NeoLabHQ/context-engineering-kit

这个仓库里有 204 个技能，GitHub 星标 ★ 1,702。

• actualizeReconcile the project's FPF state with recent repository changes
• add-taskcreates draft task file in .specs/tasks/draft/ with original user intent
• agent-evaluationEvaluate and improve Claude Code commands, skills, and agents. Use when testing
• analyseAuto-selects best Kaizen method (Gemba Walk, Value Stream, or Muda) for target
• analyse-problemComprehensive A3 one-page problem analysis with root cause and action plan
• analyze-issueAnalyze a GitHub issue and create a detailed technical specification
• apply-anthropic-skill-best-practicesComprehensive guide for skill development based on Anthropic's official best pra
• attach-review-to-prAdd line-specific review comments to pull requests using GitHub CLI API
• brainstormUse when creating or developing, before writing code or implementation plans - r
• build-mcpGuide for creating high-quality MCP (Model Context Protocol) servers that enable
• cause-and-effectSystematic Fishbone analysis exploring problem causes across six categories
• commitCreate well-formatted commits with conventional commit messages and emoji
• context-engineeringUnderstand the components, mechanics, and constraints of context in agent system
• create-agentComprehensive guide for creating Claude Code agents with proper structure, trigg
• create-commandInteractive assistant for creating new Claude commands with proper structure, pa
• create-hookCreate and configure git hooks with intelligent project analysis, suggestions, a
• create-ideasGenerate ideas in one shot using creative sampling
• create-prCreate pull requests using GitHub CLI with proper templates and formatting
• create-ruleUse when found gap or repetative issue, that produced by you or implemenataion a
• create-skillGuide for creating effective skills. This command should be used when users want
• create-workflow-commandCreate a workflow command that orchestrates multi-step execution through sub-age
• critiqueComprehensive multi-perspective review using specialized judges with debate and
• decayManage evidence freshness by identifying stale decisions and providing governanc
• design-testing-strategyUse before writing any type of tests. Distills 14 industry sources into determin
• do-and-judgeExecute a task with sub-agent implementation and LLM-as-a-judge verification wit
• do-competitivelyExecute tasks through competitive multi-agent generation, meta-judge evaluation
• do-in-parallelRun independent tasks concurrently across multiple files or targets using parall
• do-in-stepsExecute one complex task as ordered, dependent steps run sequentially, passing c
• fix-testsSystematically fix all failing tests after business logic changes or refactoring
• git-notesUse when adding metadata to commits without changing history, tracking review st
• git-worktreesUse when working on multiple branches simultaneously, context switching without
• implement-taskImplement a task step by step with automated LLM-as-Judge verification at the en
• judgeLaunch a meta-judge then a judge sub-agent to evaluate results produced in the c
• judge-with-debateEvaluate solutions through multi-round debate between independent judges until c
• kaizenUse when Code implementation and refactoring, architecturing or designing system
• launch-sub-agentLaunch an intelligent sub-agent with automatic model selection based on task com
• load-issuesLoad all open issues from GitHub and save them as markdown files
• load-pr-commentsUse to load open/unresolved PR review comments then aggregate them as tasks in .
• memorizeCurates insights from reflections and critiques into CLAUDE.md using Agentic Con
• multi-agent-patternsDesign multi-agent architectures for complex tasks. Use when single-agent contex
• plan-do-check-actIterative PDCA cycle for systematic experimentation and continuous improvement
• plan-taskRefine a draft task specification into a fully planned, implementation-ready tas
• prompt-engineeringUse this skill when you writing commands, hooks, skills for Agent, or prompts fo
• propose-hypothesesExecute complete FPF cycle from hypothesis generation to decision
• querySearch the FPF knowledge base and display hypothesis details with assurance info
• reflectReflect on previus response and output, based on Self-refinement framework for i
• resetReset the FPF reasoning cycle to start fresh
• resolve-fixed-pr-commentsVerify what PR review comments have been addressed (committed/pushed OR uncommit
• review-local-changesReview your local uncommitted working-tree changes (git diff plus untracked file
• review-prReview an existing GitHub pull request and post inline review comments on its di
• root-cause-tracingUse when errors occur deep in execution and you need to trace back to find the o
• setup-arxiv-mcpGuide for setup arXiv paper search MCP server using Docker MCP
• setup-codemap-cliGuide for setup Codemap CLI for intelligent codebase visualization and navigatio
• setup-context7-mcpGuide for setup Context7 MCP server to load documentation for specific technolog
• setup-serena-mcpGuide for setup Serena MCP server for semantic code retrieval and editing capabi
• statusDisplay the current state of the FPF knowledge base
• subagent-driven-developmentUse when executing implementation plans with independent tasks in the current se
• test-coverageUse after writing tests to assess coverage quality across structural, mutation,
• test-driven-developmentUse when implementing any feature or bugfix, before writing implementation code
• test-promptUse when creating or editing any prompt (commands, hooks, skills, subagent instr
• test-skillUse when creating or editing skills, before deployment, to verify they work unde
• thought-based-reasoningUse when tackling complex reasoning tasks requiring step-by-step logic, multi-st
• traiage-reviewThis skill should be used when need prioritize what changed code in repository h
• tree-of-thoughtsExecute tasks through systematic exploration, pruning, and expansion using Tree
• update-docsUpdate and maintain project documentation for local code changes using multi-age
• whyIterative Five Whys root cause analysis drilling from symptoms to fundamentals
• write-conciselyApply writing rules to any documentation that humans will read. Makes your writi
• write-testsAdd missing test coverage for your local code changes by generating new test fil
• agent-evaluationEvaluate and improve Claude Code commands, skills, and agents. Use when testing
• apply-anthropic-skill-best-practicesComprehensive guide for skill development based on Anthropic's official best pra
• context-engineeringUnderstand the components, mechanics, and constraints of context in agent system
• create-agentComprehensive guide for creating Claude Code agents with proper structure, trigg
• create-commandInteractive assistant for creating new Claude commands with proper structure, pa
• create-hookCreate and configure git hooks with intelligent project analysis, suggestions, a
• create-ruleUse when found gap or repetative issue, that produced by you or implemenataion a
• create-skillGuide for creating effective skills. This command should be used when users want
• create-workflow-commandCreate a workflow command that orchestrates multi-step execution through sub-age
• prompt-engineeringUse this skill when you writing commands, hooks, skills for Agent, or prompts fo
• test-promptUse when creating or editing any prompt (commands, hooks, skills, subagent instr
• test-skillUse when creating or editing skills, before deployment, to verify they work unde
• thought-based-reasoningUse when tackling complex reasoning tasks requiring step-by-step logic, multi-st
• update-docsUpdate and maintain project documentation for local code changes using multi-age
• write-conciselyApply writing rules to any documentation that humans will read. Makes your writi
• actualizeReconcile the project's FPF state with recent repository changes
• decayManage evidence freshness by identifying stale decisions and providing governanc
• propose-hypothesesExecute complete FPF cycle from hypothesis generation to decision
• querySearch the FPF knowledge base and display hypothesis details with assurance info
• resetReset the FPF reasoning cycle to start fresh
• statusDisplay the current state of the FPF knowledge base
• analyze-issueAnalyze a GitHub issue and create a detailed technical specification
• attach-review-to-prAdd line-specific review comments to pull requests using GitHub CLI API
• commitCreate well-formatted commits with conventional commit messages and emoji
• create-prCreate pull requests using GitHub CLI with proper templates and formatting
• git-notesUse when adding metadata to commits without changing history, tracking review st
• git-worktreesUse when working on multiple branches simultaneously, context switching without
• load-issuesLoad all open issues from GitHub and save them as markdown files
• load-pr-commentsUse to load open/unresolved PR review comments then aggregate them as tasks in .
• resolve-fixed-pr-commentsVerify what PR review comments have been addressed (committed/pushed OR uncommit
• analyseAuto-selects best Kaizen method (Gemba Walk, Value Stream, or Muda) for target
• analyse-problemComprehensive A3 one-page problem analysis with root cause and action plan
• cause-and-effectSystematic Fishbone analysis exploring problem causes across six categories
• kaizenUse when Code implementation and refactoring, architecturing or designing system
• plan-do-check-actIterative PDCA cycle for systematic experimentation and continuous improvement
• root-cause-tracingUse when errors occur deep in execution and you need to trace back to find the o
• whyIterative Five Whys root cause analysis drilling from symptoms to fundamentals
• build-mcpGuide for creating high-quality MCP (Model Context Protocol) servers that enable
• setup-arxiv-mcpGuide for setup arXiv paper search MCP server using Docker MCP
• setup-codemap-cliGuide for setup Codemap CLI for intelligent codebase visualization and navigatio
• setup-context7-mcpGuide for setup Context7 MCP server to load documentation for specific technolog
• setup-serena-mcpGuide for setup Serena MCP server for semantic code retrieval and editing capabi
• critiqueComprehensive multi-perspective review using specialized judges with debate and
• memorizeCurates insights from reflections and critiques into CLAUDE.md using Agentic Con
• reflectReflect on previus response and output, based on Self-refinement framework for i
• review-local-changesReview your local uncommitted working-tree changes (git diff plus untracked file
• review-prReview an existing GitHub pull request and post inline review comments on its di
• traiage-reviewThis skill should be used when need prioritize what changed code in repository h
• do-and-judgeExecute a task with sub-agent implementation and LLM-as-a-judge verification wit
• do-competitivelyExecute tasks through competitive multi-agent generation, meta-judge evaluation
• do-in-parallelRun independent tasks concurrently across multiple files or targets using parall
• do-in-stepsExecute one complex task as ordered, dependent steps run sequentially, passing c
• judgeLaunch a meta-judge then a judge sub-agent to evaluate results produced in the c
• judge-with-debateEvaluate solutions through multi-round debate between independent judges until c
• launch-sub-agentLaunch an intelligent sub-agent with automatic model selection based on task com
• multi-agent-patternsDesign multi-agent architectures for complex tasks. Use when single-agent contex
• subagent-driven-developmentUse when executing implementation plans with independent tasks in the current se
• tree-of-thoughtsExecute tasks through systematic exploration, pruning, and expansion using Tree
• add-taskcreates draft task file in .specs/tasks/draft/ with original user intent
• brainstormUse when creating or developing, before writing code or implementation plans - r
• create-ideasGenerate ideas in one shot using creative sampling
• implement-taskImplement a task step by step with automated LLM-as-Judge verification at the en
• plan-taskRefine a draft task specification into a fully planned, implementation-ready tas
• design-testing-strategyUse before writing any type of tests. Distills 14 industry sources into determin
• fix-testsSystematically fix all failing tests after business logic changes or refactoring
• test-coverageUse after writing tests to assess coverage quality across structural, mutation,
• test-driven-developmentUse when implementing any feature or bugfix, before writing implementation code
• write-testsAdd missing test coverage for your local code changes by generating new test fil
• actualizeReconcile the project's FPF state with recent repository changes
• add-taskcreates draft task file in .specs/tasks/draft/ with original user intent
• agent-evaluationEvaluate and improve Claude Code commands, skills, and agents. Use when testing
• analyseAuto-selects best Kaizen method (Gemba Walk, Value Stream, or Muda) for target
• analyse-problemComprehensive A3 one-page problem analysis with root cause and action plan
• analyze-issueAnalyze a GitHub issue and create a detailed technical specification
• apply-anthropic-skill-best-practicesComprehensive guide for skill development based on Anthropic's official best pra
• attach-review-to-prAdd line-specific review comments to pull requests using GitHub CLI API
• brainstormUse when creating or developing, before writing code or implementation plans - r
• build-mcpGuide for creating high-quality MCP (Model Context Protocol) servers that enable
• cause-and-effectSystematic Fishbone analysis exploring problem causes across six categories
• commitCreate well-formatted commits with conventional commit messages and emoji
• context-engineeringUnderstand the components, mechanics, and constraints of context in agent system
• create-agentComprehensive guide for creating Claude Code agents with proper structure, trigg
• create-commandInteractive assistant for creating new Claude commands with proper structure, pa
• create-hookCreate and configure git hooks with intelligent project analysis, suggestions, a
• create-ideasGenerate ideas in one shot using creative sampling
• create-prCreate pull requests using GitHub CLI with proper templates and formatting
• create-ruleUse when found gap or repetative issue, that produced by you or implemenataion a
• create-skillGuide for creating effective skills. This command should be used when users want
• create-workflow-commandCreate a workflow command that orchestrates multi-step execution through sub-age
• critiqueComprehensive multi-perspective review using specialized judges with debate and
• decayManage evidence freshness by identifying stale decisions and providing governanc
• design-testing-strategyUse before writing any type of tests. Distills 14 industry sources into determin
• do-and-judgeExecute a task with sub-agent implementation and LLM-as-a-judge verification wit
• do-competitivelyExecute tasks through competitive multi-agent generation, meta-judge evaluation
• do-in-parallelRun independent tasks concurrently across multiple files or targets using parall
• do-in-stepsExecute one complex task as ordered, dependent steps run sequentially, passing c
• fix-testsSystematically fix all failing tests after business logic changes or refactoring
• git-notesUse when adding metadata to commits without changing history, tracking review st
• git-worktreesUse when working on multiple branches simultaneously, context switching without
• implement-taskImplement a task step by step with automated LLM-as-Judge verification at the en
• judgeLaunch a meta-judge then a judge sub-agent to evaluate results produced in the c
• judge-with-debateEvaluate solutions through multi-round debate between independent judges until c
• kaizenUse when Code implementation and refactoring, architecturing or designing system
• launch-sub-agentLaunch an intelligent sub-agent with automatic model selection based on task com
• load-issuesLoad all open issues from GitHub and save them as markdown files
• load-pr-commentsUse to load open/unresolved PR review comments then aggregate them as tasks in .
• memorizeCurates insights from reflections and critiques into CLAUDE.md using Agentic Con
• multi-agent-patternsDesign multi-agent architectures for complex tasks. Use when single-agent contex
• plan-do-check-actIterative PDCA cycle for systematic experimentation and continuous improvement
• plan-taskRefine a draft task specification into a fully planned, implementation-ready tas
• prompt-engineeringUse this skill when you writing commands, hooks, skills for Agent, or prompts fo
• propose-hypothesesExecute complete FPF cycle from hypothesis generation to decision
• querySearch the FPF knowledge base and display hypothesis details with assurance info
• reflectReflect on previus response and output, based on Self-refinement framework for i
• resetReset the FPF reasoning cycle to start fresh
• resolve-fixed-pr-commentsVerify what PR review comments have been addressed (committed/pushed OR uncommit
• review-local-changesReview your local uncommitted working-tree changes (git diff plus untracked file
• review-prReview an existing GitHub pull request and post inline review comments on its di
• root-cause-tracingUse when errors occur deep in execution and you need to trace back to find the o
• setup-arxiv-mcpGuide for setup arXiv paper search MCP server using Docker MCP
• setup-codemap-cliGuide for setup Codemap CLI for intelligent codebase visualization and navigatio
• setup-context7-mcpGuide for setup Context7 MCP server to load documentation for specific technolog
• setup-serena-mcpGuide for setup Serena MCP server for semantic code retrieval and editing capabi
• statusDisplay the current state of the FPF knowledge base
• subagent-driven-developmentUse when executing implementation plans with independent tasks in the current se
• test-coverageUse after writing tests to assess coverage quality across structural, mutation,
• test-driven-developmentUse when implementing any feature or bugfix, before writing implementation code
• test-promptUse when creating or editing any prompt (commands, hooks, skills, subagent instr
• test-skillUse when creating or editing skills, before deployment, to verify they work unde
• thought-based-reasoningUse when tackling complex reasoning tasks requiring step-by-step logic, multi-st
• traiage-reviewThis skill should be used when need prioritize what changed code in repository h
• tree-of-thoughtsExecute tasks through systematic exploration, pruning, and expansion using Tree
• update-docsUpdate and maintain project documentation for local code changes using multi-age
• whyIterative Five Whys root cause analysis drilling from symptoms to fundamentals
• write-conciselyApply writing rules to any documentation that humans will read. Makes your writi
• write-testsAdd missing test coverage for your local code changes by generating new test fil