{
  "markdown": "<h1 align=\"center\">Wiki Knowledge Compiler</h1>\n\n<p align=\"center\">\n  <em>Turn your documents into a personal knowledge base — powered by AI, stored as simple files.</em>\n</p>\n\n<p align=\"center\">\n  <a href=\"https://github.com/bradduy/wiki-knowledge-compiler/blob/main/LICENSE\"><img src=\"https://img.shields.io/badge/license-MIT-blue.svg\" alt=\"MIT License\" /></a>\n  <a href=\"https://github.com/bradduy/wiki-knowledge-compiler\"><img src=\"https://img.shields.io/badge/platform-Claude%20Code-blueviolet.svg\" alt=\"Claude Code\" /></a>\n  <a href=\"https://obsidian.md\"><img src=\"https://img.shields.io/badge/works%20with-Obsidian-7C3AED.svg\" alt=\"Works with Obsidian\" /></a>\n  <img src=\"https://img.shields.io/badge/offline-100%25-brightgreen.svg\" alt=\"100% Offline\" />\n</p>\n\n<p align=\"center\">\n  <img src=\"./assets/demo.gif\" alt=\"Wiki Knowledge Compiler Demo\" width=\"100%\" />\n</p>\n\n<p align=\"center\">\n  <img src=\"./assets/logo.png\" alt=\"Wiki Knowledge Compiler Logo\" width=\"100%\" />\n</p>\n\n## Table of Contents\n\n- [Recent Updates](#-recent-updates)\n- [Features](#-features)\n- [Localizations](#-localizations)\n- [Quick Start](#-quick-start)\n- [Usage](#-usage)\n  - [Adding Documents](#-adding-documents)\n  - [Example: From Document to Knowledge](#-example-from-document-to-knowledge)\n- [Knowledge Graph](#-knowledge-graph)\n- [Confidence & Smart Tracking](#-confidence--smart-tracking)\n- [Browse with Obsidian](#-browse-your-wiki-with-obsidian)\n- [How Setup Works](#%EF%B8%8F-how-setup-works)\n- [Other Ways to Install](#-other-ways-to-install)\n- [Contributing](#-contributing)\n- [License](#-license)\n\n---\n\n## 🆕 Recent Updates\n\n### v0.2.0 — Knowledge graph, lifecycle management, and wiki health\n- Added knowledge graph that extracts named entities (people, projects, technologies, decisions) during ingestion and connects them with typed relationships in `.data/entities/`\n- Added typed relationships between concepts with semantic meaning: `extends`, `contradicts`, `supersedes`, `depends-on`, `generalizes`, `component-of`\n- Added confidence scoring to every page (`high`/`medium`/`low`) with `verified` dates and source `authority` tracking (`primary`/`secondary`/`commentary`)\n- Added supersession — when new info replaces old, pages are linked with `supersedes`/`superseded_by` instead of being silently overwritten\n- Added graph-aware queries to `/wiki` — walks entity relationships alongside keyword/semantic search to find connections that text search misses\n- Added `/wiki-audit` command to check wiki health: broken links, orphan pages, stale content, missing frontmatter, citation mismatches, contradictions, entity graph issues — auto-fixes what it can, reports the rest\n- Added privacy filter that auto-strips API keys, tokens, passwords, and PII from derived content during ingestion (raw sources are never modified). Pages tagged with `visibility: public/private/internal`\n- Added contradiction resolution — when sources disagree, both sides are scored by recency, authority, and evidence count. Stronger claims supersede weaker ones; close calls flagged for human review\n- Added retention decay — confidence decays over time unless pages are accessed or reinforced. Each page type has its own half-life (concepts: 180 days, insights: 60 days, drafts: 30 days)\n- Added crystallization — when `/wiki` synthesizes 3+ sources, it auto-generates a digest page capturing the question, answer, sources, key connections, and lessons\n- Added quality scoring — every page scored 0–100 across 5 dimensions (frontmatter, citations, freshness, relationships, content). Grades A–F. Pages below 60 flagged during `/wiki-update` and `/wiki-audit`\n- Improved `/wiki-ingest` to extract entities, apply privacy filtering, and set confidence/authority on summaries\n- Improved `/wiki-update` to apply retention decay and compute quality scores after rebuilding indexes\n- Improved `/wiki-insights` to read entity graph for structural connections alongside topic and concept pages\n\n> **Upgrading from v0.0.x?** Run `/wiki-setup` again to initialize the new `.data/entities/` directory. Your existing wiki pages will continue to work — new fields (`confidence`, `verified`, `entities`, typed `related`) are added incrementally as you ingest new documents or run `/wiki-update`. No migration needed.\n\n### v0.0.x — Foundation\n- 📅 **Scheduled updates** — `/wiki-schedule` auto-runs `/wiki-update` on a cron schedule via remote agents.\n- 🔀 **Mixed ingest** — `/wiki-ingest` accepts files + URLs in a single command.\n- 🔄 **Auto-ingest on update** — `/wiki-update` detects and ingests new files dropped into `raw/`.\n- 📝 **Simpler commands** — `/wiki-query` renamed to `/wiki`.\n- 📁 **Flat structure** — Removed `knowledge-base/` wrapper; `raw/` and `wiki/` live at project root.\n- 📥 **Batch ingestion** — Folders, multiple files, and glob patterns.\n- 🔍 **Search tiers** — Grep for small projects, qmd for medium/large.\n- 🔮 **Obsidian integration** — Auto-install and setup during `/wiki-setup`.\n- ⚙️ **Cross-platform auto-install** — Node.js, qmd, Obsidian on macOS, Linux, and Windows.\n\n---\n\n## ✨ Features\n\n| Feature | Description |\n|---------|-------------|\n| **Knowledge graph** | Extracts people, projects, technologies, decisions as connected entities |\n| **Typed relationships** | Concepts and entities linked with semantic meaning (`extends`, `contradicts`, `depends-on`) |\n| **Confidence tracking** | Every page rated `high`/`medium`/`low` with verified dates |\n| **Retention decay** | Confidence decays over time unless reinforced — each page type has its own half-life |\n| **Quality scoring** | Every page scored 0–100 across 5 dimensions, grades A–F |\n| **Crystallization** | Multi-source answers auto-distilled into reusable digest pages |\n| **Wiki audit** | `/wiki-audit` finds broken links, stale content, contradictions and auto-fixes them |\n| **Privacy filter** | Auto-strips API keys, tokens, passwords, PII from derived content |\n| **Contradiction resolution** | Conflicting claims scored by recency/authority/evidence, resolved or flagged |\n| **Supersession** | Old claims linked to replacements, never silently deleted |\n| **Graph-aware queries** | `/wiki` walks entity relationships, not just keywords |\n| **Multi-format ingest** | Files, folders, URLs, glob patterns, pasted text |\n| **Auto summaries** | One-page summary per source with key takeaways |\n| **Concept extraction** | 3-10 atomic concepts per source, auto-deduplicated |\n| **Smart search** | Grep (small), qmd hybrid search (medium), qmd MCP (large) |\n| **Scheduled updates** | Cron-based auto-ingestion via remote agents |\n| **Obsidian integration** | Visual graph browsing, backlinks, live sync |\n| **100% offline** | Plain markdown files, no cloud, no database |\n| **Cross-platform** | macOS, Linux, Windows with auto-install |\n\n---\n\n## 🌐 Localizations\n\n<p align=\"center\">\n  🇬🇧 <strong>English</strong> &nbsp;·&nbsp;\n  <a href=\"./README_VN.md\">🇻🇳 Tiếng Việt</a> &nbsp;·&nbsp;\n  <a href=\"./README_CN.md\">🇨🇳 简体中文</a> &nbsp;·&nbsp;\n  <a href=\"./README_KR.md\">🇰🇷 한국어</a> &nbsp;·&nbsp;\n  <a href=\"./README_JP.md\">🇯🇵 日本語</a> &nbsp;·&nbsp;\n  <a href=\"./README_DE.md\">🇩🇪 Deutsch</a> &nbsp;·&nbsp;\n  <a href=\"./README_FR.md\">🇫🇷 Français</a> &nbsp;·&nbsp;\n  <a href=\"./README_RU.md\">🇷🇺 Русский</a>\n</p>\n\n---\n\n## 🚀 Quick Start\n\nThree steps. Everything installs automatically — you don't need to set up anything yourself.\n\n**Step 1 — Install the plugin:**\n```\n/plugin marketplace add bradduy/wiki-knowledge-compiler\n```\nAfter adding, configure it in the marketplace:\n1. Run `/plugin` → select **Marketplace** tab\n2. Find **wiki-knowledge-compiler** and open it\n3. Click **Enable** and choose the scope — pick the project folder where you want your wiki to live\n4. Confirm — this grants the plugin permission to create files and run commands in that folder\n\n<p align=\"center\">\n  <img src=\"./assets/user_scope_setting.png\" alt=\"Enable scope setting\" width=\"100%\" />\n</p>\n\n**Step 2 — Run setup:**\n```\n/wiki-setup\n```\nPick your project size (small, medium, or large). The plugin handles the rest — it installs any tools you need, builds your wiki folder, and even offers to set up [Obsidian](https://obsidian.md) so you can browse your wiki visually.\n\n<p align=\"center\">\n  <img src=\"./assets/obisidian.gif\" alt=\"Obsidian visual browsing\" width=\"100%\" />\n</p>\n\n**Step 3 — Add your first document:**\n```\n/wiki-ingest ~/Documents/my-article.md\n```\n\nDone. Your wiki is live.\n\n---\n\n## 🎯 Usage\n\nEach command tells you what to try next, so you always know the next step.\n\n| Command | What it does |\n|---------|-------------|\n| `/wiki-setup` | ⚙️ One-time setup (installs everything for you) |\n| `/wiki-ingest` | 📥 Add documents — one file, a whole folder, or a URL |\n| `/wiki` | 🔍 Ask a question, get an answer with sources |\n| `/wiki-insights` | ✨ Find patterns and connections across your sources |\n| `/wiki-update` | 🔄 Sync wiki (ingest new raw files + refresh indexes) |\n| `/wiki-schedule` | 📅 Schedule /wiki-update to run automatically |\n| `/wiki-audit` | 🩺 Check wiki health — fix broken links, detect contradictions, score quality |\n\n**Recommended order:** `/wiki-setup` → `/wiki-ingest` → `/wiki` → `/wiki-insights` → `/wiki-update` → `/wiki-audit` → `/wiki-schedule`\n\n### 📥 Adding Documents\n\nYou can add documents in many ways:\n\n```bash\n# One file\n/wiki-ingest ~/Documents/article.md\n\n# A whole folder (scans everything inside)\n/wiki-ingest ~/Documents/research/\n\n# Multiple files at once\n/wiki-ingest ~/notes/meeting1.md ~/notes/meeting2.md\n\n# All PDFs in a folder\n/wiki-ingest ~/papers/*.pdf\n\n# A URL\n/wiki-ingest https://example.com/interesting-article\n\n# Or just paste text directly after the command\n```\n\nFor folders and multiple files, the plugin shows progress as it works through each one.\n\n### 🧪 Example: From Document to Knowledge\n\nSay you add an article about climate change:\n\n```\n/wiki-ingest ~/Articles/ipcc-summary-2023.md\n```\n\nThe plugin creates:\n\n| What | Description |\n|------|------------|\n| **Summary** | A one-page overview of the article with confidence rating |\n| **Concepts** | Pages for key ideas like \"carbon budget\" and \"tipping points\" |\n| **Entities** | Nodes for real things — \"IPCC\", \"Paris Agreement\", \"Dr. Smith\" |\n| **Topic page** | A \"climate science\" page linking everything together |\n\nNow add a second article about energy policy. The plugin notices both sources discuss carbon budgets and links them automatically — concepts get typed relationships (`extends`, `contradicts`, `depends-on`), and entities connect across sources.\n\nWhen you ask:\n```\n/wiki How do carbon budgets affect energy policy?\n```\n\nThe plugin doesn't just search by keywords. It walks the **knowledge graph** — starting from the \"carbon budget\" concept, following its relationships to connected entities and topics — and finds connections that keyword search would miss. You get an answer that draws from **both** sources, with links to exactly where each fact came from.\n\n---\n\n## 🧠 Knowledge Graph\n\nEvery document you add doesn't just create pages — it builds a **knowledge graph**. The plugin extracts named entities (people, projects, technologies, decisions) and connects them with typed relationships.\n\n```\nRedis ──uses──→ Auth Service ──maintained-by──→ Brad\n  │                                              │\n  └──depends-on──→ PostgreSQL           owns ←───┘\n                        │                Auth Migration\n                        └──replaces──→ MySQL\n```\n\nThis means when you ask \"what's the impact of upgrading Redis?\", the plugin walks outward from Redis through `uses`, `depends-on`, and `maintained-by` edges — finding every connected service, person, and decision. Not just pages that mention \"Redis\" by name.\n\n**What gets extracted automatically:**\n\n| Entity type | Examples |\n|-------------|---------|\n| People | \"Brad Duy\", \"Andrej Karpathy\" |\n| Projects | \"Auth Migration\", \"API Redesign\" |\n| Technologies | \"Redis\", \"PostgreSQL\", \"Kubernetes\" |\n| Libraries | \"React\", \"PyTorch\", \"Express\" |\n| Decisions | \"Switch to PostgreSQL\", \"Use microservices\" |\n| Organizations | \"Platform Team\", \"Anthropic\" |\n\n---\n\n## 📊 Confidence & Smart Tracking\n\nNot all knowledge is equally reliable. Every page in your wiki carries metadata that helps you trust it:\n\n- **Confidence** — `high`, `medium`, or `low` based on how many sources support it\n- **Verified date** — when the content was last confirmed accurate\n- **Authority** — whether a source is `primary` (original research), `secondary` (review), or `commentary` (opinion)\n- **Typed relationships** — connections between concepts aren't just links, they have meaning: `extends`, `contradicts`, `supersedes`, `depends-on`\n\nWhen two sources disagree, the plugin flags the **contradiction** and links both sides — so you see the conflict instead of getting silently wrong answers.\n\nWhen newer information replaces older claims, the old page is marked as **superseded** with a link to what replaced it. Nothing is deleted — you can always trace the history.\n\n---\n\n## 🔮 Browse Your Wiki with Obsidian\n\nDuring setup, you'll be asked if you want [Obsidian](https://obsidian.md) — a free app for browsing your wiki visually. It's installed automatically if you say yes.\n\nOpen your project folder in Obsidian and you get:\n\n- 🕸️ **Graph View** — a visual map of all your pages and how they connect\n- 🔙 **Backlinks** — click any page to see everything that links to it\n- 🔍 **Search** — find anything across your entire wiki\n- ⚡ **Live sync** — new pages from `/wiki-ingest` appear in Obsidian right away\n\nThis is optional. Your wiki works perfectly with just the commands.\n\n---\n\n## 📁 What Gets Created\n\n```\nyour-project/\n  📄 raw/           Your original documents (never modified)\n  📚 wiki/          All generated pages (summaries, concepts, topics, insights)\n  .data/            Internal working data (hidden)\n```\n\nNo wrapper folder. `raw/` and `wiki/` live right in your project. If `raw/` already exists, setup just adds `wiki/` next to it.\n\nAll files are plain Markdown. Open them with any text editor, Obsidian, or VS Code.\n\n---\n\n## 💡 Tips\n\n- 🌱 **Start small.** Add 2–3 sources and try `/wiki` before adding more.\n- 🎯 **Be specific.** \"What does source X say about Y?\" works better than vague questions.\n- ✨ **Try `/wiki-insights`** after adding several sources on a topic — it finds patterns you might miss.\n- 🔒 **Your sources are safe.** The plugin never changes your original files.\n\n---\n\n## ⚙️ How Setup Works\n\n`/wiki-setup` handles everything automatically based on your project size:\n\n| Size | Search | What gets installed |\n|------|--------|-------------------|\n| **Small** | Built-in | Nothing extra needed |\n| **Medium** | Smart search ([qmd](https://github.com/tobi/qmd)) | Node.js + qmd (auto-installed) |\n| **Large** | Fastest search | Node.js + qmd + MCP server (auto-installed and configured) |\n\nThe plugin detects your operating system (macOS, Linux, Windows) and installs everything using the right method for your machine. If anything fails, it falls back to built-in search — you're never stuck.\n\n---\n\n## 📦 Other Ways to Install\n\n### From the Claude Code Marketplace (easiest)\n\n```\n/plugin marketplace add bradduy/wiki-knowledge-compiler\n```\nAfter adding, configure it in the marketplace:\n1. Run `/plugin` → select **Marketplace** tab\n2. Find **wiki-knowledge-compiler** and open it\n3. Click **Enable** and choose the scope — pick the project folder where you want your wiki to live\n4. Confirm — this grants the plugin permission to create files and run commands in that folder\n\n<p align=\"center\">\n  <img src=\"./assets/user_scope_setting.png\" alt=\"Enable scope setting\" width=\"100%\" />\n</p>\n\n### Clone as a standalone project\n\n```bash\ngit clone https://github.com/bradduy/wiki-knowledge-compiler.git\ncd wiki-knowledge-compiler\n```\n\nOpen in Claude Code and run `/wiki-setup`.\n\n### Add to an existing project\n\n```bash\ngit clone https://github.com/bradduy/wiki-knowledge-compiler.git /tmp/wkc\ncp -r /tmp/wkc/.claude/ your-project/.claude/\ncp -r /tmp/wkc/.claude-plugin/ your-project/.claude-plugin/\ncp -r /tmp/wkc/{agents,skills,templates} your-project/\ncp /tmp/wkc/CLAUDE.md your-project/CLAUDE.md\ncp /tmp/wkc/.data/wiki.config.md your-project/.data/wiki.config.md\nbash /tmp/wkc/scripts/init-kb.sh your-project/knowledge-base\nrm -rf /tmp/wkc\n```\n\nThen run `/wiki-setup`.\n\n---\n\n## 📌 Good to Know\n\n- 📡 **Works offline.** Everything is on your computer.\n- 📄 **Plain text files.** No database or special software needed to read your wiki.\n- 📑 **PDFs supported.** Complex layouts (tables, multi-column) may not extract perfectly.\n- 👤 **Designed for personal use.** Single user.\n- 🕐 **Want version history?** Run `git init` in your knowledge base folder.\n\n---\n\n## 🤝 Contributing\n\nWant to improve the plugin? See the [contributing guide](.github/CONTRIBUTING.md) or open an issue.\n\n## 📄 License\n\nMIT\n",
  "bytes": 16883,
  "sha": "6452607c447ec64e8bb8d05794bcb4888e231f4d18a19d01ba4eccc0eedcd165",
  "repo_slug": "bradduy/wiki-knowledge-compiler",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_bradduy_wiki_knowledge_compiler_wiki_kno_b76bffbf/readme"
}