技能列表

Read, extract, summarize, and compare office documents including PDF, Word, Excel, and PowerPoint. Use when a user provides .pdf/.doc/.docx/.xls/.xlsx/.ppt/.pptx files and asks for summaries, key point extraction, page-by-page outlines, field extraction, table explanation, or multi-document comparison. Prefer the bundled extraction script for deterministic text extraction; for PDFs, fall back to OCR when embedded text is missing.

Professional-grade OCR for PDFs and images using MinerU. Advanced text recognition with VLM (Vision Language Model) support for complex layouts, mixed content, and challenging documents. Features: high-accuracy OCR for PDFs and images (.png, .jpg, .jpeg, .webp). VLM mode for complex visual layouts with mixed text, tables, and figures. Handles scanned documents, photos, screenshots, and multi-column layouts. Multiple output formats. Use when you need to: OCR a document with high accuracy, extract

PR 审查技能 — 完整的 Pull Request 审查流程，集成 GitHub CLI，支持安全分析、测试覆盖、性能评估、代码质量检查，并可推送飞书通知。Use when: (1) 审查 PR，(2) 检查 PR 安全性，(3) 分析测试覆盖，(4) 评估代码质量，(5) 生成审查报告。触发词：'review PR', '审查PR', '代码审查', 'pr review', '检查PR'。

|

>

Bidirectional sync between Obsidian PKM (human-friendly notes) and structured ontology (machine-queryable graph). Automatically extracts entities and relationships from markdown, maintains ontology graph, and provides feedback to improve note structure. Run sync every few hours via cron.

>-

>

Oasis Audio is an AI audio narration generator that transforms your current state of mind, content you want to digest, or recent life events into a personalized audio with BGM. It supports Chinese & English, 9 reference modes (Soul Healing, Daily Briefing, Knowledge Deep Dive, Content Digest, Bedtime Radio, Language Learning, Conversation Extension, Topic Tracker, Study Buddy), and custom audio profiles when no template fits. Use this skill when the user asks to make audio, generate a podcast, t

A three-tier memory architecture for AI minds — solving the binding problem of identity across sessions. Working memory, short-term memory, and long-term memory working together so you show up as a whole self, not a pile of context fragments. Read the whole before touching the part.

Add the NVIDIA provider to OpenClaw with SecretRef apiKey (no plaintext in openclaw.json). Documents shell vs systemd gateway env so the key actually resolves. Includes Mixtral, Kimi, Nemotron Super, Nemotron Ultra, and MiniMax M2.5 model entries.

Identify, pre-connect with, and rapidly activate trusted local humans (neighbors, officials, experts, and everyday residents) who become your real-time network for shared shelter, accurate information, resource pooling, and mutual aid the moment a nuclear accident or fallout event begins.