[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-93263":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":13,"stars7d":14,"stars30d":14,"stars90d":13,"forks30d":13,"starsTrendScore":13,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":16,"archived":17,"fork":17,"defaultBranch":18,"hasWiki":19,"hasPages":17,"topics":20,"createdAt":9,"pushedAt":9,"updatedAt":21,"readmeContent":22,"aiSummary":23,"trendingCount":13,"starSnapshotCount":13,"syncStatus":14,"lastSyncTime":24,"discoverSource":25},93263,"ai-content-kb","mrbear1024\u002Fai-content-kb","mrbear1024","A review-first reference architecture for AI-assisted personal content knowledge systems",null,106,6,101,0,2,43.74,"MIT License",false,"main",true,[],"2026-07-22 04:02:08","\u003Cdiv align=\"center\">\n  \u003Ch1>AI Content Knowledge Base\u003C\u002Fh1>\n  \u003Cp>\u003Cstrong>A review-first, AI-native knowledge base for people who research, write, teach, and publish.\u003C\u002Fstrong>\u003C\u002Fp>\n  \u003Cp>Markdown + Codex + Obsidian + typed YAML relationships\u003C\u002Fp>\n  \u003Cp>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmrbear1024\u002Fai-content-kb\u002Fstargazers\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmrbear1024\u002Fai-content-kb?style=flat\" alt=\"GitHub stars\">\u003C\u002Fa>\n    \u003Ca href=\"LICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fmrbear1024\u002Fai-content-kb\" alt=\"MIT license\">\u003C\u002Fa>\n    \u003Ca href=\"AGENTS.md\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FCodex-AGENTS.md-111111\" alt=\"Codex AGENTS.md\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fobsidian.md\u002F\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FObsidian-compatible-7C3AED?logo=obsidian\" alt=\"Obsidian compatible\">\u003C\u002Fa>\n  \u003C\u002Fp>\n  \u003Cp>English · \u003Ca href=\"README.zh-CN.md\">简体中文\u003C\u002Fa>\u003C\u002Fp>\n\u003C\u002Fdiv>\n\n---\n\n> [!NOTE]\n> **This project was inspired by [Andrej Karpathy's LLM Wiki pattern](https:\u002F\u002Fgist.github.com\u002Fkarpathy\u002F442a6bf555914893e9891c11519de94f):** immutable raw sources, a persistent and compounding Markdown wiki maintained by an LLM, and a repository schema such as `AGENTS.md` that defines how the agent should ingest, query, and maintain knowledge.\n>\n> `ai-content-kb` extends that idea with explicit provenance roles, a first-class published-product layer, review-first staging, typed machine-readable relationships, content-creation workflows, and safe legacy-vault migration.\n\nMost knowledge bases mix personal thinking, external evidence, published work, and AI-generated text in the same pile. This template gives each a clear role—and gives AI agents rules for working without silently polluting originals or publication workflows.\n\n> **Originals stay trustworthy. The wiki stays readable. The graph stays queryable. AI output stays reviewable.**\n\n![Example of a mature Obsidian knowledge graph](docs\u002Fassets\u002Fobsidian-graph-view.png)\n\n> The screenshot shows a mature vault using this architecture. A fresh clone starts with a small synthetic example and grows as you add reviewed notes, links, and tags.\n\n## Why this project\n\n| Need | How this project handles it |\n|---|---|\n| Keep personal ideas distinct from external claims | Separate `raw\u002F` and `sources\u002F` provenance layers |\n| Reuse published articles, courses, and scripts | Treat reviewed output as a first-class `products\u002F` layer |\n| Let people browse without turning summaries into proof | Use `wiki\u002F` as a cited human interface |\n| Give agents precise relationships | Store typed, evidenced YAML sidecars in `.kb\u002Flinks\u002F` |\n| Prevent generated text from becoming truth by accident | Route AI output through `.kb\u002Fstaging\u002F` and human review |\n| Work naturally with Codex | Ship durable workflows in root [`AGENTS.md`](AGENTS.md) |\n| Keep Obsidian optional | Store everything as ordinary Markdown, YAML, JSON, and folders |\n\n## Quickstart\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fmrbear1024\u002Fai-content-kb.git\ncd ai-content-kb\n```\n\nNo application is required to read the vault. Use plain Markdown, open the root as an Obsidian vault, or use the intended Codex workflow below.\n\n### Open in Codex\n\n1. Open the Codex desktop app.\n2. Select `+` in **Projects**.\n3. Choose **Use an existing folder**.\n4. Select the cloned repository root.\n5. Start a new task and say:\n\n```text\nInspect the project structure, read the knowledge-base rules,\nmake no changes, and list the available workflows.\n```\n\nCodex reads repository `AGENTS.md` guidance before starting work. The root instructions then route it to the rest of this project's rules. [See the official Codex documentation](https:\u002F\u002Flearn.chatgpt.com\u002Fdocs\u002Fagent-configuration\u002Fagents-md).\n\n> The phrases below are natural-language workflows defined by this repository, not native slash commands. No plugin is required.\n\n## Talk to your knowledge base\n\n| Say this | Default result |\n|---|---|\n| `加入知识库：这是我的原创笔记` | Store owner-authored input under `raw\u002F` and create staged index candidates |\n| `add to knowledge base: this attachment is an external source` | Preserve provenance under `sources\u002F` and create staged candidates |\n| `增加 Wiki 索引：刚才的材料` | Draft cited wiki pages and typed relationships in staging |\n| `query knowledge base: \u003Cquestion>` | Navigate the wiki and graph, return to originals, and cite paths |\n| `compose from knowledge base: \u003Cbrief>` | Build a source plan, outline, and sourced staging draft |\n| `migrate legacy knowledge base: inventory \u003Cpath>` | Produce a read-only inventory and mapping proposal |\n| `review and publish index: \u003Cstaging path>` | Validate and promote accepted wiki and graph candidates |\n| `lint knowledge base` | Report broken paths, citations, aliases, hash changes, and privacy risks |\n\n## How it works\n\n```mermaid\nflowchart LR\n    R[\"raw\u002F\\nOwner input\"]\n    S[\"sources\u002F\\nExternal evidence\"]\n    P[\"products\u002F\\nReviewed output\"]\n    T[\".kb\u002Fstaging\u002F\\nAI candidates\"]\n    W[\"wiki\u002F\\nHuman interface\"]\n    G[\".kb\u002Flinks\u002F\\nMachine graph\"]\n\n    R -->|ingest| T\n    S -->|ingest| T\n    P -->|backfill| T\n    T -->|human review| W\n    T -->|human review| G\n    W -->|compose| P\n    G -->|retrieve evidence| P\n```\n\n| Path | Role | Source of truth? |\n|---|---|---:|\n| `raw\u002F` | Original notes, judgments, voice transcripts, owned research, active drafts | Yes—owner intent |\n| `sources\u002F` | External clips, papers, books, reports, and media transcripts | Yes—external claims |\n| `products\u002F` | Reviewed articles, courses, scripts, and delivered work | Yes—published expression |\n| `wiki\u002F` | Concepts, entities, maps, and high-value source notes | No—must cite originals |\n| `.kb\u002Flinks\u002F` | Reviewed typed relationships with evidence and hashes | Rebuildable |\n| `.kb\u002Fstaging\u002F` | Unreviewed AI prose, wiki pages, mappings, and graph candidates | No |\n\nRead the [architecture rationale](docs\u002FARCHITECTURE.md) and [graph schema](docs\u002FGRAPH_SCHEMA.md) for details.\n\n## Core workflows\n\n### Capture and index\n\n```text\nadd to knowledge base: this attachment is an external source.\nPreserve provenance, check existing aliases, and keep generated wiki and graph data in staging.\n```\n\nReview `.kb\u002Fstaging\u002Fwiki\u002F` and `.kb\u002Fstaging\u002Flinks\u002F` before promotion.\n\n### Create with AI\n\n```text\nCompose from the knowledge base:\nWrite an article for AI product managers about Context Engineering.\nSeparate my judgments from external claims, cite repository paths,\nand create a source plan and outline before drafting.\n```\n\nUnreviewed prose goes to `.kb\u002Fstaging\u002Fdrafts\u002F`, accepted work moves to `raw\u002Fdrafts\u002F` for development, and only publication-ready work belongs in `products\u002F`.\n\n[Read the content creation guide](docs\u002FCONTENT_CREATION_GUIDE.md).\n\n### Migrate an existing vault\n\n```text\nMigrate legacy knowledge base: inventory the provided directory in read-only mode.\nDo not move, copy, delete, or modify files.\nPropose a 20–50 file pilot and stop for review.\n```\n\nMigration follows backup → inventory → mapping review → pilot → validation → reviewed batches → cutover.\n\n[Read the migration guide](docs\u002FMIGRATION_GUIDE.md).\n\n### Explore in Obsidian\n\nOpen the repository root with **Open folder as vault**, then select **Graph view** in the left ribbon. The included `.obsidian\u002Fgraph.json` colors nodes by `raw`, `sources`, `products`, and `wiki`.\n\nObsidian visualizes Markdown links and tags. Typed `.kb\u002Flinks\u002F*.yaml` relationships are a separate machine layer and do not automatically become Obsidian lines.\n\n## Review-first lifecycle\n\n| State | Destination |\n|---|---|\n| Unreviewed article or script | `.kb\u002Fstaging\u002Fdrafts\u002F` |\n| Unreviewed course outline | `.kb\u002Fstaging\u002Fcourse-drafts\u002F` |\n| Unreviewed concept, entity, map, or source note | `.kb\u002Fstaging\u002Fwiki\u002F` |\n| Unreviewed relationship sidecar | `.kb\u002Fstaging\u002Flinks\u002F` |\n| Draft selected for human development | `raw\u002Fdrafts\u002F` |\n| Reviewed knowledge and relationships | `wiki\u002F` and `.kb\u002Flinks\u002F` |\n| Reviewed, publication-ready output | `products\u002F` |\n\nThe deciding factors are **review status, intended use, human ownership, and publication readiness**—not whether AI helped write the text.\n\n## Documentation\n\n| Document | Purpose |\n|---|---|\n| [Start here](START_HERE.md) | Reading order and core boundaries |\n| [Agent rules](AGENTS.md) | Durable Codex workflows and safety rules |\n| [Knowledge base guide](KNOWLEDGE_BASE_GUIDE.md) | Human-facing operating principles |\n| [Architecture](docs\u002FARCHITECTURE.md) | Layer model, lifecycle, and scaling path |\n| [Graph schema](docs\u002FGRAPH_SCHEMA.md) | YAML sidecar fields and edge vocabulary |\n| [Content creation](docs\u002FCONTENT_CREATION_GUIDE.md) | Research, outlining, drafting, review, and backfill |\n| [Migration](docs\u002FMIGRATION_GUIDE.md) | Safe migration from an existing vault |\n| [Public release checklist](docs\u002FPUBLIC_RELEASE_CHECKLIST.md) | Privacy, secrets, rights, and release checks |\n\n## Project scope\n\nThis repository is a **reference architecture and working template**, not a hosted knowledge-management application. Automation is intentionally minimal: validate the content roles and review workflow with files first, then add manifests, search, embeddings, or a graph database only when real queries justify them.\n\nObsidian and Codex are optional interfaces. The durable system is the repository itself.\n\n## Contributing and support\n\n- Found a problem or have a proposal? [Open an issue](https:\u002F\u002Fgithub.com\u002Fmrbear1024\u002Fai-content-kb\u002Fissues).\n- Pull requests improving the information model, examples, schemas, lint rules, and documentation are welcome.\n- Do not submit private notes, copyrighted captures, credentials, absolute home paths, or generated databases containing personal metadata.\n- Before publishing a fork, use the [public release checklist](docs\u002FPUBLIC_RELEASE_CHECKLIST.md).\n\nMaintained by [mrbear1024](https:\u002F\u002Fgithub.com\u002Fmrbear1024) and contributors.\n\n## License\n\n[MIT](LICENSE). This project is not affiliated with or endorsed by Obsidian or OpenAI.\n","这是一个面向内容创作者的知识管理参考架构，专为研究、写作、教学与出版场景设计，强调‘审阅优先’原则。项目通过分层结构（如raw\u002F、sources\u002F、wiki\u002F、products\u002F）隔离原始材料、外部引用、AI生成内容与最终出版物，并利用Markdown、YAML关系描述、Obsidian兼容格式及Codex工作流实现可追溯、可审查、可查询的知识系统。核心特点是支持AI代理按明确定义的规则处理内容，同时保障原始资料可信性与发布流程安全性。适用于学者、作家、讲师等需长期积累并复用知识资产的专业用户。","2026-07-15 02:30:03","CREATED_QUERY"]