This repository is a public distribution of selected skills.
The private skill-hub repository remains the source of truth. This export contains only skills explicitly listed in public-skills.yaml.
A companion documentation site is generated under docs/ and served via GitHub Pages.
| id | Name | Version | Path | Description |
|---|---|---|---|---|
content-visual-forge |
Content Visual Forge | v2.7.2 |
content-visual-forge/ |
视觉资产制作 skill，把任意输入源（PDF/网页/文章/截图/视频转写稿/单字或词表）转化为公众号封面、知识卡、单字卡、头图、Excalidraw/p5.js/PixiJS/ASCII 创意微资产，统一视觉风格批量输出。触发语义：用户要"做封面""出卡片""生成字卡""画头图""视觉化内容""风格探索""手绘风图解"等明确视觉资产请求时启用。产物始终是图像/PDF/HTML 可视化文件，不是文字报告或长文。 |
roundtable-forge |
Roundtable Forge | v2.8.0 |
roundtable-forge/ |
Roundtable Forge. Use this skill when the user wants a multi-agent, cross-disciplinary roundtable discussion involving real people, historical figures, or fictional characters to explore a question in depth and breadth. Users may say things like "召集一个圆桌讨论", "让XX和YY讨论一下", "从不同角度分析这个问题", "跨学科圆桌", "多角色研讨", "召唤专家讨论", "每个角色一个 Agent", or "use roundtable-forge". |
deep-research-forge |
Deep Research Forge | v1.13.0 |
deep-research-forge/ |
深度研究与决策分析 skill。用于系统研究一个产品、公司、人物、概念、技术、赛道或文化现象，先选择研究方法论组合，再建立研究问题、证据账本、正式状态层级和结论级引用映射，动态组合时间轴、竞品截面、用户选择、生态地图、因果机制、反方证据、场景推演和决策模块；复杂研究可启用多 Agent 并行执行，把来源搜证、时间线、竞品截面、反方证据和决策综合拆成并行研究小队，最后输出研究报告、决策简报、复盘评分或可复用研究资产。用户会说“研究一下”“深度分析”“竞品分析”“帮我搞懂”“横纵分析”“做个 deep research”“多 agent 并行研究”“这个公司/产品/概念是什么来头”“值不值得关注/投入/学习/跟进”等。 |
self-evolving-memory |
Self-Evolving Memory | v7.3.3 |
self-evolving-memory/ |
为 AI Agent 部署分层自进化记忆系统的 mentor skill。核心状态保存在本地文件；语义检索和定时触发是可选 Host 能力。覆盖三层存储（即时 ≤20KB / 近中期 / 长期检索）、事务化巩固与回滚、多因子晋升、事件因果图谱、主题索引、回忆规划器和证据账本。当用户要初始化或修复 Agent 记忆、轻量记录、跑巩固、评估晋升、回滚校验、建索引或规划检索时触发。 |
Use a skill by naming it in your request, for example:
Use $content-visual-forge to create a WeChat cover and inline image package.
The docs/ directory contains a pre-generated static site (landing + one page per skill).
Configure GitHub Pages to serve from main branch /docs folder to publish it.
Do not edit generated exports by hand. Update the private source skill, then run:
python scripts/export_public_skills.py --output <public-repo-path> --cleanPass --no-site to skip regenerating the documentation site.