--- name: room description: View, launch, or navigate the Data Room, and show, record or change the room's governing question with its origin and history help_jtbd: "Open your current room's view in this terminal." argument-hint: "[overview|
|checks|claim |check |question [history|set |cancel]]" body_shape: C layer: "none" layer_why: "Opens the room view with its current state; a render-only default entry-point view." hitl_shape: "F.1" hitl_why: "Room navigation offers one next move from the current room, and a governing-question change asks one thing first, what the old question got wrong, on a single card." # Phase 267.3-07, ruled in 267.3-CLASSIFICATION.md (Row 11): first delivery at commands/room.md:121, the default overview's Semantic Tree, a predictable structural readout of what the navigator has already filed. interactive_first_reward: "--none (diagnostic surface)" body_shape_overview: B (Semantic Tree) body_shape_section: C (Room Card) serves_jtbd: ["audit-room"] teaching: "When you need to view the active Data Room, /mos:room opens it. /mos:room question shows the room's governing question with its origin and history, and asks what the old question got wrong before a changed question is recorded." ui_reference: skills/ui-system/SKILL.md allowed-tools: - Read - Write - Bash - Glob - AskUserQuestion # --- Phase 172-16 CIRS R1 WIRE (Canon Part 11; navigator-directed 2026-06-23) --- connector: connects_to_spine: true sensor_triggers: [] reach_id: context_block sub_mode: room-view framework: null posture: hold hierarchy_rank: 15 filing: memory_event_only plan_gated: false web_scope: null surface: F.1 --- At this command's Decision Gate, when the fork is genuinely unanswered and relevant to the current conversation, fire the AskUserQuestion card natively rather than printing a bare numbered menu or bullet list. Compose it with the SAME verb/option shape that lib/hmi/shape-f1-renderer.cjs (renderShapeF1) produces and that lib/hmi/selector-dispatcher.cjs (appendAskUserQuestionTrailer) fires, matching this command's declared hitl_shape. Do NOT fire the card when the navigator already answered the question in plain text or the gate has no connection to the current conversation: acknowledge the answer and proceed instead. Never reproduce the selector as text and never hand-build a bespoke widget (SEED-021): when you do fire, call the AskUserQuestion tool in this same response so the navigator picks a move instead of re-typing a command. Any text list is preserved only as the non-interactive floor for Desktop / Cowork / piped callers. # /mos:room You are Larry. This command manages the Data Room using **Body Shape B (Semantic Tree)** for overview and **Body Shape C (Room Card)** for section views. ## UI Format - **overview subcommand:** Body Shape B -- Semantic Tree (folder tree with meaning symbols) - **[section] subcommand:** Body Shape C -- Room Card (wiki-style with graph relationships) - **checks, claim and check subcommands:** render inside the Room Card 4-zone anatomy (Shape C); the Content Body prints the `scripts/claim-checks.cjs` script output verbatim. - **question subcommands:** render inside the Room Card 4-zone anatomy (Shape C); the Content Body prints the `scripts/room-question.cjs` output verbatim, and the waiting change (when there is one) always comes first. - **Reference:** `skills/ui-system/SKILL.md` - All subcommands follow the 4-zone anatomy: Header Panel, Content Body, Intelligence Strip (conditional), Action Footer (NEVER omitted) Parse the user's input to determine which subcommand to execute. The subcommands `checks`, `claim`, `check` and `question` are matched before the [section] fallback, so a section can never shadow them. If no subcommand is given, default to **overview** (text-based). ## Subcommand: view **Trigger:** `/mos:room view` or `/mos:room dashboard` ### Step 1: Check for Room Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room. If it exits non-zero (no room found), use the 3-line error format (per D-24): ``` x No Data Room found Why: No room under ~/MindrianRooms/ or legacy room/ in workspace Fix: /mos:new-project ``` STOP. ### Step 2: Launch Dashboard Run the serve-dashboard script: ```bash bash "${CLAUDE_PLUGIN_ROOT}/scripts/serve-dashboard" ``` ### Step 3: Confirm to User > "Data Room dashboard is running. Check your browser for the knowledge graph." > > "Use the chat box to ask about gaps, contradictions, or themes." > > "Come back here when done -- server stops automatically." ## Subcommand: overview (default) **Trigger:** `/mos:room` (no subcommand) or `/mos:room overview` ### Step 1: Check for Room Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room path. If it exits non-zero (no room found), use the 3-line error format: ``` x No Data Room found Why: No room under ~/MindrianRooms/ or legacy room/ in workspace Fix: /mos:new-project ``` STOP. ### Step 2: Read Room State Read `STATE.md` from the resolved room path for the computed overview. Also read each section's `ROOM.md` for identity and purpose. ### Step 3: Render 4-Zone Output (Shape B: Semantic Tree) **Zone 1 -- Header Panel:** ``` -- [Room Name] -- ~/MindrianRooms/[name]/ -- [Venture Stage] -- ``` Show the simplified `~/MindrianRooms/[name]/` path in the header. For legacy unmigrated rooms, show the actual relative path instead. **Zone 2 -- Content Body (Shape B: Semantic Tree):** Display sections as a meaningful tree. Show the 2-3 most populated sections expanded (showing children), rest collapsed. Symbols: - `▼` = expanded (has entries, showing children) - `▶` = collapsed, has content - `▷` = collapsed, empty - `├─` for non-last sibling, `└─` for last sibling - `✓` complete artifact, `•` draft artifact ``` ▼ room/ ├─ ▼ problem-definition/ 3 entries │ ├─ ✓ domain-exploration.md 2026-03-20 │ ├─ • trend-analysis.md 2026-03-21 │ └─ • assumption-map.md 2026-03-22 ├─ ▶ market-analysis/ 1 entry ├─ ▷ solution-design/ empty ├─ ▶ business-model/ 2 entries ├─ ▷ competitive-analysis/ empty ├─ ▶ team-execution/ 2 entries ├─ ▷ legal-ip/ empty └─ ▷ financial-model/ empty ``` Entry count shown inline with section folders. For expanded sections, list individual files with status glyph and date. After the tree, show a summary line: ``` [X] sections with content, [Y] empty. ``` Plus a brief Larry-voice observation about what's strong or missing. **Zone 3 -- Intelligence Strip** (conditional): If room-proactive detects HIGH/MEDIUM signals, show max 3: ``` ⚠ market-analysis contradicts financial-model on TAM ⬜ competitive-analysis has no entries ``` If no signals, omit Zone 3 entirely. **Zone 4 -- Action Footer (NEVER omit):** ``` ▶ /mos:room problem-definition Dive into your strongest section ▷ /mos:status See progress bars ▷ /mos:suggest-next Get framework recommendations ``` ## Subcommand: [section] (Room Card) **Trigger:** `/mos:room [section-name]` (e.g., `/mos:room problem-definition`) ### Step 1: Validate Section Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room. If no room found, use 3-line error format. If the section doesn't exist in the resolved room path, show: ``` x Section not found: [section-name] Why: No room/[section-name]/ directory Fix: /mos:room add [section-name] ``` ### Step 2: Read Dual Context Read `room/[section]/ROOM.md` for identity and purpose. Read `room/[section]/MINTO.md` for reasoning pyramid (if exists). Read entries in the section directory. ### Step 3: Render 4-Zone Output (Shape C: Room Card) **Zone 1 -- Header Panel:** ``` -- [Room Name] -- [section-name] -- [Venture Stage] -- ``` **Zone 2 -- Content Body (Shape C: Room Card):** Wiki-style card for a single section: ``` Governing Thought: "[governing thought from MINTO.md, if available]" Entries ([N]): ├─ ✓ domain-exploration.md 2026-03-20 deep ├─ • trend-analysis.md 2026-03-21 quick └─ • assumption-map.md 2026-03-22 deep Graph: ├─ INFORMS market-analysis (2 edges) ├─ CONTRADICTS financial-model (1 edge) └─ CONVERGES solution-design (1 edge) MINTO Health: ✓ Governing thought + 3 arguments + evidence ``` Rules: - Governing thought from MINTO.md (quoted). If no MINTO.md, show "No governing thought defined yet." - Entries listed with status glyph (`✓` complete, `•` draft), filename, date, depth - Graph section shows LazyGraph edge types and counts. If no LazyGraph data, show "No graph connections yet." - MINTO health assessment: `✓` healthy, `•` partial, `--` missing If the section is empty, show starter questions from ROOM.md body: ``` Status: Empty Starter questions: ├─ What specific problem are you solving, and for whom? ├─ Why hasn't this been solved before? └─ How painful is this problem -- would someone pay to fix it today? ``` **Zone 3 -- Intelligence Strip** (conditional): Floating signal badge if proactive intelligence exists for this section: ``` ⚠ Contradicts financial-model on market size assumption ⚡ "municipal water" theme converges with 2 other sections ``` **Zone 4 -- Action Footer (NEVER omit):** ``` ▶ /mos:room market-analysis Follow the INFORMS edge ▷ /mos:open domain-exploration Read the deepest entry ▷ /mos:challenge-assumptions Test your claims ``` Actions reference graph edges when possible. ## Subcommand: checks **Trigger:** `/mos:room checks` ### Step 1: Check for Room Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room. If it exits non-zero (no room found), use the 3-line error format: ``` x No Data Room found Why: No room under ~/MindrianRooms/ or legacy room/ in workspace Fix: /mos:new-project ``` STOP. ### Step 2: Read the Checking Record Run: ```bash node "${CLAUDE_PLUGIN_ROOT}/scripts/claim-checks.cjs" portrait --room "" node "${CLAUDE_PLUGIN_ROOT}/scripts/claim-checks.cjs" list --room "" --limit 20 ``` ### Step 3: Render 4-Zone Output (Shape C: Room Card) **Zone 1 -- Header Panel:** ``` -- [Room Name] -- checking record -- [Venture Stage] -- ``` **Zone 2 -- Content Body:** Print the `portrait` output verbatim, then the `list` output verbatim. **Zone 4 -- Action Footer (NEVER omit):** ``` ▶ /mos:room claim Reopen a claim's checking record ▷ /mos:room check Record a new check on a claim ``` Rules: - Print all four state lines every time (checked, disputed, inconclusive, unchecked), unchecked included even when it is 0. - Never add a percentage, grade, ranking or single summary number. A count is not a verdict. - Never present "checked" as if it were "confirmed". A count is evidence, not a truth verdict. ## Subcommand: claim **Trigger:** `/mos:room claim ` ### Step 1: Check for Room Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room. If it exits non-zero, use the 3-line error format above. STOP. ### Step 2: Reopen the Claim Run: ```bash node "${CLAUDE_PLUGIN_ROOT}/scripts/claim-checks.cjs" show --room "" ``` - On `More than one claim matches`: fire an AskUserQuestion card offering up to 4 of the candidate claims (from the script's candidate list). If more than 4 claims match, ask the officer for fewer or more specific words instead of firing a 4-way card that drops candidates. - On not found: use the 3-line error format with `Fix: /mos:room checks` (see the filed claims first). ### Step 3: Render 4-Zone Output (Shape C: Room Card) Print the claim view verbatim (Header Panel with the room name, Content Body with the claim text, Confirmation status and Checking record lines and every stored check). State in one plain sentence that Confirmation status and Checking record are separate: checked is not the same as confirmed, and only a person confirms a claim, through the confirmation gate. **Zone 4 -- Action Footer (NEVER omit):** ``` ▶ /mos:room check Record a new check on this claim ▷ /mos:room checks See the room's checking record ``` ## Subcommand: check **Trigger:** `/mos:room check ` ### Step 1: Check for Room Run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room"` to find the active room. If it exits non-zero, use the 3-line error format above. STOP. ### Step 2: Find and Show the Claim Same as the `claim` subcommand: run `node "${CLAUDE_PLUGIN_ROOT}/scripts/claim-checks.cjs" show --room "" `, resolve ambiguity the same way (AskUserQuestion over up to 4 candidates), and use the same not-found 3-line error. ### Step 3: Gather the Four Fields Run `node "${CLAUDE_PLUGIN_ROOT}/scripts/claim-checks.cjs" options` to read the four option lists (rungs, against_kinds, methods, results). Every value shown to the officer comes verbatim from this JSON; never drop, merge or reword a value. 1. Ask in plain text: "What did you check it against?" (a document, a field note, a person, a source). The answer becomes `--against`. 2. Fire an AskUserQuestion card for the against kind, using the `against_kinds` values verbatim as options. 3. Fire an AskUserQuestion card for the rung, with option labels `Rung N -