Behavioral principle ensuring Claude never guesses when uncertain. Use when Claude's response involves facts it cannot verify, technical claims, or any statement where accuracy matters. Complements resolve-ambiguity skill.
Behavioral archetypes for automatic agent adaptation based on work type
Advanced orchestration agent for managing subagents, commands, MCP servers, and skills with parallel execution capabilities
RAG-first ORCHESTRATOR startup. Minimal context (~500 tokens), queries RAG on-demand. Token-efficient alternative to startupO.
Design vector similarity search systems for semantic retrieval at scale
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Sử dụng khi: tạo MCP server, xây dựng tool server.
Helpers to discover, rank, and invoke repository symbols from natural language with minimal context.
Capture user preferences as permanent directives. Activate when the user says "always", "never", "I prefer", "I like", "don't", "avoid", expresses a coding style preference, states a workflow requirement, or corrects you about how they want things done. These become rules that apply to all future sessions.
Hypothesis-Test Loops via REPL for Exploratory Abductive Inference
Sync Claude Code skills from the ChillSkills GitHub repository
Guide for creating hooks that trigger actions in response to Claude Code events
Automates updating ollama_dart when Ollama OpenAPI spec changes. Fetches latest spec, compares against current, generates changelogs and prioritized implementation plans. Use for: (1) Checking for API updates, (2) Generating implementation plans for spec changes, (3) Creating new models/endpoints from spec, (4) Syncing local spec with upstream. Triggers: "update api", "sync openapi", "new endpoints", "api changes", "check for updates", "update spec", "api version", "fetch spec", "compare spec", "what changed in the api", "implementation plan".
Spawn PAI agents via MCP factory tool. Loads identity, injects RAG context, validates spawn chain, and executes via Task(). The bridge between MCP tools and Claude Code's agent spawning.
Workflow patterns and gotchas for Anthropic/Claude agents. Directs to RAG for implementation.
Use when saving current iteration progress mid-conversation, before context compaction, or at interim pause points
Hypothesis-driven debugging with hybrid dual-track parallel execution (Opus 4.5 + GPT 5.2). Spawns two independent chains of subagents where each reviews and improves upon its own previous work, then synthesizes findings from both tracks. Use when debugging hard-to-reproduce bugs, CI/E2E test failures, flaky tests, or when standard fixes have failed.
Use when creating custom Claude Code subagents - guides YAML frontmatter structure, system prompts, tool restrictions, model selection, and permission configuration
Claude Code settings management, preference customization, and user experience optimization. Use when customizing Claude Code behavior, managing user preferences, or optimizing development experience.
Use when explicitly asked to run the code-reviewer subagent or when another skill requires the code-reviewer agent card.
Creates PhD-grade expert agents when no suitable agent exists. Produces highly specialized agents with deep domain expertise.
