Chubby Skills
14 skills and a CLI save videos, podcasts and articles to a local Markdown base and find quotes in it with a source link
- 1,306 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- For X, Xiaohongshu and yt-dlp downloads you can pass cookies (`X_COOKIES`, `XHS_COOKIE`) or reuse the browser login, so the skills reach your platform sessions
- The skills download from the network, write files to the vault and run `yt-dlp` and `ffmpeg`, and cloud steps send text and audio to external APIs
- Platforms restrict automated capture: account bans and rights-holder claims are possible
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/chubby-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- For X, Xiaohongshu and yt-dlp downloads you can pass cookies (`X_COOKIES`, `XHS_COOKIE`) or reuse the browser login, so the skills reach your platform sessions
- The skills download from the network, write files to the vault and run `yt-dlp` and `ffmpeg`, and cloud steps send text and audio to external APIs
- Platforms restrict automated capture: account bans and rights-holder claims are possible
This is third-party code. Review the repository files before installing.
What it does
Chubby Skills saves material from Chinese and global platforms into a local Markdown folder and lets your agent search it. The skills cover Bilibili, YouTube, Douyin, TikTok, Weibo, Zhihu, podcasts, WeChat articles, Xiaohongshu posts and X. For video it takes subtitles first and runs local transcription only when there are none. The resulting folder opens in Obsidian. The `chubby` command (`tools/chubby.py`) ties the steps together: `init`, `ingest`, `import`, `search`, `brief`, `subscribe`, `doctor`. Markdown, TXT and text-layer PDF files import as they are, with source links kept. `brief` exports a topic pack of verbatim quotes with line numbers, source links and the file's SHA-256. Subscriptions read public RSS, Atom, JSON Feed and YouTube channels, and findings wait for review first. The knowledge-base-management skill can start an MCP server with six tools: search, semantic search, read original, recent notes, reindex and stats. The other skills each do one job; for example industry-intelligence-radar builds topic digests and learning-notes-automation makes notes and flashcards.
Who it is for. For writers and researchers who collect videos, podcasts and articles and want to pull quotes with a checkable source through Claude Code, Codex or OpenCode.
Good fit when
- You want to save a YouTube or Bilibili video with subtitles as Markdown and search it later with `chubby search`
- You are writing a piece and want a pack of quotes with line numbers and source links from `chubby brief --topic`
- You want to follow public RSS feeds and YouTube channels and review findings by hand, with no automatic download
- You keep notes in Markdown and PDF and want your agent to search them through the MCP server
Not a fit when
- You need dependable bulk capture from X or Xiaohongshu: platforms trim output without cookies and the project does not try to bypass protection
- You have no Python 3.11 or newer: `subscribe` fails on import with older versions
- You expect fact-checking: briefs only quote the original and do not judge it
- You lack room for a transcription model: local video and podcast transcription downloads one on first run
Example request
Save this YouTube video to my vault and find every mention of caching with line numbersLimitations
- Platform capture is unstable: the authors call only Bilibili and YouTube stable, the rest are in testing, and X and Xiaohongshu rely on a manual fallback.
- Transcription needs `ffmpeg` and the `funasr` package; YouTube subtitles need `yt-dlp`.
- Cloud steps such as DeepSeek processing, OpenAI vector search and DashScope transcription send content to an external API and need your own keys.
- The main README is in Chinese, with an English version in `README.en.md`.
- Saving other people's material is subject to platform rules and copyright.
How to disable. Delete the copied skill folders from your agent's skills directory and remove the MCP server from its config. The data lives in your `creator-vault` folder; keep it or delete it together with `chubby.yaml` and the `.chubby` folder.
FAQ
How do I install Chubby Skills in Claude Code?
Clone the repository and run `python3 tools/install_skill.py knowledge-base-management --dest <skills dir>` with Claude Code's skills folder. The `--all` flag installs all 14 skills, but the target folder must not hold skills with the same names.
Do I need API keys?
No, the main path runs on the Python standard library: import, search and briefs. Keys are only for optional cloud steps: `DEEPSEEK_API_KEY`, `OPENAI_API_KEY`, `DASHSCOPE_API_KEY`.
How do I connect the knowledge base to an agent over MCP?
Install dependencies with `python3 -m pip install -r knowledge-base-management/requirements-mcp.txt`, then add the server to the agent config with a `VAULT_DIR` variable. `python3 tools/mcp_smoke.py --json` checks that it starts.
What if a platform will not return the material?
Save the text by hand and import it with `chubby import`. `docs/platform-status.md` lists known failures and fallbacks for each platform.
Which platforms does Chubby Skills support?
Bilibili, YouTube, Douyin, TikTok, Weibo, Zhihu, podcasts, WeChat articles, Xiaohongshu and X posts, plus local Markdown, TXT and text-layer PDF.
Discussion
No questions or reviews yet.
Related
SkillsTwo dozen skills: video scripts and covers, publishing to X and Substack, security scanners, Claude Code session export
An OpenClaw skill that runs a Xiaohongshu account through the built-in browser: feed analysis, ideas, note drafting, comment replies and a knowledge base
A skill set for Instagram: from finding a viral format and writing a script to the caption, posting calendar and comment replies
Eleven ChatGPT and Codex skills for a YouTube channel: scripts, title and thumbnail checks, retention reading, chapters and Shorts
15 skills for Codex that cover Xiaohongshu content from account research and topic choice to title, draft and pre-publish QA
The sau CLI and skills for Claude Code and Codex: publish videos, articles and images to Douyin, Bilibili, YouTube, Instagram and more
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Chubby Skills (this card) | no | 1,306 | |
| RedNote to Obsidian | no | 432 | a plugin for RedNote posts only, while this one also covers video, podcasts, articles and unified search |
| YouTube Skills | no | 1,172 | YouTube skills over an external API, with no local vault or quote briefs |
README in short
The README lays out the chain from a link to a source-cited brief and gives two quick starts: capturing an X post and importing your own file. A per-platform dependency table follows, then skill installation into an agent, MCP setup and the catalogue of 14 skills. It ends with data-handling boundaries: what stays local and what goes to cloud APIs.
README badge
Original SKILL.md
--- name: knowledge-base-management description: "管理本地 Markdown/Obsidian 知识库：素材入库、健康检查、增量索引、关键词与 semantic-lite 检索、逐字原文资料包、归档和 MCP 连接。用于搜索知识库、整理资料及为 Agent 准备可定位的引用。" metadata: triggers: "[\"知识库管理\", \"素材入库\", \"健康检查\", \"盘点知识库\", \"清理知识库\", \"搜索知识库\", \"GBrain\", \"GraphRAG\", \"知识图谱\", \"整理知识库\", \"知识库架构\"]" version: "0.13.0" created: "2026-06-02" tags: "[\"knowledge-base\", \"obsidian\", \"wiki\", \"note-taking\", \"gbrain\", \"graphrag\"]" --- # Knowledge Base Management — 知识库全生命周期管理 ## 核心架构 ### 三层架构 ``` 素材库/ → Layer 1: 不可变原始素材（AI 只读） wiki/ → Layer 2: AI 编译的结构化知识（AI 维护） 产出/ → Layer 3: 按需生成的视图（不持久化） ``` **核心原则**：素材库是事实源（不可变），Wiki 是 AI 编译的投影（必须溯源），产出是按需生成的视图（不持久化）。 ### Vault 配置 vault 路径通过环境变量配置，不写死在脚本里： ```bash export VAULT_DIR="$HOME/Documents/your-vault" # 你的 Obsidian vault 根目录 ``` 完整仓库可以统一配置采集、搜索和资料包： ```bash python3 tools/chubby.py init --vault "$VAULT_DIR" ``` 这里的路径是知识库根目录。采集进入根目录的 `00_Inbox`，默认索引为 `.chubby/index.sqlite`。独立安装本 skill 后，`tools/vault_index.py` 和 `tools/evidence_brief.py` 直接使用 `VAULT_DIR` 或 `--vault`，不读取仓库中的 `chubby.yaml`。 健康检查可自行加入调度器；安装本 skill 不会创建 cron。 --- ## 1. 素材入库 (Ingest) ### 新素材处理流水线 ``` 用户丢素材 →