prompt-builder
Build complete agent prompts deterministically via Python script. Use BEFORE spawning any BAZINGA agent (Developer, QA, Tech Lead, PM, etc.).
ai-4d-framework
Apply Anthropic's 4D Framework for AI delegation: Delegation (task selection), Description (instructions), Discernment (verification), and Diligence (iteration).
interactive-feedback
Implements a proactive confirmation and continuous loop mechanism for all agent interactions. Use this skill when engaging in any task (consultation, development, debugging) to ensure step-by-step user approval and maintain continuous workflow until explicit termination.
claude-agent-ui-ts
Add a React + WebSocket UI on top of Claude Agent SDK agents with tool approval and SQLite persistence
vercel-ai-sdk
Patterns for building AI applications with Vercel AI SDK including streaming LLM responses, tool calling, structured outputs, and multi-provider support. Use when integrating LLMs into applications.
ac-master-controller
Master controller for complete autonomous operation. Use when starting full autonomous projects, managing end-to-end workflow, controlling autonomous lifecycle, or running complete implementations.
authoring-claude-md
Standards for writing CLAUDE.md and modular rules files (.claude/rules/) that maximize LLM execution efficiency. Use when creating or editing project instructions, defining conventions, or optimizing agent behavior for determinism and token efficiency. Focuses on eliminating ambiguity and producing consistent agent execution.
ensemble-content-scorer
Multi-model consensus scoring for content ideas. Scores the same idea with Claude, GPT-4o, Gemini, and Grok in parallel, then aggregates for a balanced verdict. Reduces single-model bias and improves viral predictions.
skill-list
List all available skills configured in AGENTS.md. Scan and display skills with their names, descriptions, and trigger commands. Triggers when user mentions "列出技能", "list skills", "可用技能", "show skills", "技能列表", or uses command /skill-list.
agent-creation
Guides creating Claude Code agents, subagents, and skills. Use when building new agents, optimizing existing ones, or structuring skills.
cva-patterns-cost
Cost optimization strategies for production AI pipelines in Clojure+Vertex AI. Covers multi-model routing (70% Gemini/20% Haiku/10% Sonnet), token optimization (prompt engineering, output constraints), aggressive caching (58% cost reduction), batch processing, and real-time monitoring. Includes production metrics showing $0.391 to $0.162 per pipeline (-58%). Use when optimizing production costs, implementing multi-model strategies, designing budget controls, or scaling to high volume.
agentdb-performance-optimization
Optimize AgentDB performance with quantization (4-32x memory reduction), HNSW indexing (150x faster search), caching, and batch operations. Use when optimizing memory usage, improving search speed, or scaling to millions of vectors.
architecture-review
Review an existing Python RAG codebase against an explicit RAG architecture document and produce a production-readiness backlog with priorities, rationale, and execution guidance.
subagent-builder
Create and configure custom subagents in Claude Code. Use when user asks to create a new subagent, needs specialized AI assistance with specific tool restrictions or permissions, implements task-specific workflows with isolated context, sets up domain-specific agents with custom prompts, or creates subagents with hooks or validation rules
ros-launch-files
Use this skill when the user needs help writing, understanding, or modifying ROS launch files. Covers launch XML syntax, arguments, parameters, node configuration, includes, conditionals, namespaces, remapping, and common patterns. Trigger examples: - "How do I create a launch file?" - "Add an argument to this launch file" - "What does $(find pkg) do?" - "How do I include another launch file?" - "Set up conditional launching" - "Configure node parameters in launch"
cognitive-workflows
Multi-LLM workflow patterns for complex reasoning tasks. Based on Anthropic's "Building Effective Agents" research. Use this skill when: (1) Tasks require multiple perspectives or approaches (2) Complex decisions need structured analysis (3) Problems benefit from decomposition into subtasks (4) Quality requires iterative refinement (5) Input needs routing to specialized handling Patterns: chain, parallel, route, orchestrator-workers, evaluator-optimizer
rivet-harvest
Capture decisions and requirements from conversation history. Use when user says "harvest", "capture decisions", or "save what we discussed" to persist architectural decisions into .rivet/systems.yaml.
