LLM 與 Agent

OpenCode CLI - open source AI coding agent with TUI, IDE integration, LSP support, and MCP servers.

This skill should be used when an agent is assigned to write a section of technical documentation. Provides guidance on breaking down large tasks, using doc-coauthoring workflow, and coordinating with VibeKanban task management. Use when task description mentions "write section", "document this component", or "expand documentation".

Comprehensive guide to localStorage patterns for the Plataforma B2B de treinamento técnico corporativo educational platform. This skill covers schema design, error handling strategies, quota management, and synchronization patterns essential for building resilient web applications with client-side persistence.Learn how to handle common localStorage errors (QuotaExceededError, SecurityError, DOMException), implement graceful degradation, manage storage quotas (5-10MB browser limits, 50KB per note), and sync localStorage with React state. The skill emphasizes defensive programming, user-friendly error messages, and fallback strategies.Real-world examples are taken directly from the project codebase, including CNotesView, BashNotesView, and auto-save patterns used across 5 learning systems. Each pattern is demonstrated with production code showing how localStorage is used to persist 227 educational modules' progress and user notes.Key topics include try/catch patterns for all localStorage operations, Quota

Performs pragmatic code reviews for Next.js/React projects using Perimetre framework patterns. Use when: doing code review, reviewing pull requests, checking code quality, reviewing changed files, analyzing git diffs, or when user requests code review of recent changes.

Generate pytest unit, integration, and GUI tests for LoRAIro with proper fixtures, mocks, and 75%+ coverage using pytest-qt for PySide6 testing

Delegate GitHub operations and research tasks to Copilot CLI to preserve Claude Code session limits. Automatically activates when detecting GitHub keywords (issue, PR, repo, commit) or research needs (compare, research, lookup, best practices). Offloads heavy operations to Copilot subprocess and returns compressed results only.

Expert guidance on parallel development workflows using git worktrees and multi-agent execution. Use when users mention parallel work, concurrent development, speeding up development, working on multiple features simultaneously, or scaling team productivity. Activate for questions about task decomposition, worktree management, or parallelization strategies.

TypeScriptで型安全なClaude Code Hooksを作成するためのガイド。cc-hooks-tsライブラリを使用してフックを実装する際に使用する。(1) Claude Code用の新しいhookを作成する時、(2) 既存のhookをTypeScriptで書き直す時、(3) PreToolUse/PostToolUse/SessionStart等のイベントフックを実装する時、(4) plugin.jsonにhooksを設定する時

Create, write, author, or design Agent Skills for Claude Code. Use when users want to create a new skill, update an existing skill, write SKILL.md files, design skill structure, troubleshoot skill discovery, or convert prompts/workflows into reusable Skills.

This skill should be used when creating, validating, or converting Product Requirements Documents (PRDs) to LLM-native format. Use this skill when the user asks to write a PRD, review a PRD for compliance, convert an existing PRD to machine-readable format, or ensure PRD quality for AI agent consumption. This skill is optimized for users working with agentic coding frameworks who need strict requirement specifications to prevent context poisoning, hallucination, and mesa-optimization.

Add comprehensive debugging and observability tooling for backend storage layers (PostgreSQL, ChromaDB) and startup metrics. Includes storage drift detection, raw data inspection endpoints, and a Next.js admin dashboard.

This skill guides the process of creating context-free queries for a web search agent (another Claude instance with web search tools). Use this when you need expert guidance from web research but the search agent has no access to the codebase or conversation history.

Interactive skill creator for Claude Code. Triggers when user mentions creating skills, building skills, skill templates, skill frontmatter, allowed-tools, or wants to scaffold a new skill.

Design vector database ingestion and retrieval pipelines (points + payloads, filtered similarity search, multi-stage hybrid retrieval, index maintenance). Use when building RAG/vector search flows or debugging retrieval quality; triggers: vector database, RAG, embeddings, hybrid search, filtered search, Qdrant, Weaviate, Chroma.

Risk management and constitutional compliance specialist. Ensures all trading decisions comply with constitutional limits and manages portfolio risk. Integrates Constitution Article 4 (Risk Management).

Ensures that documentation files like README.md and AGENTS.md stay in sync with the actual code and project structure. Use when adding new features or changing project architecture.

Test Driven Development (TDD) workflow automation for Java Spring Boot projects. Use for writing tests, checking coverage, generating test scaffolds, and implementing Red-Green-Refactor cycles with JUnit 5, Mockito, and Testcontainers.

Optimizes code and system performance

Guide for creating Claude Code subagents (custom agents) - specialized AI assistants with isolated context. Use when the user wants to create, add, or configure a subagent/custom agent for Claude Code.

Proactive Claude Code permission manager. Configures permissions via natural language for CLI tools (git, gcloud, aws, kubectl, maven, gradle, npm, docker), project types (Rust, Java, TypeScript, Python), and file patterns. Auto-detects project types and researches unknown tools.