LLM & Agents

Master RL theory - MDPs, value functions, Bellman equations, value/policy iteration, TD

Use when executing implementation plans with independent tasks in the current session - dispatches fresh subagent for each task with code review between tasks, enabling fast iteration with quality gates

Virtual test environment for Claude Code hooks, plugins, and agents. Use when testing or validating plugin components without affecting production.

Analyze Claude Code plugins for quality, DX, and architecture. Use when "review plugin quality", "audit plugins", "analyze the marketplace", "check plugins against Anthropic standards", "improve plugin design", or evaluating plugin architecture. Provides systematic analysis methodology with validation framework.

Use when working with unit tests in any capacity - reading, writing, updating, planning, evaluating, or reviewing tests - applies TDD principles with focused test coverage, filtering out redundant tests while ensuring critical paths, edge cases, and state changes are verified

Write Node.js CLI tools with zero dependencies. Use when creating command-line tools, argument parsing, colored output, or interactive prompts.

Use when creating or modifying Claude Code skills - SKILL.md development with frontmatter requirements, progressive disclosure, and model-invoked activation patterns for November 2025 specifications

Use this skill when creating comprehensive testing strategies for applications. Provides test planning templates, coverage targets, test case structures, and guidance for unit, integration, E2E, and performance testing. Ensures robust quality assurance across the development lifecycle.

Adversarial quality gate agent for code review - finds flaws before users do

Use this skill to discover all available SynthesisFlow skills and their capabilities. Provides a bootstrap context for AI agents by listing all skills, their descriptions, and script paths from the .claude/skills/ directory.

AI 驱动的代码审查能力，当用户提到代码审查、code review、审查代码、查看代码质量等关键词时自动触发。使用 9 个并行 Agent 进行全维度分析（Bug/质量/安全/性能）。

Mandatory design system guidelines for shadcn/ui with Tailwind v4. Enforces 4 font sizes, 2 weights, 8pt grid spacing, 60/30/10 color rule, OKLCH colors, and accessibility standards. Use when creating components, pages, or any UI elements. ALL agents MUST read and validate against design system before generating code.

Decompose complex questions into 4-6 parallel investigations with different perspectives, synthesize into phased decision framework. Use when facing architecture decisions with multiple unknowns in crypto/trading data platforms.

Evaluate Claude Code skills against best practices including size, structure, examples, and prompt engineering quality. Provides comprehensive analysis with actionable suggestions.

Implements Anthropic's Orchestrator-Workers pattern where a central LLM dynamically breaks down complex tasks, delegates to specialized worker LLMs, and synthesizes results. Use for open-ended problems where subtask number and nature are unpredictable. Ideal for complex projects (0.7+ complexity) requiring adaptive planning.

Use this skill when users want to generate images using OpenAI's image generation API (DALL-E or gpt-image-1), or extract text from images using OCR. Invoke when users request AI-generated images, artwork, logos, illustrations, visual content from text prompts, or need to extract text/data from images.

Strategic design agent focused on codebase research and implementation blueprinting.

This skill guides creating slash commands for Claude Code—reusable Markdown-based prompts with YAML configuration. Use when building custom commands, designing command workflows, or extending Claude Code functionality.

Use this skill when coordinating multiple AI models (Claude, GPT, Gemini, DeepSeek) for competitive trading analysis, running parallel model execution, aggregating decisions, or managing model performance tracking. Essential for multi-model trading arena operations.

Write minimal, strategic tests focused on core user flows and critical paths rather than comprehensive coverage. When creating or modifying test files for unit tests, integration tests, or end-to-end tests. When working on test files (*.test.ts, *.spec.ts, __tests__/*). When writing tests for critical user workflows and primary functionality. When testing behavior rather than implementation details. When mocking external dependencies or ensuring fast test execution. When deciding which parts of the codebase require testing coverage.