自然語言處理
1693 skills in 數據與 AI > 自然語言處理
building-agent-tools
Guide for creating effective tools for AI agents. Use when building MCP tools, agent APIs, or any tool interface that agents will consume. Focuses on token efficiency, meaningful context, and proper namespacing.
git-workflow
Git worktree-based workflow for parallel development with shared context. Incremental commits after logical units, feature branches with nested issue branches, squash-merge to main. Use when creating/managing worktrees, working on parallel features, committing code, creating branches, or handling branch cleanup.
tts
Implement text-to-speech (TTS) capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to convert text into natural-sounding speech, create audio content, build voice-enabled applications, or generate spoken audio files. Supports multiple voices, adjustable speed, and various audio formats.
language-precision
Use when text contains vague language ("many", "substantially"), passive voice overuse, or lacks quantitative specificity - provides 30+ before/after examples for active voice, precise terminology, and appropriate hedging
context-initializer
Automatically invokes init-explorer agent when project context is empty or unknown.
moai-alfred-context-budget
Enterprise Claude Code context window optimization with 2025 best practices: aggressive clearing, memory file management, MCP optimization, strategic chunking, and quality-over-quantity principles for 200K token context windows.
search-optimization
Guide on optimizing Bluera Knowledge search results through proper intent selection, detail level strategies, result limiting, and store filtering. Teaches when to use minimal vs contextual vs full detail, and how to choose the right search intent for different query types.
atlas
Ingest project documentation and code into persistent semantic memory (Qdrant + Voyage embeddings). Use when user wants to remember context across sessions, ingest docs, or search previous work. Requires Qdrant running locally and VOYAGE_API_KEY set.
blog-management
Manage blog posts including creating, editing, viewing, and troubleshooting blog-related features. Use when working with blog posts, articles, post forms, rich text editor, tags, thumbnails, or publishing workflows.
contacts
Get contact information from macOS. Use when texting, calling, emailing or social networking with people and companies.
ai-video-generation
AI video generation patterns using Sora, Runway, Pika, and other AI video tools. Use when generating videos from text prompts, image-to-video conversion, AI video editing, or integrating AI video APIs.
sparring
Critical thinking partner for technical concepts and strategy. Use when the user wants to explore technical ideas, validate assumptions, or develop strategy. Claude researches the user's Obsidian notes to understand their thinking patterns and context, identifies gaps and flawed assumptions, and provides constructive challenge. For AWS topics, Claude also consults AWS documentation to ensure technical accuracy.
implementing-ddd-architecture
Design and implement DDD patterns (entities, value objects, aggregates, CQRS). Use when creating new domain objects, implementing bounded contexts, designing repository interfaces, or learning proper layer separation. For fixing existing Deptrac violations, use the deptrac-fixer skill instead.
deployment-engineer
Expert deployment automation for cloud platforms. Handles CI/CD pipelines, container orchestration, infrastructure setup, and production deployments with battle-tested configurations. Specializes in GitHub Actions, Docker, HuggingFace Spaces, and GitHub Pages.
whisper-transcribe
Transcribes audio and video files to text using OpenAI's Whisper CLI with contextual grounding.Converts audio/video to text, transcribes recordings, and creates transcripts from media files.Use when asked to "whisper transcribe", "transcribe audio", "convert recording to text", or"speech to text". Uses markdown files in the same directory as context to improve transcriptionaccuracy for technical terms, proper nouns, and domain-specific vocabulary.
web-search
Search the web for up-to-date information using GPT-5.2 with web search and high reasoning. Prefer this over webfetch when you need to research complex queries, find current info across multiple sources, get recent news, up-to-date documentation, or verify facts that may have changed since your training cutoff. Returns cited sources.
moai-baas-cloudflare-ext
Enterprise Cloudflare Edge Platform with AI-powered edge computing architecture, Context7 integration, and intelligent global orchestration for scalable modern applications
guardian
Automatic quality gate and session health monitor. Spawns focused Haiku agents for code review and task planning when degradation detected. Validates suggestions against Oracle knowledge. Learns from user feedback to adjust sensitivity. Uses minimal context passing. Integrates with oracle and evaluator.
jupytext
This skill should be used when the user asks to "convert notebook to text", "use jupytext", "version control notebooks", "share data between kernels", "set up multi-kernel project", "pair notebooks with Python files", "sync ipynb and py files", or needs multi-kernel projects (Python/R/Stata/SAS) with version-control-friendly notebooks.
context-engineering
Comprehensive context engineering and management system based on the framework C = A(c₁, c₂, c₃, c₄, c₅, c₆). Use this skill when users work with context files, ask about context engineering, need to create/update/analyze contexts, validate context quality, detect discrepancies between contexts, or manage their Context Engineering system. Triggers include: "create a context", "analyze context", "check for inconsistencies", "validate context quality", "update context", "context engineering", "manage contexts", "c1-c6 components", or when working with .claude/contexts/ directories.