OKF (Open Knowledge Format) — curated catalog of tools, plugins, skills, proposals, and docs for agent-friendly knowledge. YAML-driven, agent-searchable, MCP-ready.
# Add to your Claude Code skills
git clone https://github.com/Albertchamberlain/Awesome-OKFAwesome-OKF is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Albertchamberlain. OKF (Open Knowledge Format) — curated catalog of tools, plugins, skills, proposals, and docs for agent-friendly knowledge. YAML-driven, agent-searchable, MCP-ready. It has 102 GitHub stars.
Awesome-OKF's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/Albertchamberlain/Awesome-OKF" and add it to your Claude Code skills directory (see the Installation section above).
Awesome-OKF is primarily written in Python. It is open-source under Albertchamberlain on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh Awesome-OKF against similar tools.
No comments yet. Be the first to share your thoughts!
Unlocks once the catalog security scan passes (runs nightly).
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
Open Knowledge Format (OKF) is an open specification by Google Cloud — define knowledge as a directory of Markdown files with YAML frontmatter and a small set of conventions. No runtime, no SDK.
Awesome OKF keeps the entire OKF ecosystem in one validated YAML catalog, then turns it into a browsable list, a searchable CLI, and an MCP meta-server that AI agents can query directly:
catalog.yaml
│
┌────────────────┼────────────────┐
▼ ▼ ▼
README.md CLI tools MCP server
(human-browsable) (searchable) (agent-searchable)
Edit one record. Regenerate the docs. Re-query from anywhere.
Anything readable → OKF. Every format above — and any you invent — collapses into one intermediate shape, then bundles into OKF:
┌────────── EXTRACT ──────────┐ ┌── NORMALIZE ──┐ ┌──── BUNDLE ────┐
│ Markdown · Typora · PDF │ │ title │ │ # title │
│ Obsidian · Notion · Feishu │ → │ url │ → │ description │
│ GitHub · JSON · YAML │ │ description │ │ resource │
│ CSV · key-value · OCR │ │ body / tags │ │ myokf-ready │
└─────────────────────────────┘ └────────────────┘ └────────────────┘
title / url / description (+ optional body, tags).--format anything is the programmatic form: feed it arbitrary text and it walks the parser chain (list → JSON → key-value → URLs → line-per-row) until one sticks:
python scripts/convert-to-okf.py whatever.txt --format anything -o kb/
The MCP tool convert_to_okf speaks the same idea — an agent hands it arbitrary text, it hands back OKF.
User (or Agent):
"Find an OKF plugin for Obsidian vault conversion."
Agent calls:
search_catalog({
"query": "Obsidian",
"kind": "plugin",
"limit": 3
})
Awesome-OKF responds:
┌──────────────────────────────────────────────────────────────┐
│ obsidian-to-okf plugin │
│ Convert Obsidian vaults to OKF — wikilinks become OKF links. │
│ Platform: python · Tags: obsidian, wikilink, markdown │
└──────────────────────────────────────────────────────────────┘
The catalog speaks OKF.
Add Awesome OKF to any MCP client so your agent can discover OKF resources:
pipx install awesome-okf
{
"mcpServers": {
"awesome-okf": {
"command": "awesome-okf-server"
}
}
}
awesome-okf stats
awesome-okf list --kind plugin
awesome-okf search obsidian
awesome-okf readme
Zero-dependency CLI that converts content from your favorite platforms into OKF knowledge bundles:
| 📥 Input Format | ✨ What It Does |
|---|---|
| 📋 Markdown awesome-xx lists | Extracts - [Title](URL) — Description items → OKF entries |
| 📊 JSON arrays | Converts {title, url, description} objects → OKF entries |
| 🔗 URL lists | Plain text URL collections → OKF entries |
| 🧾 YAML files | Lists/mappings of {title, url, description} — needs pyyaml (optional) |
| 📑 CSV tables | title/url/description columns; first column wins when there's no title column |
| 🗂️ Key-value text | Generic key: value blocks — frontmatter, properties files, custom formats |
| 🐙 GitHub repos | Paste a repo URL — metadata + README fetched live → OKF entry |
| 📓 Obsidian vaults | Local vault directory → one entry per note, wikilinks resolved |
| 📝 Notion exports | Notion "Export → Markdown" directory → one entry per page |
| 🦩 Feishu docs | Feishu-exported Markdown → one entry per document |
| 🖋️ Typora notes | Plain .md files — point at a file or folder, done |
| 📕 PDF documents | Text layer extracted via pymupdf (optional dep); scanned pages print an OCR recipe |
| 🖼️ Scans & images | No bundled OCR — prints a ready-to-run PaddleOCR recipe instead |
# Every platform, one command
python scripts/convert-to-okf.py README.md -o kb/ -t concept
python scripts/convert-to-okf.py https://github.com/GoogleCloudPlatform/knowledge-catalog --format github -o kb/
python scripts/convert-to-okf.py my-vault/ --format obsidian -o kb/
python scripts/convert-to-okf.py notion-export/ --format notion -o kb/
python scripts/convert-to-okf.py feishu-doc.md --format feishu -o kb/
# Structured data
python scripts/convert-to-okf.py data.yaml --format yaml -o kb/ # needs: pip install pyyaml
python scripts/convert-to-okf.py table.csv --format csv -o kb/
python scripts/convert-to-okf.py notes.properties --format kv -o kb/
# Text-layer PDFs
pip install pymupdf
python scripts/convert-to-okf.py paper.pdf --format pdf -o kb/
# Scans / image-only PDFs → OKF: OCR first (your local engine, never our model), then convert
python scripts/convert-to-okf.py scan.png --format image # prints the OCR recipe
pip install paddleocr paddlepaddle
paddleocr ppocr -i scans/ --type ocr --lang en -o ocr-text/
python scripts/convert-to-okf.py ocr-text/ --format notion -o kb/
# Output: kb/ with Markdown files, each with YAML frontmatter
# Ready for: myokf validate kb/
The data-driven catalog CLI (same architecture as Awesome-MCP):
| 🔍 Command | 📝 Purpose |
|---|---|
awesome-okf search obsidian |
Full-text search across all 29 entries |
awesome-okf list --kind plugin |
Filter by category |
awesome-okf readme |
Regenerate this README from catalog.yaml |
awesome-okf-server |
MCP meta-server — let AI agents query the catalog |
| 🏆 Repository | 📌 What It Offers |
|---|---|
| ⭐ yzfly/awesome-okf | 中文世界第一个 OKF 落点 — 7 plugins + 7 skills + 3 proposals |
| ⭐ linyiru/awesome-okf | English OKF resource hub — spec, tools, samples, guides |
| 📚 GoogleCloudPlatform/knowledge-catalog | Official OKF spec, SDK, and proposals by Google |
| 🧠 karpathy/llm-wiki | The original LLM Wiki that inspired OKF |
29 curated entries · 4 tools · 7 plugins · 7 skills · 5 proposals · 6 docs Deliberately curated — not an exhaustive index.
cli — Unified CLI for OKF — pull from GitHub, validate, and package to single-file web. — cli, python, validation, packagingcli — CLI tool to convert Markdown awesome-xx lists, JSON arrays, and URL lists into OKF knowledge bundles — zero dependencies, standard library only. — conversion, cli, markdown, jsoncli — Built-in OKF schema validator — checks YAML frontmatter, link integrity, and spec compliance. — validation, quality, schemapython — Official Python SDK for reading, validating, and writing OKF bundles — referenced by Google as the reference implementation. — sdk, python, officialpython — Unified CLI entry point wrapping all seven producer plugins. — cli, aggregator, zero-dependencypython — Extract code symbols from GitHub repositories into OKF. — github, code, symbols, zero-dependencypython — Convert Feishu (Lark) knowledge spaces and documents into OKF. — feishu, lark, document, zero-dependency