菜单
Running AI models locally with Ollama and llama.cpp. Free inference, privacy, custom model deployment. Use for cost-free routine tasks or private data processing. Don't use when quality of Opus is needed.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Machine learning model training with HuggingFace. Fine-tuning LLMs, dataset creation, GPU selection, training monitoring. Use for training custom models. Don't use for using pre-trained models via API.
Estimates project cost at market rates. Scans codebase, counts APIs, integrations, and complexity. Shows what the project would cost without AI. Use for pitches and ROI demonstrations.
Marketing and conversion optimization. Landing page CRO, copywriting, positioning, GTM strategy, A/B testing. Don't use for product development or technical decisions.
Product management frameworks. PRD writing, user stories, OKRs, roadmap, TAM/SAM/SOM analysis, competitive research. Don't use for technical implementation.
Scraping specific platforms with ready connectors. Google Maps businesses, Twitter trends, YouTube subtitles, Instagram/TikTok content. Use for platform-specific data collection.
Data analysis from external services. Connects to Google Sheets, Notion, HubSpot via Composio. Builds charts, runs queries, generates insights. Don't use for database queries.
| name | local-models |
| description | Running AI models locally with Ollama and llama.cpp. Free inference, privacy, custom model deployment. Use for cost-free routine tasks or private data processing. Don't use when quality of Opus is needed. |
| disable-model-invocation | true |
| model | sonnet |
| effort | medium |
| allowed-tools | Read, Write, Bash |
# Install model
ollama pull qwen2.5:7b
# Run
ollama run qwen2.5:7b
# API endpoint
curl http://localhost:11434/api/generate
# Start server
llama-server --model model.gguf --port 8001
# Point Claude Code to local
ANTHROPIC_BASE_URL=http://localhost:8001