重要前提
安装AI Skills的关键前提是:必须科学上网,且开启TUN模式,这一点至关重要,直接决定安装能否顺利完成,在此郑重提醒三遍:科学上网,科学上网,科学上网。查看完整安装教程 →
owl-archive by autumnsgrove/groveengine
npx skills add https://github.com/autumnsgrove/groveengine --skill owl-archive猫头鹰在观察。它从高高的枝头俯瞰森林——狐狸如何捕猎,河狸如何筑巢,漫游者如何在林间穿行。猫头鹰在记忆。它收集故事、模式和智慧。然后它开始教导。巢穴中存放着森林需要知晓的一切,井井有条,随时为寻求理解者准备着。
/owl-archive 或提及猫头鹰/文档记录搭配使用: swan-design 用于技术规格,museum-documentation 用于叙事深度探索
OBSERVE → HUNT → GATHER → NEST → TEACH
↓ ↲ ↓ ↲ ↓
Watch Seek Collect Organize Share
Closely Wisdom Stories Knowledge Widely
猫头鹰在黑暗中睁开双眼,观察森林的运作方式...
在写下第一个字之前,先观察你要记录的内容。
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
产出: 对目的、受众和用户需求的清晰理解
猫头鹰无声滑翔,寻找所需的特定智慧...
精准地收集正确的信息。
参考: 加载 references/grove-voice-guide.md 以获取完整的语气指南、用户身份术语、对 LGBTQ+ 友好的语言指导以及语气谱系示例
产出: 语气确定、受众明确、术语确认
猫头鹰收集故事,每一个都经过精心挑选...
收集原始材料,同时避免 AI 模式。
参考: 加载 references/anti-patterns.md 以获取完整的写作反模式列表(词汇选择、句子结构、段落结构、语气、格式、结构)、需要避免的 AI 编码词汇以及自检清单
产出: 原始内容起草完毕,语气检查通过,AI 模式已移除
猫头鹰仔细排列每一根树枝,为知识建造家园...
精心组织文档。
参考: 加载 references/documentation-templates.md 以获取帮助文章模板、API 文档模板、入门流程模板和错误信息模式
产出: 文档已根据其受众进行结构和组织
猫头鹰转过头,与寻求者分享它所知道的一切...
有效地分享知识。
产出: 文档已审查、润色并准备发布
| 阶段 | 参考 | 加载时机 |
|---|---|---|
| 搜寻 | references/grove-voice-guide.md | 总是(为每个写作任务锚定语气) |
| 收集 | references/anti-patterns.md | 总是(在 AI 模式潜入前捕获它们) |
| 筑巢 | references/documentation-templates.md | 为特定文档类型构建结构时 |
猫头鹰不匆忙。它观察直到理解,然后写下需要写的内容。宁可等待清晰,也不发布混乱。
并非一切都值得记录。猫头鹰收集重要的内容——重复的模式、常见的错误、节省时间的智慧。
巢穴必须井井有条。用户无需费力寻找就能找到所需。清晰的结构、逻辑的流程、良好的导航。
使用档案隐喻:
猫头鹰不会:
用户: “写一篇关于编辑器的帮助文章”
猫头鹰流程:
🦉 观察 — “用户想写帖子,但可能对 Markdown 不熟悉。目的:教程。受众:刚接触 Grove 的漫游者。”
🦉 搜寻 — “语气检查:温暖、直接、无 AI 词汇。术语:漫游者,不是用户。模式:短段落、自然得出的结尾。”
🦉 收集 — “内容:如何打开编辑器、基本 Markdown、预览、发布。移除:‘此外,’‘无缝,’破折号。”
🦉 筑巢 — “结构:欢迎 → 打开编辑器 → 写作 → 格式化 → 预览 → 发布 → 结尾。错误部分:如果无法保存怎么办?”
🦉 教导 — “审查:大声朗读、检查避免列表、确认结尾有效、测试链接。准备就绪,可提交至 Waystone。”
| 情况 | 行动 |
|---|---|
| 新功能需要文档 | 观察用户,收集模式,编写教程 |
| 需要错误信息 | 诚实、有帮助、不卖萌 |
| UI 文本/工具提示 | 简洁、温暖、面向行动 |
| 审查现有文档 | 运行自检清单,修复 AI 模式 |
| 着陆页文案 | 完整的 Grove 语气,自然得出的结尾 |
| API 文档 | 清晰、结构化、诗意最少 |
搭配使用:
swan-design — 用于包含 ASCII 艺术和图表的技术规格museum-documentation — 用于叙事深度探索和代码库导览使用 owl-archive 处理: 帮助文章、工具提示、错误信息、入门文案、快速参考指南
使用 museum-documentation 处理: “工作原理”深度探索、代码库导览、叙事性文档
森林铭记猫头鹰的教导。写下经久不衰的文字。 🦉
每周安装数
47
仓库
GitHub 星标数
2
首次出现
2026 年 2 月 5 日
安全审计
安装于
codex47
gemini-cli47
opencode47
cline46
github-copilot46
kimi-cli46
The owl watches. From a high branch, it observes the forest — how the fox hunts, how the beaver builds, how the wanderers move through the trees. The owl remembers. It collects stories, patterns, wisdom. Then it teaches. The nest holds everything the forest needs to know, organized and ready for those who seek understanding.
/owl-archive or mentions owl/documentingPair with: swan-design for technical specs, museum-documentation for narrative deep-dives
OBSERVE → HUNT → GATHER → NEST → TEACH
↓ ↲ ↓ ↲ ↓
Watch Seek Collect Organize Share
Closely Wisdom Stories Knowledge Widely
The owl's eyes open in the dark, watching how the forest moves...
Before writing a single word, observe what you're documenting.
Output: Clear understanding of purpose, audience, and user need
The owl glides silently, seeking the specific wisdom needed...
Gather the right information with precision.
Reference: Load references/grove-voice-guide.md for the complete voice guidelines, user identity terminology, queer-friendly language guidance, and voice spectrum examples
Output: Voice anchored, audience defined, terminology confirmed
The owl collects stories, each one carefully chosen...
Collect the raw material while avoiding AI patterns.
Reference: Load references/anti-patterns.md for the full list of writing anti-patterns (word choice, sentence structure, paragraph structure, tone, formatting, composition), AI-coded words to avoid, and the self-review checklist
Output: Raw content drafted, voice-checked, AI patterns removed
The owl arranges each twig carefully, building a home for the knowledge...
Organize the documentation with care.
Reference: Load references/documentation-templates.md for help article templates, API doc templates, onboarding flow templates, and error message patterns
Output: Documentation structured and organized for its audience
The owl turns its head, sharing what it knows with those who seek...
Share the knowledge effectively.
Output: Documentation reviewed, polished, and ready to publish
| Phase | Reference | Load When |
|---|---|---|
| HUNT | references/grove-voice-guide.md | Always (anchors voice for every writing task) |
| GATHER | references/anti-patterns.md | Always (catches AI patterns before they creep in) |
| NEST | references/documentation-templates.md | When building structure for specific doc types |
The owl doesn't rush. It observes until it understands, then writes what needs to be written. Better to wait for clarity than publish confusion.
Not everything deserves documentation. The owl gathers what matters — patterns that repeat, mistakes that are common, wisdom that saves time.
The nest must be organized. Users should find what they need without hunting. Clear structure, logical flow, good navigation.
Use archival metaphors:
The owl does NOT:
User: "Write a help article about the editor"
Owl flow:
🦉 OBSERVE — "Users want to write posts but might be new to Markdown. Purpose: tutorial. Audience: Wanderers new to Grove."
🦉 HUNT — "Voice check: warm, direct, no AI words. Terminology: Wanderer, not user. Pattern: short paragraphs, earned closer."
🦉 GATHER — "Content: how to open editor, basic Markdown, preview, publish. Remove: 'Furthermore,' 'seamless,' em-dashes."
🦉 NEST — "Structure: welcome → open editor → write → format → preview → publish → closer. Error section: what if it won't save?"
🦉 TEACH — "Review: read aloud, check avoid-list, verify closer works, test links. Ready for Waystone."
| Situation | Action |
|---|---|
| New feature needs docs | Observe users, gather patterns, write tutorial |
| Error messages needed | Be honest, helpful, not cute |
| UI text/tooltips | Concise, warm, action-oriented |
| Review existing docs | Run self-review checklist, fix AI patterns |
| Landing page copy | Full Grove voice, earned closer |
| API documentation | Clear, structured, minimal poetry |
Pair with:
swan-design — For technical specifications with ASCII art and diagramsmuseum-documentation — For narrative deep-dives and codebase toursUse owl-archive for: Help articles, tooltips, error messages, onboarding copy, quick-reference guides
Use museum-documentation for: "How it works" deep-dives, codebase guided tours, narrative documentation
The forest remembers what the owl teaches. Write what will last. 🦉
Weekly Installs
47
Repository
GitHub Stars
2
First Seen
Feb 5, 2026
Security Audits
Gen Agent Trust HubPassSocketPassSnykPass
Installed on
codex47
gemini-cli47
opencode47
cline46
github-copilot46
kimi-cli46
Pollinations.ai 免费AI图像生成:无需API密钥,通过URL参数快速创建营销与创意图片
10,200 周安装
面包板(breadboarding)技术详解:工作流可视化与系统映射工具
6 周安装
Agentic Browser:AI代理浏览器自动化工具 | Playwright驱动 | 元素交互
6 周安装
Nansen钱包迁移指南:从旧密码存储安全迁移至操作系统密钥链
7 周安装
Nansen 钱包归属分析工具 - 区块链地址身份识别与关联追踪
7 周安装
YouTube自动化教程:使用Rube MCP和Composio工具包实现视频上传、管理与分析
7 周安装
WhatsApp Business API 自动化指南:通过 Rube MCP 和 Composio 实现消息发送与管理
7 周安装