LLM & Agents

Autonomous agent-based web search fallback for when WebSearch API fails or hits limits

Run section orchestrators to coordinate multi-component workflows. Use when starting work on a section.

Generate images via local SDXL Lightning, OpenAI DALL·E, Replicate, or custom providers with structured prompts and safety checks.

Implements tasks from .plans/ directories by following implementation guidance, writing code and tests, and updating task status. Use when task file is in implementation/ directory and requires code implementation with comprehensive testing. Launches research agents when stuck.

Vector search and semantic memory backbone for agents and RAG pipelines.

Testing TDD & BDD - Atoll Tourisme. Use when writing tests, reviewing test coverage, or setting up testing.

Measure test coverage and identify untested code. Use when assessing test completeness.

Monitor and report status of all tmux sessions including dev environments, spawned agents, and running processes. Uses tmuxwatch for enhanced visibility.

快速在新 Next.js 项目中集成 LangGraph 后端（localhost:2024），创建极简的 /chat 页面，支持真正的流式 AI 对话。适用于需要快速搭建聊天界面的场景：1) 集成 LangGraph Studio 本地部署的后端，2) 创建支持实时流式响应的聊天页面，3) 使用 LangGraph SDK 进行客户端直接连接，4) 实现优雅的 UI 交互（避免双气泡、流畅的加载动画）。

Select the best-fit specialist agent from the registry using semantic matching, capability scoring, and tool alignment.

Advanced AgentDB Vector Search Implementation operates on 3 fundamental principles:

Generate fully-specified specialist agents with frontmatter, structured prompts, and validation hooks ready for registry inclusion.

[CLAUDE CODE ONLY] Leverage Gemini CLI for AI peer review, second opinions on architecture and design decisions, cross-validation of implementations, security analysis, alternative approaches, and holistic codebase analysis. Requires terminal access to execute Gemini CLI commands. Use when making high-stakes decisions, reviewing complex architecture, analyzing large codebases (1M token context window), or when explicitly requested for a second AI perspective. Must be explicitly invoked using skill syntax.

Generate concise CLAUDE.md files and agent documentation following best practices. Use when setting up a new project for Claude Code, auditing existing CLAUDE.md, or creating progressive disclosure documentation structure.

Testing - Principes TDD/BDD. Use when writing tests, reviewing test coverage, or setting up testing.

Coordinate multi-agent software delivery with guardrails for planning, implementation, testing, and release.

Transform extracted engineer expertise into an actionable skill with progressive disclosure, allowing agents to find and apply relevant patterns for specific tasks.

Apply quantization to reduce memory by 4-32x. Enable HNSW indexing for 150x faster search. Configure caching strategies and implement batch operations. Use when optimizing memory usage, improving search speed, or scaling to millions of vectors. Deploy these optimizations to achieve 12,500x performance gains.

Regra 07: Testes. Use when writing tests, reviewing test coverage, or setting up testing.

Determine the best Anthropic architecture for your project by analyzing requirements and recommending the optimal combination of Skills, Agents, Prompts, and SDK primitives.