[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74880":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":23,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":32,"readmeContent":33,"aiSummary":34,"trendingCount":16,"starSnapshotCount":16,"syncStatus":35,"lastSyncTime":36,"discoverSource":37},74880,"arscontexta","agenticnotetaking\u002Farscontexta","agenticnotetaking","Claude Code plugin that generates individualized knowledge systems from conversation. You describe how you think and work, have a conversation and get a complete second brain as markdown files you own.","https:\u002F\u002Farscontexta.org",null,"Shell",3395,218,31,20,0,4,25,71,12,29.02,"MIT License",false,"main",[26,27,28,29,30,31],"claude-code","claude-code-plugin","knowledge-base","knowledge-management","markdown","second-brain","2026-06-12 02:03:29","# Ars Contexta\n\n**A second brain for your agent.**\n\nA Claude Code plugin that generates complete knowledge systems from conversation.\nYou describe how you think and work. The engine derives a cognitive architecture\n-- folder structure, context files, processing pipeline, hooks, navigation maps,\nand note templates -- tailored to your domain and backed by 249 research claims.\n\nNo templates. No configuration. Just conversation.\n\n**v0.8.0** · Claude Code plugin · MIT\n\n---\n\n## Installation\n\n1. Add the marketplace to Claude Code:\n   ```\n   \u002Fplugin marketplace add agenticnotetaking\u002Farscontexta\n   ```\n\n2. Install the plugin:\n   ```\n   \u002Fplugin install arscontexta@agenticnotetaking\n   ```\n\n3. Restart Claude Code, then run:\n   ```\n   \u002Farscontexta:setup\n   ```\n\n4. Answer 2-4 questions about your domain (~20 minutes -- token-intensive but one-time)\n\n5. The engine generates your complete knowledge system\n\n6. Restart Claude Code again to activate generated hooks and skills\n\n7. Run `\u002Farscontexta:help` to see everything available\n\n---\n\n## What It Does\n\nMost AI tools start every session blank. Ars Contexta changes that by generating\na persistent thinking system derived from how you actually work.\n\n**What you get:**\n\n- **A vault** -- plain markdown files connected by wiki links, forming a traversable\n  knowledge graph. No database, no cloud, no lock-in.\n- **A processing pipeline** -- skills that extract insights, find connections, update\n  old notes with new context, and verify quality.\n- **Automation** -- hooks that enforce structure on every write, detect maintenance\n  needs, capture session state, and auto-commit.\n- **Navigation** -- Maps of Content (MOCs) at hub, domain, and topic levels.\n- **Templates** -- note templates with `_schema` blocks as single source of truth.\n- **A user manual** -- 7 pages of domain-native documentation generated alongside.\n\n**The key differentiator:** derivation, not templating. Every choice traces to\nspecific research claims. The engine reasons from principles about what your\ndomain needs and why.\n\n---\n\n## The Setup Flow\n\n`\u002Farscontexta:setup` runs a 6-phase process:\n\n| Phase | What Happens |\n|-------|-------------|\n| **Detection** | Detects Claude Code environment and capabilities |\n| **Understanding** | 2-4 conversation turns where you describe your domain |\n| **Derivation** | Maps signals to eight configuration dimensions with confidence scoring |\n| **Proposal** | Shows what will be generated and why, in your vocabulary |\n| **Generation** | Produces all files: context file, folders, templates, skills, hooks, manual |\n| **Validation** | Checks all 15 kernel primitives, runs pipeline smoke test |\n\nThe whole process takes about 20 minutes. It's token-intensive because the engine\nreads research claims, reasons about your domain, and generates substantial output.\nThis is a one-time investment -- after setup, your agent remembers.\n\nFor advanced users: `\u002Farscontexta:setup --advanced` to configure dimensions directly.\n\n---\n\n## Three-Space Architecture\n\nEvery generated system separates content into three spaces:\n\n| Space | Purpose | Growth |\n|-------|---------|--------|\n| **self\u002F** | Agent persistent mind -- identity, methodology, goals | Slow (tens of files) |\n| **notes\u002F** | Knowledge graph -- the reason the system exists | Steady (10-50\u002Fweek) |\n| **ops\u002F** | Operational coordination -- queue state, sessions | Fluctuating |\n\nNames adapt to your domain (`notes\u002F` might become `reflections\u002F`, `claims\u002F`,\nor `decisions\u002F`), but the separation is invariant.\n\n---\n\n## Commands\n\n### Plugin-Level (always available)\n\n| Command | What It Does |\n|---------|-------------|\n| `\u002Farscontexta:setup` | Conversational onboarding -- generates your full system |\n| `\u002Farscontexta:help` | Contextual guidance and command discovery |\n| `\u002Farscontexta:tutorial` | Interactive walkthrough (learn by doing) |\n| `\u002Farscontexta:ask` | Query the research graph for methodology answers |\n| `\u002Farscontexta:health` | Run diagnostic checks on your vault |\n| `\u002Farscontexta:recommend` | Get architecture advice for your use case |\n| `\u002Farscontexta:architect` | Research-backed evolution guidance |\n| `\u002Farscontexta:add-domain` | Add a new knowledge domain to an existing system |\n| `\u002Farscontexta:reseed` | Re-derive from first principles when drift accumulates |\n| `\u002Farscontexta:upgrade` | Apply plugin knowledge base updates to your system |\n\n### Generated (available after setup)\n\n| Command | What It Does |\n|---------|-------------|\n| `\u002Freduce` | Extract insights from sources |\n| `\u002Freflect` | Find connections, update MOCs |\n| `\u002Freweave` | Update older notes with new connections |\n| `\u002Fverify` | Combined quality check: description + schema + health |\n| `\u002Fvalidate` | Schema compliance checking |\n| `\u002Fseed` | Create extraction task with duplicate detection |\n| `\u002Fralph` | Queue-based orchestration with fresh context per phase |\n| `\u002Fpipeline` | End-to-end source processing |\n| `\u002Ftasks` | Queue management |\n| `\u002Fstats` | Vault metrics |\n| `\u002Fgraph` | Graph analysis |\n| `\u002Fnext` | Next-action recommendation |\n| `\u002Flearn` | Research and grow |\n| `\u002Fremember` | Mine session learnings |\n| `\u002Frethink` | Challenge system assumptions |\n| `\u002Frefactor` | Structural improvements |\n\n---\n\n## Processing Pipeline\n\nThe vault implements the **6 Rs**, extending Cornell Note-Taking's 5 Rs with a\nmeta-cognitive layer:\n\n| Phase | What Happens | Command |\n|-------|-------------|---------|\n| **Record** | Zero-friction capture into inbox\u002F | Manual |\n| **Reduce** | Extract insights with domain-native categories | `\u002Freduce` |\n| **Reflect** | Find connections, update MOCs | `\u002Freflect` |\n| **Reweave** | Update older notes with new context | `\u002Freweave` |\n| **Verify** | Description + schema + health checks | `\u002Fverify` |\n| **Rethink** | Challenge system assumptions | `\u002Frethink` |\n\n### Fresh Context Per Phase\n\nEach phase runs in its own context window via subagent spawning. LLM attention\ndegrades as context fills. By spawning a fresh subagent per phase, every phase\noperates in the \"smart zone.\"\n\n```\n\u002Fralph 5\n  |-- Read queue, find next unblocked task\n  |-- Spawn subagent (fresh context)\n  |   +-- Runs skill, updates task file, returns handoff\n  |-- Parse handoff, capture learnings\n  |-- Advance phase in queue\n  +-- Repeat for 5 tasks\n```\n\n---\n\n## Hooks\n\nFour hooks automate quality enforcement:\n\n| Hook | Event | What It Does |\n|------|-------|-------------|\n| **Session Orient** | `SessionStart` | Injects workspace tree, loads identity, surfaces maintenance signals |\n| **Write Validate** | `PostToolUse` (Write) | Schema enforcement on every note write |\n| **Auto Commit** | `PostToolUse` (Write, async) | Git auto-commit, non-blocking |\n| **Session Capture** | `Stop` | Persists session state to `ops\u002Fsessions\u002F` |\n\n---\n\n## The Research Graph\n\nThe `methodology\u002F` directory contains **249 interconnected research claims**\nabout tools for thought, knowledge management, and agent-native cognitive\narchitecture. These claims back every configuration decision.\n\n### Synthesizes\n\nZettelkasten -- Cornell Note-Taking -- Evergreen Notes -- PARA -- GTD -- Memory\nPalaces -- Cognitive Science (extended mind, spreading activation, generation\neffect) -- Network Theory (small-world topology, betweenness centrality) --\nAgent Architecture (context windows, session boundaries, multi-agent patterns)\n\n### How Claims Back Decisions\n\nEvery kernel primitive includes `cognitive_grounding` linking to specific research:\n\n- **MOC hierarchy** -- context-switching cost research (Leroy 2009)\n- **Description field** -- progressive disclosure principles\n- **Wiki links** -- spreading activation theory\n\nQuery directly: `\u002Farscontexta:ask \"Why does my system use atomic notes?\"`\n\n---\n\n## Semantic Search (optional)\n\n[qmd](https:\u002F\u002Fgithub.com\u002Ftobi\u002Fqmd) adds concept matching across vocabularies.\nNot required -- the system works fully with ripgrep + MOC traversal.\n\n`\u002Fsetup` should perform this configuration automatically when semantic search is active.\nThe commands below are manual fallback\u002Fsetup verification.\n\n```bash\n# Install qmd\nnpm install -g @tobilu\u002Fqmd\n# or\nbun install -g @tobilu\u002Fqmd\n\ncd your-vault\u002F\nqmd init\nqmd collection add . --name \u003Cnotes_directory_name> --mask \"\u003Cnotes_directory_name>\u002F**\u002F*.md\"\nqmd embed\n```\n\nCreate or merge `.mcp.json` in the vault root:\n\n```json\n{\n  \"mcpServers\": {\n    \"qmd\": {\n      \"command\": \"qmd\",\n      \"args\": [\"mcp\"],\n      \"autoapprove\": [\n        \"mcp__qmd__search\",\n        \"mcp__qmd__vector_search\",\n        \"mcp__qmd__deep_search\",\n        \"mcp__qmd__get\",\n        \"mcp__qmd__multi_get\",\n        \"mcp__qmd__status\"\n      ]\n    }\n  }\n}\n```\n\nKeep qmd MCP configuration and tool preapproval in `.mcp.json`.\n\n---\n\n## Prerequisites\n\n| Dependency | Required | Purpose |\n|-----------|----------|---------|\n| [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) v1.0.33+ | Yes | Plugin host |\n| `tree` | Yes | Workspace structure injection |\n| `ripgrep` (`rg`) | Yes | YAML queries, schema validation |\n| [qmd](https:\u002F\u002Fgithub.com\u002Ftobi\u002Fqmd) | Optional | Semantic search |\n\n---\n\n## Project Structure\n\n```\narscontexta\u002F\n|-- .claude-plugin\u002F\n|   |-- plugin.json              # Plugin manifest\n|   +-- marketplace.json         # Marketplace listing\n|-- skills\u002F                      # 10 plugin-level commands\n|   |-- setup\u002F                   # Conversational onboarding\n|   |-- help\u002F                    # Contextual guidance\n|   |-- tutorial\u002F                # Interactive walkthrough\n|   |-- ask\u002F                     # Query the research graph\n|   |-- health\u002F                  # Diagnostic checks\n|   |-- recommend\u002F               # Architecture advice\n|   |-- architect\u002F               # Evolution guidance\n|   |-- reseed\u002F                  # Re-derive from first principles\n|   |-- upgrade\u002F                 # Apply knowledge base updates\n|   +-- add-domain\u002F              # Multi-domain extension\n|-- skill-sources\u002F               # 16 generated command templates\n|   |-- reduce\u002F                  # Extract insights\n|   |-- reflect\u002F                 # Find connections\n|   |-- reweave\u002F                 # Backward pass\n|   |-- verify\u002F                  # Combined quality check\n|   +-- ...                      # 12 more processing commands\n|-- agents\u002F\n|   +-- knowledge-guide.md       # Pipeline subagent\n|-- hooks\u002F\n|   |-- hooks.json               # Hook configuration\n|   +-- scripts\u002F                 # Hook implementations\n|-- generators\u002F\n|   |-- claude-md.md             # CLAUDE.md template\n|   +-- features\u002F                # 17 composable feature blocks\n|-- methodology\u002F                 # 249 research claims\n|-- reference\u002F                   # Core reference documents\n|   |-- kernel.yaml              # 15 kernel primitives\n|   |-- three-spaces.md          # Architecture spec\n|   +-- use-case-presets.md      # Pre-validated configs\n|-- platforms\u002F                   # Platform-specific adapters\n|   |-- claude-code\u002F\n|   +-- shared\u002F\n|-- presets\u002F                     # Pre-validated configurations\n|-- scripts\u002F                     # Utility scripts\n+-- README.md\n```\n\n---\n\n## Development\n\nClone this repo and add the marketplace to Claude Code:\n\n```\n\u002Fplugin marketplace add ~\u002Fpath-to-arscontexta\n```\n\nInstall the plugin:\n\n```\n\u002Fplugin install arscontexta@agenticnotetaking\n```\n\nEvery time you make changes, re-install the plugin:\n\n```\n\u002Fplugin uninstall arscontexta@agenticnotetaking\n\u002Fplugin install arscontexta@agenticnotetaking\n```\n\n### Key Files for Contributors\n\n- `reference\u002Fkernel.yaml` -- 15 primitives every system must include\n- `generators\u002Ffeatures\u002F*.md` -- composable feature blocks\n- `skill-sources\u002F*\u002FSKILL.md` -- generated command templates\n- `skills\u002Fsetup\u002FSKILL.md` -- the derivation engine\n- `reference\u002Fuse-case-presets.md` -- preset definitions\n\n---\n\n## Presets\n\nThree pre-validated configurations for common use cases:\n\n| Preset | For | What You Get |\n|--------|-----|-------------|\n| **Research** | Academic work, literature reviews, synthesis | Atomic claims, citation tracking, methodology MOCs |\n| **Personal** | Life management, journaling, relationships | Reflective notes, goal tracking, relationship MOCs |\n| **Experimental** | Testing, iteration, rapid prototyping | Lightweight structure, fast capture, minimal ceremony |\n\nPresets provide starting defaults. The derivation engine adapts from there based\non your conversation.\n\n---\n\n## Roadmap\n\n| Feature | Status |\n|---------|--------|\n| Claude Code plugin | Available |\n| Marketplace listing | Available |\n| Multi-agent processing | In progress |\n\n---\n\n## Philosophy\n\nThe name connects to a tradition. **Ars Combinatoria**, **Ars Memoria**,\n**Ars Contexta**: the art of context.\n\nLlull's rotating wheels generated truth through combination. Bruno's memory wheels\ncreated millions of image combinations. They were external thinking systems -- tools\nto think with rather than just store in. The missing piece: they required a human\nmind to do the traversing. Now LLMs can traverse. The wheels can spin again.\n\nBuilt on [Tools for Thought for Agents](https:\u002F\u002Fgithub.com\u002Fagenticnotetaking) research.\n\n---\n\n## License\n\nMIT\n","Ars Contexta 是一个基于 Claude Code 的插件，能够通过对话生成个性化的知识管理系统。其核心功能包括根据用户的工作方式和思考模式自动生成认知架构、处理流程、自动化钩子以及导航地图等，并以 Markdown 文件形式提供给用户，构建起一个无数据库依赖的知识图谱。项目采用 Shell 语言编写，强调从原理出发为用户的特定领域定制知识系统而非简单套用模板。适用于需要深度整合个人工作流与知识管理的场景，如科研人员、写作者或任何希望提升信息组织效率的专业人士。",2,"2026-06-11 03:51:15","high_star"]