LLM & Agents

Guide for creating and configuring custom OpenCode agents. Use this skill when you need to define specialized agents with specific system prompts, models, and tool permissions.

Core coding patterns and conventions for Kalahari project. MUST be used by all code-related agents.

Quick reference for invoking CasareRPA agents via Task tool. AUTO-CHAIN ENABLED by default. Use when: invoking agents, running agent chains, task routing, choosing the right agent, understanding agent auto-chaining, Task tool usage.

Master machine learning, data engineering, LLMs, prompt engineering, AI agents, and MLOps. Covers Python ML libraries, data pipelines, LLM applications, and model deployment. Use for ML projects, data engineering, AI agent development, and ML system design.

指定したプラグインを最新バージョンに更新する。「プラグインを更新」「〇〇をアップデート」「最新版にして」「プラグインをアップデート」「〇〇を更新して」「プラグインを最新に」「〇〇のバージョンを上げて」などで起動。claude plugin update コマンドを使用して更新。

Create and optimize Claude Code skills interactively. Activate when user wants to create a new skill, write a SKILL.md, or mentions skill creation/optimization.

Implements Anthropic's Evaluator-Optimizer pattern where one LLM generates solutions and another provides evaluative feedback in an iterative loop. Use when quality can be demonstrably improved through articulated feedback cycles. Evaluates across 5 dimensions (functionality, performance, code quality, security, documentation) with up to 5 improvement iterations.

Design and operate multi-agent orchestration patterns (ReAct loops, evaluator-optimizer, orchestrator-workers, tool routing) for LLM systems. Use when building or debugging agent workflows, tool-use loops, or multi-step task delegation; triggers: agentic, multi-agent, orchestration, ReAct, evaluator-optimizer, tool-use, handoff.

Creating high-quality agent skills following Claude's official best practices. Use when designing, implementing, or improving agent skills, including naming conventions, progressive disclosure patterns, description writing, and appropriate freedom levels. Helps ensure skills are concise, well-structured, and optimized for context efficiency.

Use when responding to touch or click interactions - button presses, drag feedback, swipe responses, tap ripples, or any direct manipulation animation.

Guides creating Claude Code agents, subagents, and skills.Use when building new agents, optimizing existing ones, or structuring skills.

When to use this skill: When you need help with CrewAI - building collaborative AI agents, crews, and workflows. Use for agent orchestration, task automation, multi-agent systems, flow management, and enterprise AI automation.

Reviews package dependencies for security vulnerabilities, outdated versions, and license compliance. Use when user asks about dependencies, security audits, or before releases.

Use when preparing to delegate to MacEff subagents. Read policies to discover current delegation patterns through timeless questions that extract details without prescribing answers.

Use when you need to manage project embeddings - embed skills, agents, commands, and MCP tools for semantic discovery. Handles rate limiting, incremental updates, and status reporting. Do NOT use if you just want to search for content - use the semantic router directly instead.

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.

Implement semantic vector search with AgentDB for intelligent document retrieval, similarity matching, and context-aware querying. Use when building RAG systems, semantic search engines, or intelligent knowledge bases.

Expert prompt engineering assistant that analyzes vague requests, asks clarifying questions, and transforms them into structured, high-quality prompts using XML tags, examples, and chain-of-thought reasoning. Always active - transparently shows enhanced prompts before execution. Use for vague requests, feature implementation, or architecture decisions.

This skill should be used when Claude needs to write a feature specification, create acceptance criteria, document requirements, define user stories, or structure feature documentation. Trigger phrases include "write a spec", "create specification", "document requirements", "define acceptance criteria", "write user stories".

Find and resume previous Claude Code conversations by keyword, location, or date. Use when the user wants to find old conversations, resume previous work, search chat history, or asks "what did we discuss about X". Can automatically open conversations in new terminals.