@mindrian_os/cli 2.0.0-beta.29 → 2.0.0-beta.31

package/references/taxonomy/TAXONOMY.md

@@ -0,0 +1,192 @@
+---
+name: MindrianOS Taxonomy
+version: 1.0.0
+updated: 2026-03-31
+status: living-document
+rule: EVERY command, skill, agent, script, and Larry response MUST use these terms. No synonyms. No freestyle naming.
+---
+
+# MindrianOS Taxonomy -- The Definitive Naming System
+
+> This is a LIVING DOCUMENT. Updated whenever new concepts are added. Every piece of MindrianOS -- code, docs, Larry's voice, onboarding, error messages -- uses these exact terms. No exceptions.
+
+## Why This Exists
+
+Users, Larry, code, and docs must speak the same language. "Room" always means Room. "Entry" always means Entry. "Spark" always means Spark. Consistency builds trust. Inconsistency builds confusion.
+
+**Rule:** If a term isn't in this taxonomy, it doesn't exist in MindrianOS. Add it here first, then use it everywhere.
+
+---
+
+## The Space
+
+| Term | What It Is | Where It Lives | Icon |
+|------|-----------|----------------|------|
+| **MindrianOS** | The operating system for structured thinking | Plugin root | -- |
+| **Room** | A self-contained knowledge space for one project. The fundamental unit. | `room/` or `rooms/{name}/` | -- |
+| **Registry** | Tracks all rooms: which is active, git state, deploy URL | `.rooms/registry.json` | -- |
+
+## Inside a Room
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Section** | A domain of knowledge (problem, market, solution, team...). Like a wing of a building. | `room/{section}/` |
+| **Entry** | Any markdown knowledge artifact filed in a section. The standard unit of captured knowledge. | `room/{section}/{name}.md` |
+| **Asset** | A binary file -- PDF, image, video, audio -- with a markdown wrapper for frontmatter. | `room/assets/{section}/` |
+| **Manifest** | Auto-generated index of all assets with metadata. | `room/ASSET_MANIFEST.md` |
+| **Blueprint** | Section descriptor: what this section is for, starter questions, default methodologies. | `room/{section}/ROOM.md` |
+
+## Micro-Knowledge
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Spark** | The smallest unit of knowledge. One thought, one link, one quote, one pushback. Zero-friction capture. | `room/.sparks/{date}.md` |
+| **Catch** | The act of capturing a spark. Larry auto-detects or user says "catch that." | Passive (Larry) or active (user) |
+| **Ember** | 3+ sparks converging on the same theme. Larry detects and surfaces. | KuzuDB SPARK_CONVERGES edge |
+| **Crystallize** | Promote accumulated sparks into a full Entry with frontmatter and section filing. | `/mos:crystallize` |
+| **Residue** | 0-3 sparks extracted at session end from conversation. Auto-captured. | on-stop hook |
+| **Fragment** | Meeting "noise" segments that contain micro-knowledge. Saved instead of discarded. | Meeting filing step |
+
+## The Pulse (Context Files)
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **State** | Quantitative pulse: computed stats, entries, sections, stage, progress. | `room/STATE.md` |
+| **Thesis** | Qualitative pulse: Minto pyramid -- governing thought, arguments, evidence. | `room/MINTO.md` |
+| **Profile** | Who the user is: background, expertise, working style. Global across rooms. | `room/USER.md` |
+| **Mission** | What the user is trying to accomplish in THIS room. The room's JTBD. | `room/JTBD.md` |
+| **Decision** | A choice made at a decision point, with reasoning and alternatives considered. | `room/{section}/DECISION.md` |
+| **Claims** | Tracked assumptions with validity status (unvalidated/validated/invalidated). | `room/assumptions.json` |
+
+## People
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **People** | The human layer: team members, mentors, advisors, stakeholders. | `room/team/` |
+| **Member** | One person with role, expertise, meeting history, contact. | `room/team/{category}/{slug}/PROFILE.md` |
+| **Lenses** | AI perspective personas (De Bono Six Hats) generated from room data. | `room/personas/` |
+
+## Conversations
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Conversation** | Any dialogue that generates knowledge. Not just meetings -- calls, chats, interviews. | `room/meetings/{date}-{name}/` |
+| **Segment** | One classified piece of a conversation: insight, decision, action, advice, question. | Filed to section with frontmatter |
+| **Commitment** | Something someone promised to do. Has owner + deadline + status. | `room/action-items.md` |
+| **Echoes** | Themes recurring across multiple conversations. Cross-meeting intelligence. | `room/MEETINGS-INTELLIGENCE.md` |
+
+## The Fabric (Knowledge Graph)
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Fabric** | KuzuDB -- the invisible backbone connecting all entries. Users never see it directly. | `.lazygraph/` |
+| **Thread** | A typed relationship between two entries. The connections that make the room intelligent. | KuzuDB edge |
+| **Thread Types** | INFORMS, CONTRADICTS, CONVERGES, ENABLES, INVALIDATES | Edge labels |
+| **Surprise** | HSI-discovered non-obvious connection between entries. The hidden link nobody expected. | HSI_CONNECTION edge |
+| **Bottleneck** | Reverse Salient -- where the room's understanding lags behind its ambition. (Hughes 1983) | REVERSE_SALIENT edge |
+| **Weave** | The graph.json export that renderers consume. The Fabric made visible. | `dashboard/graph.json` |
+
+## Signals (Intelligence)
+
+| Term | What It Is | When It Fires |
+|------|-----------|---------------|
+| **Blind Spot** | An empty section or missing perspective. Something you haven't explored. | GAP detection |
+| **Tension** | Two entries that say opposite things. A contradiction worth investigating. | CONTRADICT detection |
+| **Pattern** | Same theme appearing in 3+ entries independently. Convergence = confidence growing. | CONVERGE detection |
+| **Signal** | Any intelligence Larry surfaces without being asked. Max 1-2 per session, context-relevant. | Proactive intelligence |
+| **Radar** | The ambient intelligence bar in every output (Zone 3). Where signals appear. | UI Zone 3 |
+
+## Tools (What Larry Uses)
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Tool** | One of 26+ PWS methodology frameworks as a `/mos:` slash command. | `commands/{name}.md` |
+| **Playbook** | The methodology's full reference material. The theory behind the tool. | `references/methodology/{name}.md` |
+| **Script** | Bash/CJS/Python automation that runs in background. The machinery. | `scripts/{name}` |
+| **Operator** | CJS module wrapping a script for programmatic use. | `lib/core/{name}-ops.cjs` |
+| **Skill** | Always-on rule set Larry follows every session. Auto-loaded, never invoked. | `skills/{name}/SKILL.md` |
+| **Agent** | Subagent with isolated context for specific tasks. Delegated work. | `agents/{name}.md` |
+
+## Pipelines (How Things Chain)
+
+| Term | What It Is | Where It Lives |
+|------|-----------|----------------|
+| **Pipeline** | Ordered chain of tools where each output feeds the next input. | `pipelines/{name}/` |
+| **Stage** | One tool in the chain with input/output contract. | `pipelines/{name}/{nn}-{tool}.md` |
+| **Handoff** | What one stage passes to the next. Structured output contract. | Frontmatter: requires/provides |
+| **Trail** | Metadata tracking which pipeline produced an entry. Provenance. | Frontmatter: pipeline, pipeline_stage |
+
+## Autonomous Work
+
+| Term | What It Is | How It Works |
+|------|-----------|-------------|
+| **Autopilot** | Larry acts independently -- selects + runs the best framework for current state. | `/mos:act` |
+| **Navigation** | Brain query or local routing table picks the right tool. Framework selection. | Act Step 3 |
+| **Trace** | Larry shows WHY he picked this framework before running it. Thinking transparency. | Act Step 2 (Shape E) |
+| **Sequence** | 3-5 tools in series, each output feeds the next. Chain mode. | `--chain` flag |
+| **Preview** | Show the execution plan without running anything. Safe look-ahead. | `--dry-run` flag |
+| **Sandbox** | Framework runs in subagent context. Main session stays clean. | `agents/framework-runner.md` |
+
+## The Cascade (Filing Pipeline)
+
+| Step | Name | What Happens | Timing |
+|------|------|-------------|--------|
+| 1 | **Classify** | What section does this belong to? | Sync, <100ms |
+| 2 | **Stamp** | Inject artifact hash ID into frontmatter | Sync, <50ms |
+| 3 | **Weave** | Create KuzuDB node + detect Thread edges | Sync, <500ms |
+| 4 | **Refresh** | Recompute State from filesystem | Background |
+| 5 | **Map** | Rebuild Weave (graph.json) from Fabric | Background |
+| 6 | **Discover** | HSI + Reverse Salient computation (Surprises + Bottlenecks) | Background |
+| 7 | **Version** | Git commit with provenance message | Background |
+| 8 | **Publish** | Push to GitHub, Vercel redeploys Showcase | Background |
+
+## Showcase (Presentation Views)
+
+| Term | What It Is | File |
+|------|-----------|------|
+| **Showcase** | All 6 HTML views generated from a room. The room made visible. | `/mos:export presentation` |
+| **Overview** | Dashboard: stats, cards, video, assets, opportunities | `dashboard.html` |
+| **Library** | Wiki: 3-panel article browser with search + TOC | `wiki.html` |
+| **Narrative** | Deck: fullscreen presentation slides from Thesis + entries | `deck.html` |
+| **Synthesis** | Insights: stat counters, timelines, funnels, heat maps | `insights.html` |
+| **Blueprint** | Diagrams: SVG architecture from graph edges | `diagrams.html` |
+| **Constellation** | Graph: Canvas knowledge graph with particles + glow | `graph.html` |
+
+## CLI Visual Identity (De Stijl Terminal)
+
+| Term | What It Is |
+|------|-----------|
+| **Four Zones** | Header, Body, Radar, Footer -- every output follows this anatomy |
+| **Shapes** | 5 body layouts: Board, Tree, Card, Document, Report |
+| **Glyphs** | 12 approved symbols for status/type indicators |
+| **Palette** | 5 ANSI colors: cream, red, blue, yellow, muted |
+| **Greeting** | Cold/warm/signals contract at session start |
+| **Signature** | MindrianOS logo + "Built with MindrianOS" + Mondrian color bar |
+
+## Onboarding
+
+| Term | What It Is | When It Happens |
+|------|-----------|----------------|
+| **Welcome** | First-time MindrianOS onboarding. Larry learns who you are. | First install detection |
+| **Briefing** | Per-room clarification. Larry learns what you're building HERE. | `/mos:new-project` |
+| **Quick Briefing** | 3 questions: what, who, when. Fast start. | User chooses quick |
+| **Deep Briefing** | JTBD interview: situation, motivation, outcome, anxieties. Rich context. | User chooses deep |
+| **Refresh** | After-update flow: what's new + core features reminder. | `/mos:update` |
+
+---
+
+## Usage Rules
+
+1. **Larry ALWAYS uses these terms** in conversation. "I'll file this as an Entry in your market-analysis Section." Not "I'll save this document to the folder."
+
+2. **Error messages use these terms.** "No Room found in this workspace" not "No room/ directory."
+
+3. **Onboarding teaches these terms.** The Welcome and Briefing flows introduce the vocabulary naturally.
+
+4. **Code comments use these terms.** `// Weave: create KuzuDB node + detect Threads` not `// index artifact in graph database`
+
+5. **This file is the source of truth.** If a term isn't here, it doesn't exist in MindrianOS. Add it here FIRST, then use it everywhere.
+
+---
+
+*Living document. Last updated: 2026-03-31. Version 1.0.0.*

package/references/templates/MINTO.md

@@ -0,0 +1,36 @@
+---
+section: "{section-name}"
+governing_thought: ""
+mece_check: pending
+last_updated: ""
+---
+
+# MINTO Reasoning Pyramid
+
+## Governing Thought
+
+[The single sentence that captures what this section proves or argues. This is the answer-first statement that everything below supports.]
+
+## Supporting Arguments
+
+1. [First supporting argument -- one clear claim]
+2. [Second supporting argument -- independent from the first]
+3. [Third supporting argument -- completes the coverage]
+
+## Evidence
+
+### Argument 1 Evidence
+- [Evidence item from room entries, meetings, or research]
+
+### Argument 2 Evidence
+- [Evidence item from room entries, meetings, or research]
+
+### Argument 3 Evidence
+- [Evidence item from room entries, meetings, or research]
+
+## MECE Check
+
+- [ ] Arguments are Mutually Exclusive (no overlap between arguments)
+- [ ] Arguments are Collectively Exhaustive (no gaps in coverage)
+- [ ] Each argument has at least one evidence item
+- [ ] Governing thought follows logically from arguments

package/references/user-research/2026-04-05-leah-lawrence-session.md

@@ -0,0 +1,202 @@
+# User Session: Lawrence + an admin-key holder -- Blueprint Phase AI Consultancy
+
+**Date:** 2026-04-05
+**Users:** Lawrence Aronhime (professor, power user), an admin-key holder (daughter, new user)
+**Environment:** macOS, Sonnet 4.6, Claude Pro, v1.7.1
+**Project:** Blueprint Phase AI Consultancy
+**Duration:** CLI session ~60min, meeting ~1h57m
+**Outcome:** Hit Pro daily limit during chain execution. 5/6 chain steps completed.
+
+---
+
+## Part 1: CLI Session Transcript Analysis
+
+### Session Timeline
+
+| Time | Event | Issue? |
+|---|---|---|
+| Start | /mos:onboard attempted | node: command not found (PATH issue on macOS) |
+| +5:48 | Session-start finally completes | **CRITICAL: 5m48s churn** |
+| +6:00 | Room status displayed correctly | Working |
+| +7:00 | /mos:lean-canvas started | Working -- Larry challenged answers well |
+| +12:00 | User typed "you tell me" | Larry handled gracefully, quoted user's own prior work |
+| +15:00 | /mos:act dispatched | Selected map-unknowns (correct choice) |
+| +18:00 | map-unknowns completed (3m6s) | Filed to market-analysis, 2 cross-refs found |
+| +20:00 | User said "first 1 then 2" | Larry understood sequential request |
+| +25:00 | /mos:validate requested mid-flow | User interrupted challenge-assumptions for validate |
+| +28:00 | User: "this is an ill-defined problem" | **KEY MOMENT: Lawrence self-classifying the problem type** |
+| +30:00 | /mos:act --chain built 3-step chain | macro-trends -> user-needs -> lean-canvas |
+| +35:00 | User: "add online research" | Chain modified mid-execution to 6 steps |
+| +38:00 | User: "do systems thinking and then find analogies next" | Further chain modification |
+| +45:00 | User: "how much longer?" | **UX gap: no progress estimation shown proactively** |
+| +47:00 | Snapshot of findings requested | Larry delivered excellent mid-session synthesis |
+| +55:00 | Step 4/6 (systems-thinking) completed | Filed to problem-definition |
+| +60:00 | Step 5/6 (find-analogies) started | **HIT PRO DAILY LIMIT** -- session terminated |
+
+### Teaching Quality Assessment
+
+**Excellent moments:**
+- "That's not the problem -- that's the symptom." (Box 1, rejecting surface answer)
+- "You're sliding back into skills-gap territory." (Box 3, catching regression)
+- "They discover they're MORE valuable because of AI -- not despite it." (Quoting user's own prior work)
+- "Your IR segment choice is relationship-driven, not market-driven." (map-unknowns surfacing)
+- "Is your brother scared because he's an interventional radiologist -- or because he's your brother?" (challenge-assumptions killing question)
+- Mid-session synthesis: 3 frameworks independently validated the imagination ceiling
+
+**Larry's voice worked:** Concise, challenging, built on user's own language. The Ask-Tell dial was well-calibrated -- investigative early, then shifted to insight delivery after enough evidence accumulated.
+
+### Critical UX Issues
+
+| # | Issue | Severity | Impact |
+|---|---|---|---|
+| 1 | **node: command not found** | HIGH | Session-start fails, detect-integrations breaks. macOS PATH not finding node. |
+| 2 | **5m48s startup churn** | HIGH | User asked "why does my request for onboard take so long?" First impression ruined. |
+| 3 | **No chain progress indicator** | MEDIUM | User had to ask "how much longer?" after 14 minutes of silence. |
+| 4 | **Pro limit hit mid-chain** | HIGH | 5 of 6 chain steps completed, lean-canvas (the synthesis) never ran. User's business model -- the whole point -- was never built. |
+| 5 | **No cost estimation before chain** | MEDIUM | User didn't know a 6-step chain would consume their entire daily Pro budget. |
+| 6 | **Chain modification UX** | LOW | User could modify chain mid-run (good!) but no confirmation of updated chain structure shown. |
+
+---
+
+## Part 2: Family Meeting Transcript Analysis
+
+### Participants
+- **Jonathan Sagir** -- developer, demonstrating MindrianOS
+- **Lawrence Aronhime** -- professor (Larry model), power user, demonstrating to Leah
+- **an admin-key holder** -- daughter, new user, fashion/design background, has her own Blueprint Phase project
+
+### Key User Experience Observations
+
+#### 1. Room Confusion (Lawrence)
+> "I don't know where my rooms are even."
+> "I've just been dumping everything into here."
+> "I was stressing me out, the file structure, because I didn't understand it."
+> "Most people are just gonna wanna use it without worrying about all this."
+
+**Finding:** Even the professor the system is modeled on finds room organization stressful. File structure IS the orchestration, but users don't want to think about file structure.
+
+#### 2. Multi-Room Context Contamination
+> Jonathan: "If you talk about a few things in this pool of context, it will get scrambled and diluted."
+> Lawrence: "So my gluten-free recipes have nothing to do with any of this research."
+> Leah: "So I put it in a separate room... and I think that was the right thing to do."
+
+**Finding:** Leah intuitively understood room separation. Lawrence mixed things. The system needs to proactively detect topic drift and suggest new rooms.
+
+#### 3. New Terminal / New Session Confusion (Leah)
+> Leah: "Every time I do this I have to ask one of my active chats, 'How do I start a new conversation?'"
+> Leah: "I just downloaded it to my desktop and pulled it in from there."
+> Jonathan: "Something's wrong in your machine."
+
+**Finding:** macOS admin restrictions + Claude Code installation creates recurring friction. Users need a one-click way to start fresh.
+
+#### 4. Slash Commands Discovery
+> Jonathan: "When you have a dead end, just tell me what commands can be appropriate."
+> Lawrence: "I could just do this at any point during any conversation?"
+> Jonathan: "Any point."
+
+**Finding:** The /mos:help command works, but users don't know to ask for it. Commands should surface contextually, not wait to be asked.
+
+#### 5. Grant Discovery Wow Moment
+> Lawrence: "This is crazy. I didn't tell it to do any of this."
+> Jonathan: "It searched the grants.gov API. You never gave it the API."
+> Lawrence: "Who would ever look there? The Kyrgyz Republic."
+> [They validated the grant on Google -- it was real]
+
+**Finding:** Opportunity scanning is the most visually impressive feature for non-technical users. Real grants, real URLs, real relevance scoring. The "unfair advantage" framing resonates.
+
+#### 6. Snapshot Export Issues
+> Lawrence: "These links in these tabs... they do not display properly."
+> [Dashboard tabs were empty/broken]
+> Jonathan: "It's just need to wire something probably."
+
+**Finding:** Export/snapshot quality is a first-impression feature. Broken tabs undermine the "wow" moment.
+
+#### 7. Context Window Awareness
+> Jonathan: "You see the skeleton? Eighty-nine percent."
+> Lawrence: "When I see these numbers hit the nineties, what do I do? Retire?"
+> Jonathan: "Just let it."
+
+**Finding:** Users don't understand autocompact. The skull icon at 89% creates anxiety without explaining what happens next. Need graceful messaging.
+
+#### 8. Teaching Quality Validated by Lawrence
+> Lawrence: "I don't think the lean canvas was helpful."
+> Lawrence: "I'm not convinced that scenario analysis is the way to go here."
+> Lawrence: "I think system thinking and then find analogies."
+
+**Finding:** Lawrence actively steered methodology selection -- exactly as designed. He used his domain expertise to override/redirect Larry's suggestions. The system accommodated this gracefully.
+
+#### 9. Case Study Evidence
+> Lawrence: "Lea won two grants using Mindry v1."
+> Jonathan: "That's a case study right there."
+> Lawrence: "An IRIS group used PWS, filed with Mindry, got a grant. The whole cycle."
+
+**Finding:** Real success stories exist. Leah's 2 grants + IRIS group grant = 3 validated case studies.
+
+#### 10. Speed / Autonomous Mode
+> Lawrence: "I created the student, it had the conversation, and it did all the work for the student. Two minutes."
+> Jonathan: "That's too easy then. It has to be an interaction."
+> Lawrence: "Students will figure that out in two minutes."
+
+**Finding:** /mos:act is TOO autonomous when used without interaction. Students bypass learning. Solution: students must do exercises by hand first, then use MindrianOS. Defense via deck presentation is the validation gate.
+
+#### 11. Leah as New User Archetype
+> Leah: "I have no idea what any of this means." (on PEST analysis)
+> Lawrence explains PEST to Leah in 10 seconds
+> Leah immediately understands after one-sentence explanation
+
+**Finding:** Non-business users need one-sentence explanations of frameworks. Larry could provide these inline. "PEST = political, economic, social, technological forces shaping your market."
+
+#### 12. Cross-Domain Opportunity
+> Jonathan: "What I'm aiming for is a researcher just feeds it his research, and then it opens up a world of possibility for commercialization."
+> Lawrence: "If this can do what I think it can do and what I'm hoping my science colleagues tell me it can do, this is it."
+> Lawrence: "We need our case studies. I think the tool is built, it's the case studies now."
+
+**Finding:** The product is ready for validation. The blocker is case studies, not features.
+
+---
+
+## Actionable Items for v1.8.0
+
+### Critical (P0)
+1. **Fix macOS PATH for node** -- detect-integrations and hook scripts fail when node isn't in PATH
+2. **Reduce session-start time** -- 5m48s is unacceptable. Target: <3s
+3. **Add chain progress estimation** -- show "Step 2/6, ~15 minutes remaining" proactively
+4. **Add chain cost warning** -- "This 6-step chain will use approximately 300K tokens (~$X). Your Pro limit may be reached. Continue?"
+
+### Important (P1)
+5. **Proactive room separation** -- detect topic drift and suggest "This seems like a new project. Want to open a new room?"
+6. **Graceful autocompact messaging** -- replace skull icon anxiety with "Context is getting full. Your room data is safe -- I'll compress our conversation and keep working."
+7. **Fix export/snapshot tab rendering** -- dashboard tabs must work on first impression
+8. **One-sentence framework explanations** -- when a framework runs, lead with "PEST analyzes political, economic, social, and technological forces."
+
+### Nice to Have (P2)
+9. **Unfinished chain recovery** -- when Pro limit kills a chain, offer to resume on next session
+10. **Room location helper** -- "Your rooms are at /Users/laronhime2/room/" with Finder-openable path
+11. **Context percentage explanation** -- tooltip or inline explanation of what 89% means
+
+---
+
+## Quotes Worth Preserving
+
+### On MindrianOS Value
+> "I didn't tell it to do any of this." -- Lawrence, on grant discovery
+> "This is the real Mindrian dream." -- Jonathan, on the chain running autonomously
+> "It's like really unbelievable when you think about it, that it does all this on its own." -- Leah
+
+### On Teaching Quality
+> "That's not the problem -- that's the symptom." -- Larry (AI), redirecting Lawrence
+> "You're sliding back into skills-gap territory." -- Larry (AI), catching regression
+> "Your IR segment choice is relationship-driven, not market-driven." -- Larry (AI), map-unknowns insight
+
+### On User Friction
+> "I was stressing me out, the file structure, because I didn't understand it." -- Lawrence
+> "Every time I do this I have to ask how to start a new conversation." -- Leah
+> "When I see these numbers hit the nineties, what do I do? Retire?" -- Lawrence
+
+### On Product Readiness
+> "We need our case studies. I think the tool is built, it's the case studies now." -- Lawrence
+> "Lea won two grants using Mindry v1." -- Lawrence
+> "If we unlock hard science, everything else is by definition unlockable." -- Jonathan
+
+### On the Identity Ceiling
+> "People can't imagine a different version of themselves on the other side of AI -- so they either panic or stay shallow." -- Larry (AI), distilling Lawrence's own work

package/references/vault-kit/README.md

@@ -0,0 +1,35 @@
+# Obsidian Vault Kit
+
+This directory is the canonical, version-controlled source for the Obsidian
+configuration kit used by `/mos:vault` (Phase 78). It contains the static
+assets that must be copied verbatim into any exported vault's `.obsidian/`
+folder so the vault immediately renders with MindrianOS De Stijl theme,
+section-colored graph view, and four ready-to-use note templates.
+
+## Contents
+
+| Source                                 | Destination in vault                        | Purpose                                      |
+|----------------------------------------|---------------------------------------------|----------------------------------------------|
+| `snippets/mindrian-destijl.css`        | `.obsidian/snippets/mindrian-destijl.css`   | De Stijl dark theme (red/blue/gold/Mondrian) |
+| `appearance.json`                      | `.obsidian/appearance.json`                 | Enables snippet + dark mode + fonts          |
+| `app.json`                             | `.obsidian/app.json`                        | Sane defaults (live preview, shortest links) |
+| `graph.json`                           | `.obsidian/graph.json`                      | Section color groups + force layout          |
+| `templates/new-artifact.md`            | `.obsidian/templates/new-artifact.md`       | Content artifact template (RULES-08)         |
+| `templates/new-meeting-note.md`        | `.obsidian/templates/new-meeting-note.md`   | Meeting note template (RULES-08)             |
+| `templates/new-team-profile.md`        | `.obsidian/templates/new-team-profile.md`   | Team profile template (RULES-08)             |
+| `templates/new-xref.md`                | `.obsidian/templates/new-xref.md`           | Cross-reference template (RULES-08)          |
+
+## Requirement Traceability
+
+- **KIT-01** De Stijl CSS snippet -- `snippets/mindrian-destijl.css` (verbatim from align-ecosystem, 297 lines)
+- **KIT-02** Obsidian app defaults -- `app.json`
+- **KIT-03** Section-colored graph view -- `graph.json` (>=11 colorGroups, hideUnresolved, showArrow)
+- **KIT-04** Appearance enables snippet + dark mode -- `appearance.json`
+- **KIT-05** Graph force layout tuning -- `graph.json` (linkDistance=180, centerStrength=0.4, repelStrength=12)
+- **RULES-08** Four note templates matching `FRONTMATTER_SCHEMA` -- `templates/*.md`
+
+## Phase 78 Contract
+
+`/mos:vault` must rsync this directory's contents into the target vault's
+`.obsidian/` folder (never mutate the files during copy). The kit is
+intentionally a pure static asset set -- zero generators, zero templating.

package/references/vault-kit/app.json

@@ -0,0 +1,12 @@
+{
+  "promptDelete": false,
+  "alwaysUpdateLinks": true,
+  "newLinkFormat": "shortest",
+  "useMarkdownLinks": false,
+  "attachmentFolderPath": "./attachments",
+  "newFileLocation": "root",
+  "showLineNumber": true,
+  "strictLineBreaks": false,
+  "livePreview": true,
+  "defaultViewMode": "preview"
+}

package/references/vault-kit/appearance.json

@@ -0,0 +1,12 @@
+{
+  "baseFontSize": 16,
+  "theme": "obsidian",
+  "cssTheme": "",
+  "enabledCssSnippets": [
+    "mindrian-destijl"
+  ],
+  "interfaceFontFamily": "Inter",
+  "textFontFamily": "Inter",
+  "monospaceFontFamily": "JetBrains Mono",
+  "translucency": false
+}

package/references/vault-kit/graph.json

@@ -0,0 +1,35 @@
+{
+  "collapse-filter": true,
+  "search": "",
+  "showTags": false,
+  "showAttachments": false,
+  "hideUnresolved": true,
+  "showOrphans": false,
+  "collapse-color-groups": false,
+  "colorGroups": [
+    { "query": "path:problem-definition",   "color": { "a": 1, "rgb": 13123887 } },
+    { "query": "path:business-model",       "color": { "a": 1, "rgb": 2841509  } },
+    { "query": "path:financial-model",      "color": { "a": 1, "rgb": 15247928 } },
+    { "query": "path:competitive-analysis", "color": { "a": 1, "rgb": 4889263  } },
+    { "query": "path:solution-design",      "color": { "a": 1, "rgb": 4885084  } },
+    { "query": "path:team",                 "color": { "a": 1, "rgb": 9131516  } },
+    { "query": "path:team-execution",       "color": { "a": 1, "rgb": 9131516  } },
+    { "query": "path:market-analysis",      "color": { "a": 1, "rgb": 13936195 } },
+    { "query": "path:meetings",             "color": { "a": 1, "rgb": 7039851  } },
+    { "query": "path:legal-ip",             "color": { "a": 1, "rgb": 10177482 } },
+    { "query": "path:opportunity-bank",     "color": { "a": 1, "rgb": 15247928 } },
+    { "query": "path:sub-rooms",            "color": { "a": 1, "rgb": 4889263  } }
+  ],
+  "collapse-display": false,
+  "showArrow": true,
+  "textFadeMultiplier": 0,
+  "nodeSizeMultiplier": 1.4,
+  "lineSizeMultiplier": 1,
+  "collapse-forces": false,
+  "centerStrength": 0.4,
+  "repelStrength": 12,
+  "linkStrength": 1,
+  "linkDistance": 180,
+  "scale": 1,
+  "close": true
+}