LLM & Agents

QA engineer and test automation specialist with deep expertise in Flutter testing. Use for designing test strategies, writing unit/widget/integration tests, improving test coverage, and ensuring code reliability.

Orchestrate comprehensive 7-phase RBMAS research for multi-dimensional queries (4+ dimensions). Coordinates multiple specialist agents through SCOPE → PLAN → RETRIEVE → TRIANGULATE → DRAFT → CRITIQUE → PACKAGE methodology. Use for thorough internet research requiring multiple sources, iterative refinement, and quality gates. Triggers - research, investigate, analyze with 4+ distinct dimensions or comprehensive depth requirements.

Generate pytest tests with parametrization, shared fixtures, minimal mocking. Use for unit tests and test coverage. Follows 1-1 file mapping and real object testing.

Analyze code changes for security vulnerabilities using LLM reasoning and threat model patterns. Use for PR reviews, pre-commit checks, or branch comparisons.

Plan what to measure in AI agent systems using tiered approach

Conduct comprehensive research on any topic by coordinating 2-4 specialized researcher agents in parallel, then synthesizing findings into a detailed report via mandatory report-writer agent delegation

Create worktrees and launch Claude Code agents. USE THIS SKILL when user says "create worktree", "spin up worktree", "new worktree", "worktree for X", or wants parallel development branches. Also handles worktree status, cleanup, and agent launching.

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends agent' capabilities with specialized knowledge, workflows, or tool integrations.

USE when preparing to backup agent consciousness, planning consciousness transplant to new system, or restoring from backup archive. Extracts policy guidance for strategic backup triggers, cross-OS migration, virgin system restore, and safety protocols.

Claude SDK development with Messages API, Tool Use, Extended Thinking, streaming, and prompt caching

Use when working with Quetrex's voice interface, OpenAI Realtime API, WebRTC, or echo cancellation. Knows Quetrex's specific voice architecture decisions and patterns. CRITICAL - prevents breaking working voice system.

Orchestrator agent that coordinates all BarqNet specialized agents (backend, integration, client, documentation, audit, testing) to execute complete end-to-end workflows. Plans multi-agent deployments, manages task dependencies, tracks progress across all platforms, and ensures comprehensive completion. Use for complex multi-component tasks, full-stack features, or production deployments.

Full implementation mode - end-to-end feature implementation with parallel agent orchestration

Pull request swarm management for multi-agent code review and validation. Use for coordinated PR reviews, automated validation, PR-based swarm creation, and intelligent merge workflows.

This skill provides complete coverage of Google Gemini embeddings API (gemini-embedding-001) for building RAG systems, semantic search, document clustering, and similarity matching. Use when implementing vector search with Google's embedding models, integrating with Cloudflare Vectorize, or building retrieval-augmented generation systems. Covers SDK usage (@google/genai), fetch-based Workers implementation, batch processing, 8 task types (RETRIEVAL_QUERY, RETRIEVAL_DOCUMENT, SEMANTIC_SIMILARITY, etc.), dimension optimization (128-3072), and cosine similarity calculations. Prevents 8+ embedding-specific errors including dimension mismatches, incorrect task types, rate limiting issues (100 RPM free tier), vector normalization mistakes, text truncation (2,048 token limit), and model version confusion. Includes production-ready RAG patterns with Cloudflare Vectorize integration, chunking strategies, and caching patterns. Token savings: ~60%. Production tested. Keywords: gemini embeddings, gemini-embedding-001, google embeddings, semantic search, RAG, vector search, document clustering, similarity search, retrieval augmented generation, vectorize integration, cloudflare vectorize embeddings, 768 dimensions, embed content gemini, batch embeddings, embeddings api, cosine similarity, vector normalization, retrieval query, retrieval document, task types, dimension mismatch, embeddings rate limit, text truncation, @google/genai

Analyzes codebases by impact scope (low/medium/high), generates refactoring checklists, supports interactive review, executes refactors with QA, and produces detailed optimization reports.

Create API agents that wrap external HTTP services (n8n, LangGraph, CrewAI, OpenAI endpoints). Configure request/response transforms, webhook status tracking, A2A protocol compliance. CRITICAL: Request transforms use template variables ({{userMessage}}, {{conversationId}}, etc.). Response transforms use field extraction. Status webhook URL must read from environment variables.

Generate and edit images using AI models (Gemini, OpenAI, BFL FLUX, or Poe). Use this skill when you need to create images from text prompts, edit existing images, or generate variations. Supports multiple providers with automatic API key detection from environment variables.

多 Agent 協作討論模式。當用戶說「party mode」、「bmad party mode」、「開始/啟動 party mode」、「召喚團隊」、「召喚專家」、「團隊討論」、「專家會議」時啟動。適用於技術架構討論、產品策略規劃、專案腦力激盪等需要多角度專家觀點的場景。

Execute and manage multi-step workflows with progress tracking, error handling, and validation. Use when running complex development workflows, coordinating multi-phase tasks, or managing workflow state and recovery.