LLM 與 Agent

Multi-agent coordination patterns for OpenCode swarm workflows. Use when working on complex tasks that benefit from parallelization, when coordinating multiple agents, or when managing task decomposition. Do NOT use for simple single-agent tasks.

Crypto wallet assistant for generating payment links for supported tokens (USDC, USDT, vUSD, ICP, CKUSDC) across multiple networks (Solana, Ethereum, Base, BSC, Vly, ICP), X402 redirect links for accessing payment-protected content, and providing access to the vly.money wallet interface. Use when users request to send/transfer crypto, mention 402 errors or payment-required content, or explicitly ask to access vly.money wallet.

Implement large language model (LLM) chat completions using the z-ai-web-dev-sdk. Use this skill when the user needs to build conversational AI applications, chatbots, AI assistants, or any text generation features. Supports multi-turn conversations, system prompts, and context management.

Brief description of what this skill does. Should be clear enough that Claude knows when to invoke it. Max 1024 characters.

Create exceptional, deterministic Claude skills that guide AI behavior with precision. Use this skill when the user asks to create new skills, document capabilities, or codify expertise into reusable instructions. Generates structured skill files that produce consistent, high-quality outputs.

Quantum-resistant, self-learning version control for AI agents with ReasoningBank intelligence and multi-agent coordination

Structure LLM context using XML tags for maximum attention efficiency. Use when writing system prompts, organizing complex instructions, or engineering context for Claude agents. Produces semantically meaningful XML that creates clear attention anchors.

Drafts a TIL blog post in the user's voice and creates it in Notion with Status="Claude Draft". Contains voice guide for matching the user's writing style. Use when user approves a TIL topic and wants a draft created.

This skill should be used when the user asks to "create an agent", "write a subagent", "generate agent definition", "add agent to plugin", "write agent frontmatter", "create autonomous agent", "build subagent", needs agent structure guidance, YAML frontmatter configuration, invocation criteria with examples, or wants to add specialized subagents to Claude Code plugins with proper capabilities lists and tool access definitions.

Creates Agentforce agents using Agent Script syntax with 100-point scoring. Use when building AI agents, defining topics and actions, or deploying AiAuthoringBundle (v65+) or GenAiPlannerBundle (v65+) metadata.

Comprehensive Salesforce testing skill with test execution, code coverage analysis, and agentic test-fix loops. Run Apex tests, analyze coverage, generate test patterns, and automatically fix failing tests with 120-point scoring.

Second-pass context optimization that analyzes user prompts and removes irrelevant specs, agents, and skills from loaded context. Achieves 80%+ token reduction through smart cleanup. Activates for optimize context, reduce tokens, clean context, smart context, precision loading.

This skill should be used when creating a Claude Code slash command. Use when users ask to "create a command", "make a slash command", "add a command", or want to document a workflow as a reusable command. Essential for creating optimized, agent-executable slash commands with proper structure and best practices.

Expert PM validation for increment closure with 3-gate quality checks (tasks, tests, documentation). Acts as final quality gate before closing increments. Validates task completion, test coverage, documentation updates, and detects scope creep. Activates for /done command, increment closure, close increment, validate increment, completion check, quality gate, PM approval, ready to close, scope creep detection, release validation, increment validation.

Instant status commands that bypass LLM processing. Execute scripts directly for /sw:status, /sw:progress, /sw:jobs. Activates for specweave status, specweave progress, specweave jobs, increment status, show status, show progress, background jobs, job status.

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 Claude's capabilities with specialized knowledge, workflows, or tool integrations.

Explains how SpecWeave achieves context efficiency through Claude's native progressive disclosure mechanism and sub-agent parallelization. Skills load only when relevant, sub-agents isolate context. Activates when users ask about context loading, token usage, or how SpecWeave scales. Keywords: context loading, progressive disclosure, token efficiency, sub-agents, context management.

Expert Claude Code plugin development covering plugin structure, slash commands, auto-activating skills, sub-agents, plugin.json configuration, YAML frontmatter, activation keywords, directory structure, and plugin best practices. Activates for plugin development, create plugin, claude plugin, slash command, skill activation, SKILL.md, plugin.json, claude code plugin, how to make plugin.

Node.js/TypeScript backend developer. Builds Express.js, Fastify, NestJS APIs with Prisma ORM, TypeORM, Mongoose. Implements REST APIs, GraphQL, authentication (JWT, session, OAuth), authorization, database operations, background jobs, WebSockets, real-time features, API validation, error handling, middleware. Activates for: Node.js, NodeJS, Express, Fastify, NestJS, TypeScript backend, API, REST API, GraphQL, Prisma, TypeORM, Mongoose, MongoDB, PostgreSQL with Node, MySQL with Node, authentication backend, JWT, passport.js, bcrypt, async/await, promises, middleware, error handling, validation, Zod, class-validator, background jobs, Bull, BullMQ, Redis, WebSocket, Socket.io, real-time.

Creates and refines Claude agent skills following best practices. Use when creating new skills, improving existing ones, or learning about skill structure and conventions.