@mindrian_os/cli 2.0.0-beta.3 → 2.0.0-beta.33

package/skills/scenario-plan/SKILL.md

@@ -0,0 +1,76 @@
+---
+name: scenario-plan
+description: Build a 2x2 scenario matrix of plausible futures
+license: BSL-1.1. See LICENSE for complete terms (Business Source License 1.1, Change Date 2030-04-16 to Apache License 2.0).
+help_jtbd: "Branch into scenarios and compare outcomes."
+body_shape: "methodology"
+hitl_shape: "F.5"
+hitl_why: "The 2x2 produces four parallel scenario branches the navigator resolves among."
+serves_jtbd: ["compare-options", "plan-execution"]
+# Phase 265-13 reward-before-investment backfill (field only; grounded in the
+# shipped Session Flow: naming the two key uncertainties previews the 2x2
+# scenario matrix structure before the navigator invests in the full narrative).
+interactive_first_reward: schema_preview
+teaching: "When the future could go two different ways on two key uncertainties, /mos:scenario-plan builds the 2x2 matrix and names each quadrant. Forces you to plan for the world you do not expect."
+# --- Phase 122 workflow-layer frontmatter ---
+kind: methodology
+frameworks: ["Scenario Planning"]
+produces: "room/**/scenarios/*"
+inputs: []
+autonomous_safe: true
+allowed-tools: Read Write Bash Glob AskUserQuestion
+# --- Phase 143.3 connector frontmatter ---
+connector:
+  connects_to_spine: true
+  sensor_triggers: []
+  reach_id: context_block
+  sub_mode: scenario-plan
+  framework: "Scenario Planning"
+  posture: hold
+  hierarchy_rank: 45
+  filing: fileEvidenceWithReadback
+  plan_gated: false
+  web_scope: null
+  surface: F.2
+---
+
+<!-- mos:firing-block v2 -->
+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:firing-block -->
+
+# /mos:scenario-plan
+
+You are Larry. This command guides the user through the Scenario Planning framework (the PWS canon also calls this "Scenario Analysis" -- same tool). It is the natural next move after domain exploration: when `/mos:explore-domains` surfaces an uncertain, undefined space, scenario planning maps the plausible futures inside it.
+
+The arc follows the canonical method: Define the domain -> STEEP/PESTEL trend sweep -> pick two independent critical uncertainties -> build the 2x2 matrix -> develop vivid quadrant narratives (PARTS-tested) -> identify opportunities -> cross-scenario synthesis -> prioritize and bank -> robust strategies -> iterate. The full phase guide lives in the reference file.
+
+## Setup
+
+1. Read `${MINDRIAN_OS_ROOT:-${CLAUDE_PLUGIN_ROOT:?MindrianOS install root not found. Set MINDRIAN_OS_ROOT (see lib/core/active-plugin-root.cjs) or run from Claude Code.}}/references/methodology/scenario-plan.md` for framework details
+2. Read `${MINDRIAN_OS_ROOT:-${CLAUDE_PLUGIN_ROOT:?MindrianOS install root not found. Set MINDRIAN_OS_ROOT (see lib/core/active-plugin-root.cjs) or run from Claude Code.}}/references/personality/voice-dna.md` for Larry's voice
+3. Read `room/STATE.md` for venture context (if exists)
+
+## Session Flow
+
+Ask: "Quick pass or deep dive?"
+
+Then follow the framework phases from the reference file, adapting to the user's responses. You are NOT following a rigid script -- the phases are a guide. If the user provides rich context, skip ahead. If they need more exploration, slow down.
+
+## When Complete
+
+Create the artifact using the template from the reference file.
+Ask: "File this to market-analysis?" before writing.
+
+Then bank the harvest: the prioritized robust problems are the deliverable, not the matrix. Offer to ADD the top opportunities to the Opportunity Bank (with an HSI score and domain tags) so scenario work converts undefined futures into banked, prioritized opportunities.
+
+If the conversation reveals a connection to another methodology, suggest it. The canonical chains: scenario work feeds the Futures Wheel for second-order cascades, and the cross-scenario robust problems feed the Opportunity Bank. "This scenario work connects to /mos:research. Want to explore that next?"

package/skills/scheduled-tasks/SKILL.md

@@ -0,0 +1,301 @@
+---
+name: scheduled-tasks
+description: Define Cowork scheduled tasks for the room
+license: BSL-1.1. See LICENSE for complete terms (Business Source License 1.1, Change Date 2030-04-16 to Apache License 2.0).
+help_jtbd: "View and manage scheduled background sweeps."
+body_shape: E (Action Report)
+hitl_shape: "F.1"
+hitl_why: "Scheduled tasks offers one next move to confirm a schedule change."
+# Phase 267.3-07, ruled in 267.3-CLASSIFICATION.md (Row 15): first delivery inferred from hitl_why plus the Configuration section at commands/scheduled-tasks.md:290, a view-then-confirm shape over the plugin's own scheduler configuration.
+interactive_first_reward: "--none (diagnostic surface)"
+serves_jtbd: ["plan-execution"]
+teaching: "When you want Cowork to run something on a schedule against this room, /mos:scheduled-tasks defines the recurring job. Best for nightly grant sweeps or weekly meeting digests."
+ui_reference: skills/ui-system/SKILL.md
+surface: cowork
+allowed-tools: Read Write Glob Bash WebSearch mcp__tavily__tavily-search mcp__mindrian-brain__brain_query AskUserQuestion
+# --- Phase 172-06 CIRS R1 exclude (Canon Part 11) ---
+connector:
+  excluded: true
+  reason: "Lifecycle command. Manages scheduled / background sweeps; a scheduler-config surface run by the navigator or cron, not a contextual reach."
+---
+
+<!-- mos:firing-block v2 -->
+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:firing-block -->
+
+# Cowork Scheduled Tasks
+
+These are task definitions for Cowork's built-in scheduler. Each task runs autonomously at its configured interval, producing room artifacts in `room/intelligence/`.
+
+Users configure these in Cowork's task management UI. Larry references these definitions to register and execute scheduled work.
+
+---
+
+## Task 1: Daily Briefing
+
+**Schedule:** Every 24 hours (recommended: 6:00 AM user's timezone)
+**Trigger:** `daily-briefing`
+**Requirements:** SCHED-02
+
+### What It Does
+
+Generates a daily intelligence briefing from room state:
+- Approaching prediction deadlines (overdue, urgent, approaching)
+- New contradictions detected in intelligence files
+- Stale sections (7+ days without updates)
+- Room health summary (section coverage, artifact counts)
+
+### Execution
+
+```javascript
+const { writeBriefing } = require('lib/core/daily-briefing.cjs');
+const result = writeBriefing(roomDir);
+// result.outputPath = room/intelligence/briefing-YYYY-MM-DD.md
+```
+
+### Output
+
+Filed to: `room/intelligence/briefing-YYYY-MM-DD.md`
+
+Frontmatter includes: type, date, predictions_tracked, contradictions_found, stale_sections, generated timestamp.
+
+### Recovery
+
+If missed (Cowork was offline), the session catch-up module detects the gap and generates the briefing on next session start.
+
+---
+
+## Task 2: Prediction Deadline Tracker
+
+**Schedule:** Every 12 hours
+**Trigger:** `prediction-tracker`
+**Requirements:** SCHED-02
+
+### What It Does
+
+Scans `.predictions/REGISTRY.json` for:
+- OVERDUE predictions (deadline passed)
+- URGENT predictions (due within 3 days)
+- APPROACHING predictions (due within 14 days)
+
+Alerts are included in the daily briefing. This task runs more frequently to catch deadline transitions between briefings.
+
+### Execution
+
+```javascript
+const { getPredictionDeadlines } = require('lib/core/daily-briefing.cjs');
+const deadlines = getPredictionDeadlines(roomDir);
+const overdue = deadlines.filter(p => p.status === 'OVERDUE');
+const urgent = deadlines.filter(p => p.status === 'URGENT');
+```
+
+If overdue or urgent predictions exist, surface them via resource notification:
+> "You have N overdue predictions and M predictions due within 3 days."
+
+### Output
+
+No separate file -- results feed into the daily briefing. Notifications are surfaced to the user in the Cowork conversation.
+
+---
+
+## Task 3: Competitor Watch
+
+**Schedule:** Weekly (recommended: Monday 8:00 AM)
+**Trigger:** `competitor-watch`
+**Requirements:** SCHED-03, SCHED-07
+
+### What It Does
+
+1. Extracts tracked competitors from `room/competitive-analysis/` and `.config.json`
+2. Builds search queries for each competitor (max 5)
+3. Executes web searches for recent competitor activity
+4. Checks findings against room assumptions for contradictions
+5. Files results to `room/intelligence/competitors-YYYY-MM-DD.md`
+
+### Execution
+
+```javascript
+const { buildCompetitorQueries, fileCompetitorResults } = require('lib/core/scheduled-scanner.cjs');
+const queryPlan = buildCompetitorQueries(roomDir);
+
+if (queryPlan.insufficient) {
+  // No competitors tracked -- skip
+  return;
+}
+
+// For each query, use WebSearch or Tavily
+for (const q of queryPlan.queries) {
+  // Execute: WebSearch(q.searchQuery)
+  // Collect findings and check for contradictions against room/competitive-analysis/
+}
+
+// File results
+fileCompetitorResults(roomDir, results);
+```
+
+### Search Strategy
+
+For each competitor:
+- Query: `"[name]" funding OR launch OR pivot OR acquisition OR partnership`
+- Time range: last 30 days
+- Extract: key developments, funding rounds, product launches, pivots
+- Flag: any finding that contradicts claims in room/competitive-analysis/
+
+### Output
+
+Filed to: `room/intelligence/competitors-YYYY-MM-DD.md`
+
+Frontmatter includes: type, date, competitors_scanned, contradictions_found, generated timestamp, source.
+
+---
+
+## Task 4: Grant & Funding Discovery
+
+**Schedule:** Weekly (recommended: Wednesday 8:00 AM)
+**Trigger:** `grant-discovery`
+**Requirements:** SCHED-04, SCHED-07
+
+### What It Does
+
+1. Reads room context (domain keywords, geography, venture stage from STATE.md)
+2. Queries Grants.gov API and Simpler Grants API
+3. Scores results for relevance against room context
+4. Files top results to `room/intelligence/grants-YYYY-MM-DD.md`
+
+### Execution
+
+```javascript
+const { scanAndFileGrants } = require('lib/core/scheduled-scanner.cjs');
+const result = await scanAndFileGrants(roomDir);
+// result.outputPath, result.resultCount, result.apiErrors
+```
+
+### Prerequisites
+
+Room STATE.md should contain `domain_keywords` for accurate matching. Without keywords, the scanner falls back to problem-definition/ content.
+
+### Output
+
+Filed to: `room/intelligence/grants-YYYY-MM-DD.md`
+
+Frontmatter includes: type, date, results count, api_errors count, generated timestamp, source.
+
+---
+
+## Task 5: Domain News Scan
+
+**Schedule:** Weekly (recommended: Friday 8:00 AM)
+**Trigger:** `news-scan`
+**Requirements:** SCHED-05, SCHED-07
+
+### What It Does
+
+1. Reads domain keywords from room STATE.md
+2. Builds news queries: domain developments, regulatory updates, market/investment
+3. Executes web searches
+4. Files results to `room/intelligence/news-YYYY-MM-DD.md`
+
+### Execution
+
+```javascript
+const { buildNewsQueries, fileNewsResults } = require('lib/core/scheduled-scanner.cjs');
+const queryPlan = buildNewsQueries(roomDir);
+
+if (queryPlan.insufficient) {
+  // No domain keywords -- skip
+  return;
+}
+
+// For each query, use WebSearch or Tavily
+// Collect results per topic
+
+fileNewsResults(roomDir, results);
+```
+
+### Output
+
+Filed to: `room/intelligence/news-YYYY-MM-DD.md`
+
+Frontmatter includes: type, date, topics_scanned, items_found, generated timestamp, source.
+
+---
+
+## Task 6: Scout Sentinel
+
+**Schedule:** Weekly (recommended: Sunday midnight)
+**Trigger:** `scout-sentinel`
+**Requirements:** SCHED-06, SCHED-01
+
+### What It Does
+
+Runs the full scout cadence via the single Phase-145 composer. One runner composes all six scout sub-sensors PLUS the four SCHED-02 sensors behind the Phase-140 safe-auto-fire guard:
+1. State snapshot
+2. Health check (compare STATE.md vs last snapshot; exit captured for the HARD-01 invariant)
+3. Deadline monitor (funding/ and opportunity-bank/ deadlines)
+4. Whitespace recompute (SCHED-02)
+5. Reverse-salient (SCHED-02)
+6. Opportunity-bank scan (SCHED-02)
+7. Competitor watch (SCHED-02; emitted as a public-SIGNAL query plan, never fetched inside the runner)
+8. HSI recomputation (if dependencies available)
+
+### Execution
+
+This task invokes the single cadence runner. The Cowork scheduler owns the cadence, so `--force` bypasses the per-session throttle (the scheduler interval IS the throttle):
+
+```bash
+PLUGIN_ROOT="$(dirname "$(readlink -f "$0")")/.."
+node "${PLUGIN_ROOT}/scripts/scout-cadence-runner.cjs" "$ROOM_DIR" --force
+```
+
+The runner composes the Phase-140-hardened sentinel scripts, the HSI / whitespace / reverse-salient Python pipeline, the opportunity-bank ops, and the competitor query plan -- all behind the Phase-140 safe-auto-fire guard. It is Canon Part 8 zero-egress (inherited from Plan 01): no Brain query, no web fetch. See `commands/scout.md` (Scheduled Cadence) for the full Tri-Polar cadence model.
+
+### Output
+
+A single structured summary plus the composed sensor outputs:
+- The runner's structured summary (fired/throttled, per-step results, findings-to-graph, the competitor public-SIGNAL query plan)
+- The `safeAutoFireCheck` violations surface (HARD-01 health exit / HARD-02 NULL source_path / HARD-03 backup pollution), surfaced prominently but non-fatally
+- `room/.snapshots/STATE-YYYY-MM-DD.md`
+- `room/.intelligence/health-YYYY-MM-DD.md` (if drift detected)
+- HSI results updated in `room/.hsi-results.json`
+
+---
+
+## Tri-Polar Notes
+
+**CLI:** These tasks don't run on CLI. Users invoke `/mos:scout` manually. Session-start hook handles catch-up scanning.
+
+**Desktop:** These tasks don't run on Desktop (no scheduler). Desktop users get the session catch-up summary when reconnecting after a gap.
+
+**Cowork:** Primary surface. Tasks register in Cowork's built-in scheduler. Users configure frequency in task management UI. Missed tasks are recovered by session catch-up on next connection.
+
+---
+
+## Configuration
+
+Users can customize task schedules by setting preferences in `room/.config.json`:
+
+```json
+{
+  "scheduled_tasks": {
+    "daily_briefing": { "enabled": true, "interval_hours": 24 },
+    "prediction_tracker": { "enabled": true, "interval_hours": 12 },
+    "competitor_watch": { "enabled": true, "interval_days": 7 },
+    "grant_discovery": { "enabled": true, "interval_days": 7 },
+    "news_scan": { "enabled": true, "interval_days": 7 },
+    "scout_sentinel": { "enabled": true, "interval_days": 7 }
+  },
+  "tracked_competitors": ["Competitor A", "Competitor B"]
+}
+```
+
+All tasks are enabled by default. Users can disable individual tasks or adjust frequency.

package/skills/score-innovation/SKILL.md

@@ -0,0 +1,74 @@
+---
+name: score-innovation
+description: Score cross-domain innovation via HSI
+license: BSL-1.1. See LICENSE for complete terms (Business Source License 1.1, Change Date 2030-04-16 to Apache License 2.0).
+help_jtbd: "Score the innovation potential of your venture."
+body_shape: "methodology"
+hitl_shape: "F.8"
+hitl_why: "The six HSI components are graded independently, an unordered basket of scoring jobs."
+# Phase 267.3-04, ruled in 267.3-CLASSIFICATION.md (Row 14): first delivery at commands/score-innovation.md:66, the named intersection between two domains nobody thought to combine, judgment rather than computation.
+interactive_first_reward: methodology_reframe
+serves_jtbd: ["compare-options", "validate-idea"]
+teaching: "When you are choosing between cross-domain innovation candidates, /mos:score-innovation runs HSI scoring to rank them by semantic surprise. The math reveals which idea is actually novel."
+# --- Phase 122 workflow-layer frontmatter ---
+kind: methodology
+frameworks: ["HSI Semantic Surprise Analysis Assistant"]
+produces: "room/opportunity-bank/hsi-scores/*"
+inputs: []
+autonomous_safe: true
+allowed-tools: Read Write Bash Glob AskUserQuestion
+# --- Phase 143.3 connector frontmatter ---
+connector:
+  connects_to_spine: true
+  sensor_triggers: []
+  reach_id: context_block
+  sub_mode: hsi
+  framework: "HSI Semantic Surprise Analysis Assistant"   # MUST match the existing frameworks: value
+  posture: hold
+  hierarchy_rank: 3
+  filing: fileEvidenceWithReadback
+  plan_gated: false
+  web_scope: null
+  surface: F.1
+---
+
+<!-- mos:firing-block v2 -->
+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:firing-block -->
+
+# /mos:score-innovation
+
+You are Larry. This command guides the user through a qualitative cross-domain innovation opportunity assessment.
+
+## Setup
+
+1. Read `${MINDRIAN_OS_ROOT:-${CLAUDE_PLUGIN_ROOT:?MindrianOS install root not found. Set MINDRIAN_OS_ROOT (see lib/core/active-plugin-root.cjs) or run from Claude Code.}}/references/methodology/score-innovation.md` for the HSI framework, phases, and artifact template
+2. Read `${MINDRIAN_OS_ROOT:-${CLAUDE_PLUGIN_ROOT:?MindrianOS install root not found. Set MINDRIAN_OS_ROOT (see lib/core/active-plugin-root.cjs) or run from Claude Code.}}/references/personality/voice-dna.md` for Larry's voice
+3. Read `room/STATE.md` for venture context (if exists)
+4. Read all room sections to identify domains the user has explored
+
+## Session Flow
+
+Ask: "Quick pass or deep dive?"
+
+This is a CONVERSATIONAL assessment -- Larry guides the user through qualitative cross-domain opportunity scoring. No computation, no algorithms. Larry's judgment and Socratic questioning drive the assessment.
+
+Follow the framework phases from the reference file. The best innovations happen at the intersection of two domains nobody thought to combine.
+
+## When Complete
+
+Create the artifact using the template from the reference file.
+Ask: "File this to problem-definition?" before writing.
+
+If domain pairs reveal high-opportunity intersections, suggest:
+"The [Domain A] x [Domain B] intersection looks promising. Want to run /mos:explore-domains to map it deeper?"