Skip to content
Back to search

agenticnotetaking

MarketplaceWorks with Claude CodeGitHub

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.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/agenticnotetaking-arscontexta/badge.svg)](/listings/agenticnotetaking-arscontexta)

HTML

<a href="/listings/agenticnotetaking-arscontexta"><img src="/api/listings/agenticnotetaking-arscontexta/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

1. Add the marketplace to Claude Code:

   /plugin marketplace add agenticnotetaking/arscontexta

2. Install the plugin:

   /plugin install arscontexta@agenticnotetaking

3. Restart Claude Code, then run:

   /arscontexta:setup

4. Answer 2-4 questions about your domain (~20 minutes -- token-intensive but one-time)

5. The engine generates your complete knowledge system

6. Restart Claude Code again to activate generated hooks and skills

7. Run `/arscontexta:help` to see everything available

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 37 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  3. "name": "agenticnotetaking",
  4. "owner": {
  5. "name": "Heinrich",
  6. "email": "arscontexta@proton.me"
  7. },
  8. "metadata": {
  9. "description": "Conversational derivation engine for agent-native knowledge systems",
  10. "version": "0.8.0"
  11. },
  12. "plugins": [
  13. {
  14. "name": "arscontexta",
  15. "description": "Generate research-backed knowledge systems from natural conversation. 15 kernel primitives, 26 commands, 249 research claims, 3 presets.",
  16. "version": "0.8.0",
  17. "source": "./",
  18. "strict": false,
  19. "author": {
  20. "name": "Heinrich",
  21. "email": "arscontexta@proton.me"
  22. },
  23. "homepage": "https://github.com/agenticnotetaking/arscontexta",
  24. "repository": "https://github.com/agenticnotetaking/arscontexta",
  25. "license": "MIT",
  26. "keywords": [
  27. "knowledge-management",
  28. "tools-for-thought",
  29. "agent-memory",
  30. "derivation-engine",
  31. "zettelkasten",
  32. "cognitive-architecture"
  33. ],
  34. "category": "productivity"
  35. }
  36. ]
  37. }
README.md · README · 387 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Ars Contexta
  2. **A second brain for your agent.**
  3. A Claude Code plugin that generates complete knowledge systems from conversation.
  4. You describe how you think and work. The engine derives a cognitive architecture
  5. -- folder structure, context files, processing pipeline, hooks, navigation maps,
  6. and note templates -- tailored to your domain and backed by 249 research claims.
  7. No templates. No configuration. Just conversation.
  8. **v0.8.0** · Claude Code plugin · MIT
  9. ---
  10. ## Installation
  11. 1. Add the marketplace to Claude Code:
  12. ```
  13. /plugin marketplace add agenticnotetaking/arscontexta
  14. ```
  15. 2. Install the plugin:
  16. ```
  17. /plugin install arscontexta@agenticnotetaking
  18. ```
  19. 3. Restart Claude Code, then run:
  20. ```
  21. /arscontexta:setup
  22. ```
  23. 4. Answer 2-4 questions about your domain (~20 minutes -- token-intensive but one-time)
  24. 5. The engine generates your complete knowledge system
  25. 6. Restart Claude Code again to activate generated hooks and skills
  26. 7. Run `/arscontexta:help` to see everything available
  27. ---
  28. ## What It Does
  29. Most AI tools start every session blank. Ars Contexta changes that by generating
  30. a persistent thinking system derived from how you actually work.
  31. **What you get:**
  32. - **A vault** -- plain markdown files connected by wiki links, forming a traversable
  33. knowledge graph. No database, no cloud, no lock-in.
  34. - **A processing pipeline** -- skills that extract insights, find connections, update
  35. old notes with new context, and verify quality.
  36. - **Automation** -- hooks that enforce structure on every write, detect maintenance
  37. needs, capture session state, and auto-commit.
  38. - **Navigation** -- Maps of Content (MOCs) at hub, domain, and topic levels.
  39. - **Templates** -- note templates with `_schema` blocks as single source of truth.
  40. - **A user manual** -- 7 pages of domain-native documentation generated alongside.
  41. **The key differentiator:** derivation, not templating. Every choice traces to
  42. specific research claims. The engine reasons from principles about what your
  43. domain needs and why.
  44. ---
  45. ## The Setup Flow
  46. `/arscontexta:setup` runs a 6-phase process:
  47. | Phase | What Happens |
  48. |-------|-------------|
  49. | **Detection** | Detects Claude Code environment and capabilities |
  50. | **Understanding** | 2-4 conversation turns where you describe your domain |
  51. | **Derivation** | Maps signals to eight configuration dimensions with confidence scoring |
  52. | **Proposal** | Shows what will be generated and why, in your vocabulary |
  53. | **Generation** | Produces all files: context file, folders, templates, skills, hooks, manual |
  54. | **Validation** | Checks all 15 kernel primitives, runs pipeline smoke test |
  55. The whole process takes about 20 minutes. It's token-intensive because the engine
  56. reads research claims, reasons about your domain, and generates substantial output.
  57. This is a one-time investment -- after setup, your agent remembers.
  58. For advanced users: `/arscontexta:setup --advanced` to configure dimensions directly.
  59. ---
  60. ## Three-Space Architecture
  61. Every generated system separates content into three spaces:
  62. | Space | Purpose | Growth |
  63. |-------|---------|--------|
  64. | **self/** | Agent persistent mind -- identity, methodology, goals | Slow (tens of files) |
  65. | **notes/** | Knowledge graph -- the reason the system exists | Steady (10-50/week) |
  66. | **ops/** | Operational coordination -- queue state, sessions | Fluctuating |
  67. Names adapt to your domain (`notes/` might become `reflections/`, `claims/`,
  68. or `decisions/`), but the separation is invariant.
  69. ---
  70. ## Commands
  71. ### Plugin-Level (always available)
  72. | Command | What It Does |
  73. |---------|-------------|
  74. | `/arscontexta:setup` | Conversational onboarding -- generates your full system |
  75. | `/arscontexta:help` | Contextual guidance and command discovery |
  76. | `/arscontexta:tutorial` | Interactive walkthrough (learn by doing) |
  77. | `/arscontexta:ask` | Query the research graph for methodology answers |
  78. | `/arscontexta:health` | Run diagnostic checks on your vault |
  79. | `/arscontexta:recommend` | Get architecture advice for your use case |
  80. | `/arscontexta:architect` | Research-backed evolution guidance |
  81. | `/arscontexta:add-domain` | Add a new knowledge domain to an existing system |
  82. | `/arscontexta:reseed` | Re-derive from first principles when drift accumulates |
  83. | `/arscontexta:upgrade` | Apply plugin knowledge base updates to your system |
  84. ### Generated (available after setup)
  85. | Command | What It Does |
  86. |---------|-------------|
  87. | `/reduce` | Extract insights from sources |
  88. | `/reflect` | Find connections, update MOCs |
  89. | `/reweave` | Update older notes with new connections |
  90. | `/verify` | Combined quality check: description + schema + health |
  91. | `/validate` | Schema compliance checking |
  92. | `/seed` | Create extraction task with duplicate detection |
  93. | `/ralph` | Queue-based orchestration with fresh context per phase |
  94. | `/pipeline` | End-to-end source processing |
  95. | `/tasks` | Queue management |
  96. | `/stats` | Vault metrics |
  97. | `/graph` | Graph analysis |
  98. | `/next` | Next-action recommendation |
  99. | `/learn` | Research and grow |
  100. | `/remember` | Mine session learnings |
  101. | `/rethink` | Challenge system assumptions |
  102. | `/refactor` | Structural improvements |
  103. ---
  104. ## Processing Pipeline
  105. The vault implements the **6 Rs**, extending Cornell Note-Taking's 5 Rs with a
  106. meta-cognitive layer:
  107. | Phase | What Happens | Command |
  108. |-------|-------------|---------|
  109. | **Record** | Zero-friction capture into inbox/ | Manual |
  110. | **Reduce** | Extract insights with domain-native categories | `/reduce` |
  111. | **Reflect** | Find connections, update MOCs | `/reflect` |
  112. | **Reweave** | Update older notes with new context | `/reweave` |
  113. | **Verify** | Description + schema + health checks | `/verify` |
  114. | **Rethink** | Challenge system assumptions | `/rethink` |
  115. ### Fresh Context Per Phase
  116. Each phase runs in its own context window via subagent spawning. LLM attention
  117. degrades as context fills. By spawning a fresh subagent per phase, every phase
  118. operates in the "smart zone."
  119. ```
  120. /ralph 5
  121. |-- Read queue, find next unblocked task
  122. |-- Spawn subagent (fresh context)
  123. | +-- Runs skill, updates task file, returns handoff
  124. |-- Parse handoff, capture learnings
  125. |-- Advance phase in queue
  126. +-- Repeat for 5 tasks
  127. ```
  128. ---
  129. ## Hooks
  130. Four hooks automate quality enforcement:
  131. | Hook | Event | What It Does |
  132. |------|-------|-------------|
  133. | **Session Orient** | `SessionStart` | Injects workspace tree, loads identity, surfaces maintenance signals |
  134. | **Write Validate** | `PostToolUse` (Write) | Schema enforcement on every note write |
  135. | **Auto Commit** | `PostToolUse` (Write, async) | Git auto-commit, non-blocking |
  136. | **Session Capture** | `Stop` | Persists session state to `ops/sessions/` |
  137. ---
  138. ## The Research Graph
  139. The `methodology/` directory contains **249 interconnected research claims**
  140. about tools for thought, knowledge management, and agent-native cognitive
  141. architecture. These claims back every configuration decision.
  142. ### Synthesizes
  143. Zettelkasten -- Cornell Note-Taking -- Evergreen Notes -- PARA -- GTD -- Memory
  144. Palaces -- Cognitive Science (extended mind, spreading activation, generation
  145. effect) -- Network Theory (small-world topology, betweenness centrality) --
  146. Agent Architecture (context windows, session boundaries, multi-agent patterns)
  147. ### How Claims Back Decisions
  148. Every kernel primitive includes `cognitive_grounding` linking to specific research:
  149. - **MOC hierarchy** -- context-switching cost research (Leroy 2009)
  150. - **Description field** -- progressive disclosure principles
  151. - **Wiki links** -- spreading activation theory
  152. Query directly: `/arscontexta:ask "Why does my system use atomic notes?"`
  153. ---
  154. ## Semantic Search (optional)
  155. [qmd](https://github.com/tobi/qmd) adds concept matching across vocabularies.
  156. Not required -- the system works fully with ripgrep + MOC traversal.
  157. `/setup` should perform this configuration automatically when semantic search is active.
  158. The commands below are manual fallback/setup verification.
  159. ```bash
  160. # Install qmd
  161. npm install -g @tobilu/qmd
  162. # or
  163. bun install -g @tobilu/qmd
  164. cd your-vault/
  165. qmd init
  166. qmd collection add . --name <notes_directory_name> --mask "<notes_directory_name>/**/*.md"
  167. qmd embed
  168. ```
  169. Create or merge `.mcp.json` in the vault root:
  170. ```json
  171. {
  172. "mcpServers": {
  173. "qmd": {
  174. "command": "qmd",
  175. "args": ["mcp"],
  176. "autoapprove": [
  177. "mcp__qmd__search",
  178. "mcp__qmd__vector_search",
  179. "mcp__qmd__deep_search",
  180. "mcp__qmd__get",
  181. "mcp__qmd__multi_get",
  182. "mcp__qmd__status"
  183. ]
  184. }
  185. }
  186. }
  187. ```
  188. Keep qmd MCP configuration and tool preapproval in `.mcp.json`.
  189. ---
  190. ## Prerequisites
  191. | Dependency | Required | Purpose |
  192. |-----------|----------|---------|
  193. | [Claude Code](https://docs.anthropic.com/en/docs/claude-code) v1.0.33+ | Yes | Plugin host |
  194. | `tree` | Yes | Workspace structure injection |
  195. | `ripgrep` (`rg`) | Yes | YAML queries, schema validation |
  196. | [qmd](https://github.com/tobi/qmd) | Optional | Semantic search |
  197. ---
  198. ## Project Structure
  199. ```
  200. arscontexta/
  201. |-- .claude-plugin/
  202. | |-- plugin.json # Plugin manifest
  203. | +-- marketplace.json # Marketplace listing
  204. |-- skills/ # 10 plugin-level commands
  205. | |-- setup/ # Conversational onboarding
  206. | |-- help/ # Contextual guidance
  207. | |-- tutorial/ # Interactive walkthrough
  208. | |-- ask/ # Query the research graph
  209. | |-- health/ # Diagnostic checks
  210. | |-- recommend/ # Architecture advice
  211. | |-- architect/ # Evolution guidance
  212. | |-- reseed/ # Re-derive from first principles
  213. | |-- upgrade/ # Apply knowledge base updates
  214. | +-- add-domain/ # Multi-domain extension
  215. |-- skill-sources/ # 16 generated command templates
  216. | |-- reduce/ # Extract insights
  217. | |-- reflect/ # Find connections
  218. | |-- reweave/ # Backward pass
  219. | |-- verify/ # Combined quality check
  220. | +-- ... # 12 more processing commands
  221. |-- agents/
  222. | +-- knowledge-guide.md # Pipeline subagent
  223. |-- hooks/
  224. | |-- hooks.json # Hook configuration
  225. | +-- scripts/ # Hook implementations
  226. |-- generators/
  227. | |-- claude-md.md # CLAUDE.md template
  228. | +-- features/ # 17 composable feature blocks
  229. |-- methodology/ # 249 research claims
  230. |-- reference/ # Core reference documents
  231. | |-- kernel.yaml # 15 kernel primitives
  232. | |-- three-spaces.md # Architecture spec
  233. | +-- use-case-presets.md # Pre-validated configs
  234. |-- platforms/ # Platform-specific adapters
  235. | |-- claude-code/
  236. | +-- shared/
  237. |-- presets/ # Pre-validated configurations
  238. |-- scripts/ # Utility scripts
  239. +-- README.md
  240. ```
  241. ---
  242. ## Development
  243. Clone this repo and add the marketplace to Claude Code:
  244. ```
  245. /plugin marketplace add ~/path-to-arscontexta
  246. ```
  247. Install the plugin:
  248. ```
  249. /plugin install arscontexta@agenticnotetaking
  250. ```
  251. Every time you make changes, re-install the plugin:
  252. ```
  253. /plugin uninstall arscontexta@agenticnotetaking
  254. /plugin install arscontexta@agenticnotetaking
  255. ```
  256. ### Key Files for Contributors
  257. - `reference/kernel.yaml` -- 15 primitives every system must include
  258. - `generators/features/*.md` -- composable feature blocks
  259. - `skill-sources/*/SKILL.md` -- generated command templates
  260. - `skills/setup/SKILL.md` -- the derivation engine
  261. - `reference/use-case-presets.md` -- preset definitions
  262. ---
  263. ## Presets
  264. Three pre-validated configurations for common use cases:
  265. | Preset | For | What You Get |
  266. |--------|-----|-------------|
  267. | **Research** | Academic work, literature reviews, synthesis | Atomic claims, citation tracking, methodology MOCs |
  268. | **Personal** | Life management, journaling, relationships | Reflective notes, goal tracking, relationship MOCs |
  269. | **Experimental** | Testing, iteration, rapid prototyping | Lightweight structure, fast capture, minimal ceremony |
  270. Presets provide starting defaults. The derivation engine adapts from there based
  271. on your conversation.
  272. ---
  273. ## Roadmap
  274. | Feature | Status |
  275. |---------|--------|
  276. | Claude Code plugin | Available |
  277. | Marketplace listing | Available |
  278. | Multi-agent processing | In progress |
  279. ---
  280. ## Philosophy
  281. The name connects to a tradition. **Ars Combinatoria**, **Ars Memoria**,
  282. **Ars Contexta**: the art of context.
  283. Llull's rotating wheels generated truth through combination. Bruno's memory wheels
  284. created millions of image combinations. They were external thinking systems -- tools
  285. to think with rather than just store in. The missing piece: they required a human
  286. mind to do the traversing. Now LLMs can traverse. The wheels can spin again.
  287. Built on [Tools for Thought for Agents](https://github.com/agenticnotetaking) research.
  288. ---
  289. ## License
  290. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.