@mindrian_os/install 1.13.0-beta.17 → 1.13.0-beta.21

package/commands/help.md

@@ -1,6 +1,7 @@
 ---
 name: help
 description: List commands grouped by flow (tldr-style)
+help_jtbd: "List commands grouped by flow, with one-line outcomes."
 argument-hint: [command-name]
 body_shape: B (Semantic Tree)
 body_shape_detail: -- (inline, no zones)
@@ -15,316 +16,46 @@ allowed-tools:
 
 # /mos:help
 
-You are Larry. This command helps users discover what they can do. Uses **Body Shape B (Semantic Tree)** with **De Stijl color-coded job categories** and **JTBD outcome descriptions**.
+You are Larry. List commands grouped by job category, with one-line outcome descriptions ("what's in it for me").
 
-## Design Rules
+## How to render the default `/mos:help`
 
-1. **Every description is a JTBD outcome.** Not what the tool does -- what the user GETS. "Build reasoning that survives a boardroom challenge" not "Minto Pyramid".
-2. **Every command gets a colored block.** The color tells you what KIND of thinking before you read the description.
-3. **Groups are by job category, not alphabetical.** Users scan for what they need to DO.
+Invoke `scripts/help-renderer.cjs` and emit its output verbatim. The script handles all terminal-capability dispatch (truecolor / 256color / ASCII) via `lib/core/terminal-capability.cjs`. The 4-zone format is preserved by the renderer's design (body_shape: B Semantic Tree).
 
-## De Stijl Color System
-
-Six Mondrian accent colors, each mapped to a thinking job:
-
-| ANSI Code | Color | Job Category | What It Means |
-|-----------|-------|-------------|---------------|
-| `\033[38;2;166;61;47m` | RED | Problem Discovery | Find what's worth solving |
-| `\033[38;2;30;58;110m` | BLUE | Structured Thinking | Build reasoning that holds |
-| `\033[38;2;107;78;139m` | AMETHYST | Perspectives + Creativity | See through lenses you'd never pick |
-| `\033[38;2;200;164;60m` | YELLOW | Intelligence + Brain | Answers from 23K nodes of teaching data |
-| `\033[38;2;45;107;74m` | GREEN | Output + Export | Ship investor-ready work |
-| `\033[38;2;42;107;94m` | TEAL | Infrastructure | Setup, manage, maintain |
-
-Supporting ANSI codes:
-- Cream (headers): `\033[38;2;245;240;232m`
-- Muted (descriptions): `\033[38;2;160;154;144m`
-- Reset: `\033[0m`
-
-Use these EXACT hex-mapped ANSI codes. They match the website and dashboard palette.
-
-## UI Format
-
-- **Default (`/mos:help`):** 4-zone anatomy with color-coded Semantic Tree
-- **Per-command (`/mos:help [command]`):** tldr-style inline -- 1 description line + 3 examples max, no zones
-- **`--all` flag:** Full command list, no truncation, same color system
-
-## Brain Enhancement (Optional)
-
-Try calling Brain: first `mcp__mindrian-brain__brain_schema`, then `mcp__mindrian-brain__get_neo4j_schema` as fallback. If it succeeds, Brain mode is active. If it fails or errors, skip this section entirely and proceed to Step 1 below.
-
-**If Brain connected:**
-
-1. Read `references/brain/query-patterns.md` for `brain_framework_chain` and `brain_gap_assess` patterns
-2. Read `room/STATE.md` for current frameworks used and venture stage
-3. Run `brain_framework_chain` with the user's current frameworks and inferred problem type to get graph-informed personalized recommendations
-4. Run `brain_gap_assess` with `$room_frameworks` to identify specific missing prerequisites and natural next-step frameworks
-5. Use these Brain results to personalize the command recommendations beyond stage-based defaults. Brain data shows what actually works for this user's specific situation.
-
-Proceed to Step 1 below with this additional context.
-
-## Step 1: Determine Venture Stage
-
-Read `room/STATE.md` to find the current venture stage. If `room/` does not exist or STATE.md is missing, the stage is **Pre-Opportunity**.
-
-Extract `venture_stage` from the YAML frontmatter of STATE.md.
-
-## Step 1.5: Check Admin Visibility
-
-Check if the current user is an admin. This determines whether admin-only commands are visible.
-
-**Check in order:**
-
-1. Environment variable `MOS_ADMIN=true` is set
-2. Username contains "jsagi" or "jonathan" (check `$USER`, `$USERNAME`, or `whoami`)
-3. Home directory matches `/home/jsagi` (check `$HOME`)
-
-Set an internal flag `is_admin` to true if ANY condition is met, false otherwise.
-
-**Generic visibility filtering rule:** When listing commands, check each command file's YAML frontmatter for a `visibility` field. If `visibility: admin` is set and `is_admin` is false, skip that command entirely. This makes filtering generic -- any future hidden command just needs `visibility: admin` in its frontmatter.
-
-## Step 2: Load References
-
-Read `references/methodology/index.md` for the full command routing table.
-Read `references/methodology/problem-types.md` for problem type classification and methodology routing by definition level and complexity.
-
-These are your source of truth for all commands, descriptions, stage mappings, and framework recommendations.
-
-## Step 3: Default Behavior (No Flags)
-
-If the user ran `/mos:help` with no flags, render the 4-zone output using ANSI colors.
-
-Define these variables at the top of your output generation:
-
-```
-R = \033[38;2;166;61;47m    (red)
-B = \033[38;2;30;58;110m    (blue)
-A = \033[38;2;107;78;139m   (amethyst)
-Y = \033[38;2;200;164;60m   (yellow)
-G = \033[38;2;45;107;74m    (green)
-T = \033[38;2;42;107;94m    (teal)
-C = \033[38;2;245;240;232m  (cream -- for headers)
-M = \033[38;2;160;154;144m  (muted -- for descriptions)
-X = \033[0m                  (reset)
-```
-
-### Zone 1 -- Header Panel
-
-```
-{M}╭─ MindrianOS ── Help ─────────────────────────────────────╮{X}
-{M}│                                                            │{X}
-{M}│{X}  {R}■{X} Problem    {B}■{X} Reasoning   {Y}■{X} Intelligence                {M}│{X}
-{M}│{X}  {A}■{X} Perspective  {G}■{X} Output    {T}■{X} Infrastructure              {M}│{X}
-{M}│                                                            │{X}
-```
-
-If room exists, use room name instead of "MindrianOS".
-
-### Zone 2 -- Content Body (Color-Coded JTBD Semantic Tree)
-
-Each command gets a colored ■ block matching its job category. Descriptions are JTBD outcomes -- what the user GETS, never what the tool DOES.
-
-**MANDATORY: Every description must pass the "so what?" test.** If a user reads it and thinks "so what?", rewrite it. The description should make them think "I need that."
+Specifically:
 
-```
-  {M}▼ Getting Started{X}
-  ├─ {T}■{X} /mos:new-project            {M}Build your Data Room from a conversation about your idea{X}
-  ├─ {T}■{X} /mos:onboard                {M}Learn what matters for YOUR stage in 5 minutes{X}
-  ├─ {T}■{X} /mos:setup                  {M}Connect Brain, transcription, or graph integrations{X}
-  ├─ {R}■{X} /mos:diagnose               {M}Know which methodology fits before you waste time guessing{X}
-  └─ {T}■{X} /mos:help [command]         {M}See exactly how any command works with examples{X}
-
-  {M}▼ Problem Discovery{X}
-  ├─ {R}■{X} /mos:beautiful-question     {M}Turn a vague idea into a question investors want answered{X}
-  ├─ {R}■{X} /mos:explore-domains        {M}Find the 2-3 domains where your innovation actually lives{X}
-  ├─ {R}■{X} /mos:explore-trends         {M}See where your market is heading before competitors do{X}
-  ├─ {R}■{X} /mos:map-unknowns           {M}Know what you don't know -- before it kills your venture{X}
-  ├─ {R}■{X} /mos:analyze-needs          {M}Discover what job your customer is actually hiring for{X}
-  ├─ {R}■{X} /mos:user-needs             {M}Watch how people actually behave, not what they say they want{X}
-  ├─ {R}■{X} /mos:root-cause             {M}Trace a problem to its source instead of treating symptoms{X}
-  ├─ {R}■{X} /mos:macro-trends           {M}Map the large-scale shifts reshaping your domain right now{X}
-  └─ {R}■{X} /mos:analyze-timing         {M}Know if you're too early, too late, or right on time{X}
-
-  {M}▼ Structured Thinking{X}
-  ├─ {B}■{X} /mos:structure-argument     {M}Build reasoning that survives a boardroom challenge{X}
-  ├─ {B}■{X} /mos:lean-canvas            {M}Get your business model on one page that makes sense{X}
-  ├─ {B}■{X} /mos:analyze-systems        {M}Find where leverage actually lives in a complex system{X}
-  ├─ {B}■{X} /mos:systems-thinking       {M}See the feedback loops driving your market's behavior{X}
-  ├─ {B}■{X} /mos:find-bottlenecks       {M}Identify the one lagging component holding everything back{X}
-  ├─ {B}■{X} /mos:validate               {M}Check if your evidence actually supports your claims{X}
-  ├─ {B}■{X} /mos:build-knowledge        {M}Separate your data from your opinions from your wisdom{X}
-  ├─ {B}■{X} /mos:build-thesis           {M}Have the investment narrative that makes VCs lean in{X}
-  └─ {B}■{X} /mos:grade                  {M}Get an honest score calibrated against 100+ real ventures{X}
-
-  {M}▼ Perspectives + Creativity{X}
-  ├─ {A}■{X} /mos:think-hats             {M}See your problem through 6 lenses you'd never pick yourself{X}
-  ├─ {A}■{X} /mos:challenge-assumptions  {M}Find the assumption that will break you -- before it does{X}
-  ├─ {A}■{X} /mos:scenario-plan          {M}Prepare for 4 plausible futures instead of betting on one{X}
-  ├─ {A}■{X} /mos:explore-futures        {M}See 10-year signals that change what you build today{X}
-  ├─ {A}■{X} /mos:dominant-designs       {M}Know if the market standard is locked in or cracking open{X}
-  ├─ {A}■{X} /mos:find-analogies         {M}Discover your problem was solved in another field years ago{X}
-  ├─ {A}■{X} /mos:score-innovation       {M}Score how promising a cross-domain opportunity really is{X}
-  ├─ {A}■{X} /mos:persona                {M}Get 6 persistent perspectives that challenge your blind spots{X}
-  └─ {A}■{X} /mos:leadership             {M}Know what kind of leader your team actually needs right now{X}
-
-  {M}▼ Intelligence + Brain{X}
-  ├─ {Y}■{X} /mos:query                  {M}Ask a question and get answers from your knowledge graph{X}
-  ├─ {Y}■{X} /mos:graph                  {M}Explore connections and patterns across your entire room{X}
-  ├─ {Y}■{X} /mos:research               {M}Get web evidence cross-referenced against Brain intelligence{X}
-  ├─ {Y}■{X} /mos:find-connections       {M}Discover links between your work and fields you never checked{X}
-  ├─ {Y}■{X} /mos:compare-ventures       {M}Learn what happened when others tried something similar{X}
-  ├─ {Y}■{X} /mos:scout                  {M}Run overnight scans for competitors, grants, and domain news{X}
-  ├─ {Y}■{X} /mos:opportunities          {M}Discover grants matched to your room context automatically{X}
-  ├─ {Y}■{X} /mos:funding                {M}Track non-dilutive funding from discovery to submission{X}
-  ├─ {Y}■{X} /mos:suggest-next           {M}Get a graph-informed recommendation on what to do next{X}
-  ├─ {Y}■{X} /mos:deep-grade             {M}Get percentile-ranked scoring against real student ventures{X}
-  ├─ {Y}■{X} /mos:diagnostics            {M}Wave-1 algorithmic fingerprint -- disruption, coverage, novelty, surprise{X}
-  └─ {Y}■{X} /mos:brain-derive           {M}Force-refresh BRAIN.md for section(s) now -- single, --all, or --cross-room{X}
-
-  {M}▼ Working Sessions{X}
-  ├─ {R}■{X} /mos:act                    {M}Larry picks the right framework -- you just describe the problem{X}
-  ├─ {B}■{X} /mos:pipeline               {M}Chain 3-5 frameworks where each output feeds the next{X}
-  ├─ {Y}■{X} /mos:file-meeting           {M}Turn a messy transcript into structured intelligence in 60s{X}
-  ├─ {Y}■{X} /mos:reanalyze              {M}Find patterns in old meetings you missed the first time{X}
-  └─ {Y}■{X} /mos:speakers               {M}See who attended your meetings and what expertise they brought{X}
-
-  {M}▼ Parallel Power{X}
-  ├─ {A}■{X} /mos:act --swarm            {M}Get answers from 3 frameworks at once -- minutes, not hours{X}
-  ├─ {A}■{X} /mos:persona --parallel     {M}Hear from all 6 perspectives simultaneously on one question{X}
-  ├─ {B}■{X} /mos:grade --full           {M}Grade every room section at once -- full venture health check{X}
-  ├─ {Y}■{X} /mos:research --broad       {M}Research from 3 angles at once and get a merged synthesis{X}
-  └─ {T}■{X} /mos:models                 {M}Control which AI model handles which type of work{X}
-
-  {M}▼ Output + Export{X}
-  ├─ {G}■{X} /mos:export                 {M}Generate investor-ready PDFs from your room in seconds{X}
-  ├─ {G}■{X} /mos:present                {M}Open a 6-view visual presentation of your entire venture{X}
-  ├─ {G}■{X} /mos:dashboard              {M}See your knowledge graph with chat in the browser{X}
-  ├─ {G}■{X} /mos:wiki                   {M}Browse your Data Room like Wikipedia -- linked, searchable{X}
-  ├─ {G}■{X} /mos:visualize              {M}See your venture as a knowledge graph, timeline, or diagram{X}
-  ├─ {G}■{X} /mos:publish                {M}Deploy your presentation to a live URL for stakeholders{X}
-  └─ {G}■{X} /mos:reason                 {M}Structure any room section into airtight MECE logic{X}
-
-  {M}▼ Infrastructure{X}
-  ├─ {T}■{X} /mos:status                 {M}See exactly where your project stands and what's missing{X}
-  ├─ {T}■{X} /mos:room                   {M}Browse your Data Room sections and launch the dashboard{X}
-  ├─ {T}■{X} /mos:rooms                  {M}Manage multiple ventures -- switch, park, archive{X}
-  ├─ {T}■{X} /mos:organize               {M}Restructure your room hierarchy with graph-informed proposals{X}
-  ├─ {T}■{X} /mos:explain-decision       {M}Show why Larry made the calls he made on the last turn{X}
-  ├─ {T}■{X} /mos:update                 {M}Check if a newer version of MindrianOS is available{X}
-  └─ {T}■{X} /mos:splash                 {M}Show the Mondrian banner{X}
+```bash
+node "${CLAUDE_PLUGIN_ROOT}/scripts/help-renderer.cjs"
 ```
 
-**If `is_admin` is true**, append:
+If `CLAUDE_PLUGIN_ROOT` is not set, fall back to the repo-relative path: `node ./scripts/help-renderer.cjs`.
 
-```
-  {M}▼ Admin (owner only){X}
-  └─ {T}■{X} /mos:admin                  {M}Manage Brain API keys and access tiers{X}
-```
+DO NOT compose the help text inline. DO NOT hardcode color escapes in the LLM response. The renderer is the single source of truth (per Phase 121.5-07 D-05 / D-06 LOCKED).
 
-**If `is_admin` is false**, do NOT render the Admin group. No trace of `/mos:admin` should appear anywhere in the output.
+## How the renderer composes output
 
-### Color Legend (ALWAYS render after the tree)
+- `data/help-groups.json` declares the 11 canonical groups (10 user-facing + Infrastructure) and which commands belong to each (D-07 LOCKED: Export / Publish / Hub trio replaces legacy "Output + Export").
+- Each `commands/<name>.md` frontmatter declares `help_jtbd:` -- a one-line "what's in it for me" outcome description.
+- The renderer joins them: groups on the outside, per-command JTBD lines on the inside.
 
-```
-  ┌──────────────────────────────────────────────────────────┐
-  │                                                          │
-  │  {R}■{X} {C}RED{X}        {M}Problem Discovery{X}                          │
-  │               {M}Find what's worth solving{X}                  │
-  │                                                          │
-  │  {B}■{X} {C}BLUE{X}       {M}Structured Thinking{X}                        │
-  │               {M}Build reasoning that holds{X}                 │
-  │                                                          │
-  │  {A}■{X} {C}AMETHYST{X}   {M}Perspectives + Creativity{X}                  │
-  │               {M}See through lenses you'd never pick{X}        │
-  │                                                          │
-  │  {Y}■{X} {C}YELLOW{X}     {M}Intelligence + Brain{X}                       │
-  │               {M}Answers from 23K nodes of teaching data{X}    │
-  │                                                          │
-  │  {G}■{X} {C}GREEN{X}      {M}Output + Export{X}                            │
-  │               {M}Ship investor-ready work{X}                   │
-  │                                                          │
-  │  {T}■{X} {C}TEAL{X}       {M}Infrastructure{X}                             │
-  │               {M}Setup, manage, maintain{X}                    │
-  │                                                          │
-  └──────────────────────────────────────────────────────────┘
-```
-
-### Count Line
+## Bulletproof contract
 
-After the legend:
-```
-  {M}67 commands -- the color tells you what kind of thinking{X}
-  {M}you're about to do before you read a single word.{X}
-```
+The renderer is bulletproof across CLI / Desktop / Cowork:
 
-### Contextual Recommendations
+- CLI with truecolor (`COLORTERM=truecolor` AND TTY): full 6-color De Stijl rendering.
+- CLI with 256color (`COLORTERM=256color` OR `TERM` contains `256color`): truecolor escapes (modern 256-color terminals accept them).
+- CLI without TTY (piped, redirected): ASCII rendering with 12-glyph vocabulary.
+- Desktop (`CLAUDE_DESKTOP=1`): ASCII rendering -- hard-override regardless of TTY.
+- Cowork: same as CLI; renderer probes the live capability per invocation.
 
-Below the count, add 2-3 lines of Larry-voice recommendations based on the current venture stage.
-
-**No room exists:**
-```
-  {M}No project yet. Start with /mos:new-project or just describe your venture.{X}
-```
-
-**Room exists but mostly empty (Pre-Opportunity):**
-```
-  {M}Room is mostly empty. /mos:beautiful-question or /mos:explore-domains to start filling it.{X}
-```
-
-**Problem defined (Discovery):**
-```
-  {M}Problem defined. /mos:analyze-needs maps your customer's real jobs.{X}
-```
-
-**Market explored (Validation):**
-```
-  {M}Market is clear. /mos:challenge-assumptions before the market does it for you.{X}
-```
-
-**Solution designed (Design):**
-```
-  {M}Strong foundation. /mos:structure-argument builds your reasoning pyramid.{X}
-```
-
-**Full coverage (Investment):**
-```
-  {M}Comprehensive room. /mos:grade for honest feedback, /mos:build-thesis for the narrative.{X}
-```
-
-### Meeting-Aware Addition
-
-If `room/meetings/` exists or user mentions meetings:
-```
-  {M}[N] meetings filed. /mos:file-meeting to add another.{X}
-```
-
-### Zone 3 -- Intelligence Strip (conditional)
-
-If room-proactive signals exist, show max 2:
-```
-  ⬜ competitive-analysis has no entries
-  ⚡ "infrastructure" converges across 3 sections
-```
-
-If no signals, omit Zone 3.
-
-### Zone 4 -- Action Footer (NEVER omit)
-
-```
-  {G}▶{X} /mos:new-project               {M}Start your first Data Room{X}
-  {M}▷{X} /mos:diagnose                  {M}Not sure? Classify your problem first{X}
-```
-
-Actions grounded in what the user likely needs based on their stage.
-
-## Step 4: Per-Command Help (`/mos:help [command]`)
+## Per-command help (`/mos:help [command]`)
 
 If the user ran `/mos:help [command]` (e.g., `/mos:help explore-domains`):
 
-Render tldr-style. NO zones. NO header panel. Just the command help:
+Render tldr-style. NO zones. NO header panel. Just the command help card:
 
 ```
-/mos:explore-domains -- Find the 2-3 domains where your innovation actually lives
+/mos:explore-domains -- Get the 5-lens decomposition of your problem domain.
 
   /mos:explore-domains                    Interactive session
   /mos:explore-domains --deep             Deep exploration (longer)
@@ -332,69 +63,58 @@ Render tldr-style. NO zones. NO header panel. Just the command help:
 ```
 
 Rules:
-- First line: command name + ` -- ` + JTBD outcome description (not tool name)
-- Examples indented 2 spaces, command left-aligned + brief annotation
-- Max 3 examples
-- No flags documentation, no option tables, no verbose descriptions
-- No zones, no header, no footer -- just the help card
+
+- First line: command name + ` -- ` + the command's `help_jtbd:` value (preferred) or `description:` fallback.
+- Examples indented 2 spaces, command left-aligned + brief annotation.
+- Max 3 examples.
+- No flags documentation, no option tables, no verbose descriptions.
+- No zones, no header, no footer -- just the help card.
 
 Load the command's `.md` file from `commands/` to get accurate description and usage patterns.
 
-**Admin visibility guard:** Before loading a command file, check its YAML frontmatter for `visibility: admin`. If the command has `visibility: admin` and `is_admin` is false, treat the command as nonexistent -- render the unknown command error below. This ensures `/mos:help admin` reveals nothing to non-admin users.
+**Admin visibility guard:** Before loading a command file, check its YAML frontmatter for `visibility: admin`. If the command has `visibility: admin` and the current user is not an admin (see admin detection below), treat the command as nonexistent -- render the unknown command error below.
 
 If the command doesn't exist (or is hidden by the visibility guard):
+
 ```
 x Unknown command: [command]
   Why: No matching /mos: command found
-  Fix: /mos:help --all
+  Fix: /mos:help
 ```
 
-## Step 5: With `--all` Flag
-
-If the user included `--all` (e.g., `/mos:help --all`):
-
-Show the **full command list** in the same color-coded tree format as Step 3 but with ALL commands listed (no truncation). Include all methodology commands explicitly.
-
-Keep the same job-category groupings and color assignments. Apply the same visibility filtering as Step 3.
-
-End with the color legend, count line, and Zone 4 footer.
+## Admin detection
 
-## Command-to-Color Mapping Reference
+Set an internal flag `is_admin` to true if ANY of these is met:
 
-Use this table to assign the correct ■ color to each command:
+1. Environment variable `MOS_ADMIN=true` is set.
+2. Username contains "jsagi" or "jonathan" (check `$USER`, `$USERNAME`, or `whoami`).
+3. Home directory matches `/home/jsagi` (check `$HOME`).
 
-### RED (Problem Discovery)
-beautiful-question, explore-domains, explore-trends, map-unknowns, analyze-needs, user-needs, root-cause, macro-trends, analyze-timing, diagnose, act
+The renderer reads visibility from each command's frontmatter; admin-gated commands are excluded by default and surface only when `is_admin` is true.
 
-### BLUE (Structured Thinking)
-structure-argument, lean-canvas, analyze-systems, systems-thinking, find-bottlenecks, validate, build-knowledge, build-thesis, grade, pipeline, grade --full
+## `--all` flag
 
-### AMETHYST (Perspectives + Creativity)
-think-hats, challenge-assumptions, scenario-plan, explore-futures, dominant-designs, find-analogies, score-innovation, persona, leadership, act --swarm, persona --parallel
-
-### YELLOW (Intelligence + Brain)
-query, graph, research, find-connections, compare-ventures, scout, opportunities, funding, suggest-next, deep-grade, diagnostics, brain-derive, file-meeting, reanalyze, speakers, research --broad
-
-### GREEN (Output + Export)
-export, present, dashboard, wiki, visualize, publish, reason
-
-### TEAL (Infrastructure)
-new-project, onboard, setup, help, status, room, rooms, organize, update, splash, admin, models
+If the user included `--all`, render the same output as the default (the renderer already includes all groups including Infrastructure; there is no truncation in the canonical layout).
 
 ## Troubleshooting
 
 If the user mentions any error, Brain issue, Pinecone quota, Neo4j connection problem, or plugin issue:
 
-1. Read `docs/TROUBLESHOOTING.md`
-2. Present the relevant fix using the 3-line error format
-3. The #1 fix for Brain issues: `rm -f .mcp.json` and restart Claude Code
+1. Read `docs/TROUBLESHOOTING.md`.
+2. Present the relevant fix using the 3-line error format.
+3. The #1 fix for Brain issues: `rm -f .mcp.json` and restart Claude Code.
 
 ## Voice Rules
 
 - Terse, structural, confident. Commands are the content.
-- **Banned phrases (per D-23):** "Great question!", "I'd be happy to help", "It's important to note", "Let me explain", sentences starting with "I", "Here's what I found"
-- Lead with structure, not commentary. The tree IS the help.
-- End with agency -- give the user a choice of what to do next via Zone 4.
-- NO EMOJI. Use only the 12 glyphs from the symbol vocabulary: ■ ▼ ▶ ▷ ├─ └─ ✓ • ⚠ ⚡ ⬜ →
-- ALL descriptions must be JTBD outcomes. Never name the framework -- describe what the user walks away with.
-- ALL commands must have a colored ■ block. Never render a command without its color.
+- Banned phrases (per D-23): "Great question!", "I'd be happy to help", "It's important to note", "Let me explain", sentences starting with "I", "Here's what I found".
+- Lead with structure, not commentary. The renderer IS the help.
+- NO EMOJI. The renderer uses only the 12-glyph vocabulary from SKILL.md Section 3.
+- ALL descriptions are JTBD outcomes (each command's `help_jtbd:` frontmatter), never names of the framework.
+
+## See also
+
+- `data/help-groups.json` -- the canonical groups + commands declaration.
+- `scripts/help-renderer.cjs` -- the bulletproof renderer.
+- `scripts/check-help-coverage.cjs` -- the CI tripwire that enforces 100% `help_jtbd` coverage + 100% group membership.
+- `lib/core/terminal-capability.cjs` -- the truecolor/ASCII probe (D-05 LOCKED; reusable across /mos:status, /mos:doctor, /mos:splash).

package/commands/hmi-status.md

@@ -1,18 +1,16 @@
 ---
 name: hmi-status
-description: Show the latest HMI compliance poll - operator-aware UI Ruling System drift summary, JTBD-priority-weighted violation list, read-only inspection (recovery via /mos:doctor --ui-compliance --fix)
+description: "[Deprecated] Show the latest HMI compliance poll (use /mos:doctor --ui-compliance --json)"
+help_jtbd: "Audit UI Ruling System compliance (deprecated: use /mos:doctor --ui-compliance --json)."
 argument-hint: "[--json]"
 body_shape: E (Action Report)
-body_shape_detail: 4-zone Shape E rendering of the side-channel at <roomDir>/.mindrian/last-hmi-poll.json (status + counts + top-5 priorities + operator-shape mismatches); --json emits raw envelope; graceful Shape E when poll absent or tier 0
 serves_jtbd: ["audit-room"]
-teaching: "When you need to know how well the UI is honoring the ruling system, /mos:hmi-status shows the latest compliance poll with operator-aware drift summaries. Read-only; fix via /mos:doctor."
-locks_operator: null
-min_tier: 0
-concurrency: sequential
-streams_events: false
-disable-model-invocation: false
+deprecated: true
+deprecated_redirect: "doctor --ui-compliance --json"
+deprecated_removal: "v1.14.0"
+teaching: "Deprecated alias. Use /mos:doctor --ui-compliance --json to audit UI Ruling System compliance; the standalone hmi-status command folds into doctor class F. Scheduled removal: v1.14.0."
 canon_parts: [3, 7, 8]
-phase: 105
+phase: 121.5-08
 ui_reference: skills/ui-system/SKILL.md
 allowed-tools:
   - Bash
@@ -21,153 +19,34 @@ allowed-tools:
 
 # /mos:hmi-status
 
-Read-only inspection surface for the Phase 105 HMI compliance poll. Renders the side-channel at `<roomDir>/.mindrian/last-hmi-poll.json` as a 4-zone Shape E (Action Report) per `skills/ui-system/SKILL.md`. The command NEVER auto-fixes - recovery is surfaced via Zone 4 action footer pointing at `/mos:doctor --ui-compliance --fix`, which the user invokes deliberately (Canon Part 3 - every choice is graph data).
+> Deprecated. /mos:hmi-status now redirects to /mos:doctor --ui-compliance --json. Scheduled removal: v1.14.0. Use /mos:doctor --ui-compliance --json going forward.
 
-The polling primitive at `scripts/hmi-compliance-poll.cjs` (Phase 105-01) is what writes the side-channel; the Stop hook at `scripts/hmi-compliance-update.cjs` (Phase 105-03) is what fires the poll on every user turn. This command only READS what those write. Per Canon Part 8 the read is purely LOCAL - no Brain queries, no network.
+You are Larry. The user invoked /mos:hmi-status. Per D-11 (LOCKED 2026-05-16, Phase 121.5-08 Sub-plan J) /mos:hmi-status is a soft-alias stub for the v1.13.x window. The canonical surface is /mos:doctor --ui-compliance --json. Doctor's class F (UI Ruling System scan, shipped Phase 106) handles all HMI-compliance logic.
 
-## What it shows
+## Steps
 
-- **Doctor status + violation counts** - aggregated from the latest `node scripts/doctor.cjs --ui-compliance --json` run, broken down per kind (missing-body-shape, unauthorized-box-char, unauthorized-glyph, renderer-missing-zone1, renderer-missing-zone4).
-- **Top 5 JTBD-weighted priorities** - violations on commands matched against the active JTBD's `methodology_hooks` are weighted 1.0; non-matches keep base weight 0.3; null JTBD uses uniform 0.5. Shape E rows show the file, weight, and matched JTBD when applicable.
-- **Operator-shape mismatches** - commands whose declared `body_shape` does not fall within the family the active operator expects (Phase 99 substrate). Informational only - never auto-rewrites a command file.
-- **Provenance footer** - timestamp + elapsed_ms + Mode A/B/0 marker so the navigator knows how fresh the poll is and which operating mode produced it.
-- **Zone 4 action footer** - 2-3 grounded /mos: commands. Primary uses `▶`, alternatives use `▷`. The canonical recovery action is `▶ /mos:doctor --ui-compliance --fix`.
+1. Emit the deprecation note above as a single cyan line (Larry voice; no em-dash; one sentence per skills/ui-system/SKILL.md Section 6).
 
-## How it works
-
-The command is a pure side-channel reader. It does NOT run the doctor. It does NOT mutate state. It does NOT call the Brain. The execution path is:
-
-1. Resolve the active room from `~/MindrianRooms/.rooms/registry.json` (Phase 83 substrate). If none: graceful Shape E pointing at `/mos:setup`.
-2. Read `<roomDir>/.mindrian/last-hmi-poll.json`. If missing: graceful Shape E with `▶ /mos:doctor --ui-compliance` + `▷ node scripts/hmi-compliance-poll.cjs --once` to force a fresh poll.
-3. Branch on `envelope.status`:
-   - `ok` -> full Shape E with Zone 2 status + priorities + mismatches + Zone 4 recovery action.
-   - `tier-0-skip` -> minimal Shape E pointing at `/mos:setup graph` (Brain unreachable AND local graph unusable).
-   - `doctor-error` -> Shape E acknowledging the doctor failure + manual retry footer.
-   - `no-active-room` -> defensive Shape E (same render as missing side-channel).
-4. `--json` short-circuits to `JSON.stringify(envelope, null, 2)` for hooks and regression tests.
-
-The script always exits 0. Read-only commands never block the user turn.
-
-To force a fresh poll (the Stop hook runs once per turn, but you can re-poll on demand):
-
-```bash
-node scripts/hmi-compliance-poll.cjs --once
-```
-
-To recover from drift after inspecting:
-
-```bash
-/mos:doctor --ui-compliance --fix
-```
-
-## Step 1: Execute
-
-Run via Bash:
-
-```bash
-node "${CLAUDE_PLUGIN_ROOT}/scripts/hmi-status-command.cjs" $ARGUMENTS
-```
-
-If `CLAUDE_PLUGIN_ROOT` is unset (older Claude Code versions), fall back to:
+2. Invoke /mos:doctor --ui-compliance --json with the user's original arguments. Run:
 
 ```bash
-node ~/.claude/plugins/mindrian-os/scripts/hmi-status-command.cjs $ARGUMENTS
-```
-
-## Step 2: Render the output
-
-The script outputs a 4-zone Shape E. Display the script's stdout directly. Do not re-format. Do not strip ANSI color codes.
-
-## Examples
-
-### Default show (status: ok with 47 violations)
-
-```
--- mindrianos -- hmi-status -- BUILD_ROOM/find-bottleneck --
-
-  ■ Doctor status: warn (47 violations)
-  ■ Counts: 47 missing-body-shape, 0 unauthorized-box-char, 0 unauthorized-glyph
-
-  ■ Top 5 priorities (JTBD-weighted)
-     ✓ commands/find-bottlenecks.md  weight 1.0  matched: find-bottleneck via methodology_hooks
-     • commands/whitespace.md  weight 0.3
-     • commands/explore-domains.md  weight 0.3
-     • commands/diagnostics.md  weight 0.3
-     • commands/find-analogies.md  weight 0.3
-
-  ■ Operator shape mismatches
-     ⚠ commands/foo.md  declared: B  operator BUILD_ROOM expects: E
-
-  (polled 2026-05-01T15:50:47Z, elapsed 187ms, mode B)
-
-  ▶ /mos:doctor --ui-compliance --fix
-  ▷ /mos:doctor --ui-compliance --json
-  ▷ /mos:hmi-status --json
-```
-
-### Default show (status: ok, zero violations)
-
-```
--- mindrianos -- hmi-status -- JUST_TALK/no-jtbd --
-
-  ■ Doctor status: ok (0 violations)
-
-  (polled 2026-05-01T15:50:47Z, elapsed 92ms, mode B)
-
-  ▷ /mos:hmi-status --json
-  ▷ /mos:doctor --all
-```
-
-### No poll yet
-
+node "${CLAUDE_PLUGIN_ROOT}/scripts/soft-alias-runner.cjs" --from hmi-status --to "doctor --ui-compliance --json" --remaining-args $ARGUMENTS
 ```
--- test-room -- hmi-status -- no-poll-yet --
 
-  ■ No poll has fired yet for this room
+The runner emits `{redirect, deprecation_note, args, ok}`. Use the redirect to confirm the target, then proceed with /mos:doctor --ui-compliance behavior. The user sees ONE deprecation note + the doctor compliance scan output.
 
-  ▶ /mos:doctor --ui-compliance
-  ▷ node scripts/hmi-compliance-poll.cjs --once
-```
-
-### --json passthrough
-
-```bash
-$ /mos:hmi-status --json
-{
-  "schema_version": 1,
-  "status": "ok",
-  "polled_at": "2026-05-01T15:50:47.000Z",
-  "operator": "BUILD_ROOM",
-  "jtbd": "find-bottleneck",
-  "tier": 1,
-  "mode": "B",
-  "doctor": { "status": "warn", "totalViolations": 47, "counts": { ... } },
-  "operator_shape_mismatches": [ ... ],
-  "priorities": [ ... ],
-  "elapsed_ms": 187,
-  "_provenance": { "phase": "105-01", "doctor_version": "1.12.3", "skill_ref": "skills/ui-system/SKILL.md" }
-}
-```
+3. Pass through doctor's --ui-compliance output verbatim. Read-only by design (per the original hmi-status contract); recovery is surfaced via /mos:doctor --ui-compliance --fix.
 
-## Voice rules
+## Why this is a soft-alias
 
-When Larry surfaces the output conversationally:
+Cluster 5 audit (2026-05-15) flagged that /mos:hmi-status was a thin read-only wrapper over doctor's class F. Two commands, one substrate. Folding hmi-status into /mos:doctor --ui-compliance --json gives one canonical compliance surface.
 
-- "47 commands missing body_shape. The top one matches your active JTBD - want to fix?"
-- "Run /mos:doctor --ui-compliance --fix to auto-recover."
-- "The poll is read-only - I never auto-fix without you saying so."
+Per D-11 the standalone command goes away (soft-alias for one release; removal v1.14.0). Scripts and hooks calling the old form keep working through v1.13.x.
 
-NEVER:
-- Run /mos:doctor --ui-compliance --fix without the user asking.
-- Re-poll silently when a fresh poll is < 5 minutes old.
-- Re-show the example blocks above when speaking conversationally.
+Per Canon Part 8: doctor's --ui-compliance class is purely LOCAL (zero network, zero Brain) -- the Part 8 invariant is preserved through the fold.
 
-## See also
+## Cross-references
 
-- `/mos:doctor --ui-compliance` - the underlying detector this command reads from.
-- `/mos:operator` (Phase 99) - the operator state stratum the priority weighting reads.
-- `/mos:jtbd` (Phase 100) - the JTBD state the weighting matches against `methodology_hooks`.
-- `skills/ui-system/SKILL.md` - the UI Ruling System spec this command's renderer self-complies with.
-- `docs/MINDRIAN-CANON.md` Part 3 - the tri-context Decision Gate. Recovery is user-driven, not auto.
-- `docs/MINDRIAN-CANON.md` Part 7 - reuse-before-build. The renderer borrows the 4-zone scaffold from `scripts/doctor.cjs renderHumanReport`.
-- `docs/MINDRIAN-CANON.md` Part 8 - graph boundary. Zero Brain queries; pure LOCAL read of side-channel + registry.
+- `commands/doctor.md` -- the canonical target; class F handles --ui-compliance.
+- `scripts/soft-alias-runner.cjs` -- the shared runner.
+- Canon Part 7 (Reuse Before Build) + Canon Part 3 (Decision Gate -- recovery is /mos:doctor --ui-compliance --fix, user-driven, not auto).