自然語言處理

Orchestrate comprehensive assessment of newly created skills to determine if they should auto-trigger using context-gathering, code-analyzer, and optional research-expert agents with prioritized evaluation criteria

This skill should be used when the user asks to "create a diagram", "draw a flowchart", "generate a sequence diagram", "make a state diagram", "create a class diagram", "draw an ER diagram", "update diagram", "edit mermaid", "improve diagram layout", or mentions mermaid diagrams in any context.

Verify each generated claim against provided sources, enforce inline citations, and prune unsupported text. Use when outputs must be strictly grounded.

Pre-sales intelligence system. Ingests incoming emails, analyzes sender context, clusters related conversations, detects business opportunities, builds relationship timelines, and surfaces high-potential prospects for outreach.

Create hyper-detailed task specifications with embedded context, eliminating implementation lookups. This skill should be used for breaking features into implementable tasks with complete architecture references.

Model Context Protocol (MCP) server implementation specialist for Claude Desktop integration. Handles TypeScript/Node.js server scaffolding, endpoint creation, telemetry setup, npx distribution, and comprehensive documentation. Follows MCP specification and best practices for production-grade server deployment.

Calculate readability scores (Flesch-Kincaid, Gunning Fog, SMOG) and grade levels for text. Analyze writing complexity and get improvement suggestions.

Persistent project documentation system that maintains context across sessions. Creates structured Memory Bank files to preserve project knowledge, decisions, and progress.

Generates LLM-optimized code context with function call graphs, side effect detection, and incremental updates. Processes JavaScript/TypeScript codebases to create compact semantic representations including multi-level summaries, entry point identification, and hash-based change tracking. Provides 74-97% token reduction compared to reading raw source files. Useful for understanding code architecture, debugging complex systems, reviewing pull requests, and onboarding to unfamiliar projects.

Guide for building MCP (Model Context Protocol) servers that integrate external APIs/services with LLMs. Covers Python (FastMCP) and TypeScript (MCP SDK) implementations.

Manage Model Context Protocol (MCP) servers - discover, analyze, and execute tools/prompts/resources from configured MCP servers. Use when working with MCP integrations, need to discover available MCP capabilities, filter MCP tools for specific tasks, execute MCP tools programmatically, access MCP prompts/resources, or implement MCP client functionality. Supports intelligent tool selection, multi-server management, and context-efficient capability discovery.

Decide how to codify knowledge for reuse. Use when evaluating whether a pattern, solution, or insight should be formalized—and in what form (types, code structure, CLAUDE.md, README, skill, or code comments).

Model Context Protocol (MCP) integration specialist. Use when creating MCP server configurations, implementing MCP integrations, or optimizing MCP performance. Specializes in MCP server architecture and integration patterns.

Draft professional emails for various business and personal contexts

Activates automatically when designing or implementing Cloudflare Workers APIs, orchestrating multi-provider API calls, implementing D1 database operations, or managing KV-based caching strategies. Ensures proper error handling, provider tagging, fallback chains, and orchestration patterns for BooksTrack backend infrastructure. Auto-activates on: - File patterns: *.worker.js, *-api.ts, *-service.ts in backend code - Keywords: "api orchestration", "multi-provider", "fallback chain", "D1 query" - Context: Cloudflare Workers development, API design discussions

Stage specific line ranges in git by constructing and applying patches programmatically. Use when: (1) asked to create atomic commits but a file contains multiple unrelated changes, (2) part of a change is ready for commit while another part needs more work, (3) separating a bug fix from surrounding refactoring in the same file, (4) the user asks to commit only specific lines or hunks from modified files, (5) git add -p is unavailable due to non-interactive context, (6) building tooling that needs programmatic control over what gets staged

Provides comprehensive guidelines for writing and formatting MDX content, frontmatter, and Docusaurus-specific components for the Physical AI & Humanoid Robotics textbook.

Extract text, tables, and metadata from PDF files. Use when working with PDFs, document extraction, or parsing PDF content.

Processes incoming emails for Unite-Hub. Extracts sender data, identifies communication intents, links to CRM contacts, analyzes sentiment, and updates contact records with AI insights.

Create Architecture Decision Records (ADRs) documenting key architectural choices. Analyzes context, evaluates alternatives, and captures decisions using standard ADR template. Use when documenting database design, technology selection, pattern choices, or architectural trade-offs.