"keywords":111,"publishedAt":19},"01M7AGNE44ZPV7VVMCZCDQ0X55","ck:worktree","Create isolated git worktree for parallel development in monorepos.","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fworktree\u002FSKILL.md",[],[],{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":116,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":119,"publishedAt":19},"01M7AGNB7EGT5ZKNRAF5M1V6D0","ck:web-testing","Web testing with Playwright, Vitest, k6. E2E\u002Funit\u002Fintegration\u002Fload\u002Fsecurity\u002Fvisual\u002Fa11y testing. Use for test automation, flakiness, Core Web Vitals, mobile gestures, cross-browser.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fweb-testing\u002FSKILL.md",[],[],{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":126,"publishedAt":19},"01M7AGN7W938QCYZ70A8B0X70G","ck:web-frameworks","Build with Next.js (App Router, RSC, SSR, ISR), Turborepo monorepos. Use for React apps, server rendering, build optimization, caching strategies, shared dependencies.","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fweb-frameworks\u002FSKILL.md",[],[],{"skillNo":128,"name":129,"nameZh":6,"description":130,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":131,"categories":132,"keywords":133,"publishedAt":19},"01M7AGN5HKJM49Z49N6XFDQB82","ck:web-design-guidelines","Review UI code for Web Interface Guidelines compliance. Use when asked to \"review my UI\", \"check accessibility\", \"audit design\", \"review UX\", or \"check my site against best practices\".","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fweb-design-guidelines\u002FSKILL.md",[],[],{"skillNo":135,"name":136,"nameZh":6,"description":137,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":138,"categories":139,"keywords":140,"publishedAt":19},"01M7AGN4D60DGGYM2YGZ80NT2A","ck:watzup","Review recent changes and wrap up the current work session.","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fwatzup\u002FSKILL.md",[],[],{"skillNo":142,"name":143,"nameZh":6,"description":144,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":145,"categories":146,"keywords":147,"publishedAt":19},"01M7AGN308V7338SEEX8A6DZ53","ck:use-mcp","Utilize MCP server tools with intelligent discovery and execution.","https:\u002F\u002Fgithub.com\u002Fmanhvann\u002Fcodexkit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fuse-mcp\u002FSKILL.md",[],[]]