[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-92454":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":18,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":22,"readmeContent":23,"aiSummary":24,"trendingCount":14,"starSnapshotCount":14,"syncStatus":25,"lastSyncTime":26,"discoverSource":27},92454,"infinite-brain-os","starmynd-org\u002Finfinite-brain-os","starmynd-org","A git-backed operating system for running a business with AI agents. Plain Markdown and YAML, readable by any file-reading agent, owned by you.",null,"Shell",141,49,1,0,6,45.7,"MIT License",false,"main",true,[],"2026-07-22 04:02:06","# Infinite Brain OS\n\nA git-backed operating system for running a business with AI agents. Plain Markdown and\nYAML, readable by any file-reading agent, owned by you.\n\nThe Infinite Brain is a knowledge OS: it makes what your business knows, decides, and does\nreliably retrievable and safely executable by AI agents, today and after the tools change.\nKnowledge lives in namespaces with an explicit promotion path to operator-approved canon.\nWork lives in projects with typed entities (commands, agents, skills, rules, workflows,\ntools). A contract layer (`_system\u002F`) plus a validator keeps the whole graph honest. Your\nAI coding agent (Claude Code, Codex, or any file-reading agent) is the runtime.\n\nNo database, no server, no vendor lock-in. If you can read this repo, so can your agents.\n\n## Quickstart\n\n```bash\ngit clone \u003Cthis-repo> my-brain && cd my-brain\nbash _system\u002Fvalidate.sh     # the contract holds: zero errors on a fresh clone\nclaude                       # or your agent of choice; CLAUDE.md orients it\n```\n\nThen say to your agent:\n\n```\nRead START-HERE.md and give me the tour.\n```\n\nOr open the folder as an Obsidian vault (config ships in `.obsidian\u002F`) and read\n`START-HERE.md` yourself. The full walkthrough is `docs\u002Fgetting-started.md`.\n\n## What ships in the box\n\n- **The contract layer.** `_system\u002F` holds the schemas, rules, registries, and\n  `validate.sh`: what must be true in this repo and how it is checked.\n- **The doctrine.** `knowledge\u002Fai-architecture\u002F` is the full reference architecture: the\n  control spine, the namespace model, canon versus synthesis, retrieval doctrine, surface\n  boundaries, and the agent-authority limits. It is the \"why\" behind every folder here.\n- **A complete worked example.** A fictional candle studio (Emberline) threads through one\n  example of every entity type, all cross-linked: a command, an agent, a skill, a rule, two\n  workflows (agentic and deterministic), a tool, a tiny knowledge namespace, a data pointer,\n  a memory, an output, a filled project, and an assembled department.\n- **Builders.** Skills and workflows that scaffold new namespaces, departments, projects,\n  agents, and knowledge bases, plus an onboarding interview (`docs\u002Fonboard-business.md`)\n  that maps your business onto the architecture.\n- **Eight profile references.** `knowledge\u002F_examples\u002F` shows the eight namespace profiles\n  (doctrine, data-system, design-system, content-strategy, tool-contract, and more) as\n  copyable scaffolds.\n\n## Folder map\n\n```\n_system\u002F           The operative contract: schemas, rules, registries, validate.sh\nknowledge\u002F         Namespace-first knowledge graph (the doctrine and your domains)\nentities\u002F          Canonical executable entities: commands, agents, skills, rules\n.claude\u002F .codex\u002F   Runtime adapter shims (regenerate with sync-adapters.sh)\nworkflows\u002F         Agentic reasoning pipelines\nautomations\u002Fn8n\u002F   Deterministic workflows (JSON plus a brain-record companion)\ntools\u002F             Pointer nodes over bounded capabilities\ndepartments\u002F       Assemblies over the entities: one folder per operating lane\nprojects\u002F          One PLAN.md per project, with inline tasks\nintake\u002F            Inbound flow: source captures, routing, processed receipts\ndata\u002F              Pointers to where numbers live (never the numbers)\nmemory\u002F            Reviewed learnings\noutputs\u002F           Produced artifacts with lineage\nsessions\u002F          The audit trail of AI work sessions\nswarms\u002F            Multi-agent sprint packages\ndocs\u002F              Setup, retrieval, and onboarding docs\n```\n\n## The example tour (fifteen minutes)\n\n1. `knowledge\u002Femberline-studio\u002Fcanon\u002Fbrand-essentials.md`: a canon node, the studio's\n   source of truth. Note the verification fields and the changelog.\n2. `entities\u002Frules\u002Fstudio-brand-voice.md`: a rule derived from that canon.\n3. `entities\u002Fskills\u002Fwrite-product-description.md`: a skill that applies the rule.\n4. `entities\u002Fagents\u002Fstudio-inbox-triage.md`: an agent that uses both and escalates what it\n   must not decide.\n5. `workflows\u002Fweekly-studio-review.md`: the weekly loop reading `data\u002Forders-ledger.md`\n   (a pointer, never live numbers) and `memory\u002Fphotograph-before-listing.md` (a lesson).\n6. `outputs\u002F2026-06-05-spring-collection-brief.md`: what the loop produced, with lineage.\n7. `projects\u002F_example\u002FPLAN.md`: the work container that ties it together.\n8. `departments\u002Fexample-studio-ops\u002FINDEX.md`: the assembly of all of the above.\n\nEvery file is under two minutes' reading. Follow the edges in the frontmatter; that is the\ngraph.\n\n## The rules that keep it honest\n\n- Every node-bearing file carries typed YAML frontmatter; `bash _system\u002Fvalidate.sh` must\n  exit 0.\n- Canon is operator-approved, always. Agents draft; you sign.\n- The repo never stores live numbers, live queues, or secrets: pointers only.\n- Sessions that touch the repo are registered, logged, and closed out in `sessions\u002F`.\n\nShipped doctrine occasionally refers to Paperclip, the task runtime of the deployment this\nstarter derives from: treat it as a placeholder for whatever runtime you adopt. Nothing\nhere requires it.\n\n## Make it yours\n\nRun the onboarding interview (`docs\u002Fonboard-business.md`), or go manual: build your first\nnamespace with `entities\u002Fskills\u002Fbuild-namespace.md`, assemble your first department from\n`departments\u002F_template\u002F`, fill in `knowledge\u002Fpersonal-operator\u002Fpillars\u002Foperator-profile.md`,\nand retire the candle studio when you no longer need the training wheels.\n\n## License\n\nMIT. See `LICENSE`. Contributions welcome: see `CONTRIBUTING.md`.\n","Infinite Brain OS 是一个基于 Git 的业务操作系统，用于协调 AI 代理执行企业级知识管理与自动化任务。它采用纯 Markdown 和 YAML 文件组织知识、命令、代理、规则、工作流等实体，通过 `_system\u002F` 合约层和验证脚本确保结构一致性，无需数据库或服务器，所有数据由用户完全自主掌控。核心特点是命名空间驱动的知识图谱、显式的知识晋升路径（至审核通过的权威版本）、可被任意文件读取型 AI 代理（如 Claude Code）直接解析执行。适用于中小团队构建可审计、可迁移、免厂商锁定的 AI 增强型业务运营体系。",2,"2026-07-09 02:30:03","CREATED_QUERY"]