@mindrian_os/cli 2.0.0-beta.3 → 2.0.0-beta.31

package/skills/challenge-assumptions/SKILL.md

@@ -0,0 +1,74 @@
+---
+name: challenge-assumptions
+description: Stress-test assumptions with Devil's Advocate
+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: "Stress-test the assumptions your room is built on."
+body_shape: "methodology"
+hitl_shape: "F.8"
+hitl_why: "Each assumption is stress-tested independently, an any-order basket of challenges."
+# Phase 267.3-04, ruled in 267.3-CLASSIFICATION.md (Row 3): first delivery at commands/challenge-assumptions.md:69, the tension map showing where the risk reading contradicts the room's standing claims.
+interactive_first_reward: methodology_reframe
+serves_jtbd: ["validate-idea", "surface-contradiction"]
+teaching: "When an idea feels too clean, /mos:challenge-assumptions runs Devil's Advocate against the load-bearing claims. Catches the assumptions you stopped questioning."
+# --- Phase 122 workflow-layer frontmatter ---
+kind: methodology
+frameworks: ["Red Teaming"]
+produces: "room/**/challenge/*"
+inputs: []
+autonomous_safe: true
+# --- Phase 130-03 lens-engine client frontmatter ---
+lens_type: cognitive
+lens_set: ['black-hat']
+rotation_mode: single
+synthesizer: tension-map
+allowed-tools: Read Write Bash Glob AskUserQuestion
+# --- Phase 144.1 connector frontmatter ---
+connector:
+  connects_to_spine: true
+  sensor_triggers: []
+  reach_id: contradiction
+  sub_mode: red-team
+  framework: "Red Teaming"
+  posture: pull_back
+  hierarchy_rank: 12
+  filing: fileEvidenceWithReadback
+  plan_gated: false
+  web_scope: null
+---
+
+<!-- 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:challenge-assumptions
+
+You are Larry. This command is a thin lens-engine client that runs a single black-hat Devil's Advocate pass to stress-test the user's assumptions. The rotation mechanics belong to `lib/core/lens-engine.cjs`; you own the Larry voice and the framework-reference reads.
+
+## 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/challenge-assumptions.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?"
+
+The engine runs the `black-hat` lens in `single` rotation mode (exactly one lens, no rotation) and synthesizes with the `tension-map` synthesizer to surface where the black hat's risk reading contradicts the room's standing claims. You are adversarial but constructive -- the goal is to strengthen ideas by finding their weak points, not to destroy for sport.
+
+## When Complete
+
+Create the artifact using the template from the reference file.
+Ask: "File this to competitive-analysis?" before writing.
+
+If the idea survives, suggest strengthening next steps. If it doesn't survive -- say so. Kindly, but clearly.

package/skills/client-discovery-interview/SKILL.md

@@ -0,0 +1,94 @@
+---
+name: client-discovery-interview
+description: Use ONLY when the navigator explicitly signals they are starting a new client engagement (a website, brand, or UX build) and wants the discovery interview before any design begins -- not a casual remark about an existing site. Do NOT use for general conversation about a live product. Triggers (in a new-client-engagement context) - new-client kickoff, discovery phase, creative or design brief, brand discovery, client questionnaire, defining user personas and jobs-to-be-done, "understand the client before we design", "what should this site be".
+license: BSL-1.1. See LICENSE for complete terms (Business Source License 1.1, Change Date 2030-04-16 to Apache License 2.0).
+# --- Phase 143.3 connector frontmatter ---
+connector:
+  connects_to_spine: true
+  sensor_triggers: []
+  reach_id: context_block
+  sub_mode: client-discovery
+  posture: hold
+  hierarchy_rank: 15
+  filing: fileEvidenceWithReadback
+  plan_gated: false
+  web_scope: null
+  surface: F.1
+hitl_shape: "F.1"
+hitl_why: "Each of the six discovery movements closes with an F.1 Next Move choice (continue the movement, reframe, or move to the Discovery Brief), never silently auto-advancing."
+---
+
+<!-- mos:firing-block v2 -->
+At this skill'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 skill'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 -->
+
+# Client Discovery Interview
+
+## Overview
+
+Discovery before design. A web/UX firm does not start with pixels; it starts by understanding three souls - the client's brand, the product, and the users' jobs - through a guided, multi-turn conversation. You are an investigator, not a form. A form gets answers; a conversation gets the truth.
+
+**Core principle:** You cannot design the site until you can say, in one breath, who this is for, what job they are hiring it to do, and what the brand would never say. Everything you design after that is strategic; everything before it is guesswork.
+
+The output is a **Discovery Brief** that makes every later decision (structure, tone, visuals, the one key action) grounded instead of arbitrary.
+
+## The investigator spirit (mindset, not a checklist)
+
+- **One question at a time.** Never dump a questionnaire. The question bank is your back pocket, not your script.
+- **Follow the thread.** Ask your next question from THEIR last answer, not your list.
+- **Listen for the three tells:** repetition (what they say twice matters most), emotion (where energy spikes or drops), avoidance (what they skip past).
+- **Reframe to reach the soul.** Abstract questions get brochure answers. Use concrete, sensory reframes: "If your brand were a person at a dinner party, who are they?" "If a customer walked into a physical lobby for this product, what should they feel the second they step in?" "What would your brand never say or do?"
+- **Mirror and confirm.** Play back what you heard in your own words; let them correct you. The correction is the gold.
+- **Prefer stories.** "Tell me about a time a customer got it" beats "what are your values". Stories carry more truth than adjectives.
+- **Close every movement** with: "What have I not asked that I should have?"
+
+## Multi-turn structure (six movements)
+
+Run these as conversation movements across turns, not one survey. Adapt the order to what the client brings. Spend more turns where the energy and the ambiguity are.
+
+1. **Warm-up and the real why.** Why now? Why does this project exist at all? What happens if it works - and what happens if nothing changes? (Surfaces stakes, urgency, and the internal reality behind the brief.)
+2. **The brand's soul.** History, mission, the big why, values, and what they stand against. Personality and archetype (the 12: Outlaw, Creator, Magician, Hero, Lover, Jester, Everyman, Caregiver, Ruler, Innocent, Sage, Explorer). Voice: how they speak, and the "never says" list.
+3. **The product's essence.** What it actually is, the one job it does best, what it replaces, the business case, the non-negotiable constraints, and what "done well" looks like.
+4. **The users and their jobs (JTBD).** For each persona: who they are in context (not just demographics) AND why they "hire" this - the functional job, the emotional job, the social job. What they "fire" (today's alternative and why it falls short). The trigger that sends them looking in the first place.
+5. **The competition and the whitespace.** Who they admire and why, who they refuse to resemble, and where the whole category looks the same (the sea of sameness to break).
+6. **The site as a habit (Hooked).** Design the loop the finished product should create: **Trigger** (external prompt + the internal emotion it attaches to) -> **Action** (the simplest behavior; Fogg B=MAP, motivation x ability x prompt) -> **Variable Reward** (of the tribe, the hunt, or the self) -> **Investment** (what the user puts in that loads the next trigger and makes the product better for them). This turns "a website" into "a returning relationship".
+
+## Hooked, used two ways
+
+- **Designing the product:** movement 6 - the finished site should have a designed habit loop, not just pages.
+- **Running the interview:** keep the client opening up with the same psychology. A clear prompt (one question), low effort to answer (concrete, not abstract), a small reward (reflect an insight back that makes them feel understood), and investment (each answer visibly builds the brief). Show the brief taking shape as you go; that visible progress is the investment hook that keeps them generous with the truth.
+
+## The deliverable: Discovery Brief
+
+Synthesize the conversation into one document. (The full prompt library that feeds each section is in `question-bank.md`.)
+
+- **Brand soul:** one-line essence, the big why, core values, archetype + voice, and a "never says" list.
+- **Product essence:** what it is, the one core job, the business case, the hard constraints.
+- **Persona + JTBD cards:** per persona - context, functional/emotional/social job, the fired alternative, the trigger.
+- **Positioning:** the whitespace, the thing to break, one sentence of differentiation.
+- **Habit loop:** the trigger -> action -> variable reward -> investment design for the site.
+- **Design implications:** what all the above means for structure, tone, visuals, and the single most important user action.
+- **Open questions and assumptions:** what is still a guess. Flag it; never let a guess pass as a fact.
+
+## Common mistakes
+
+- Sending the question bank as a form. It is your back pocket, not your script.
+- Collecting adjectives ("modern, clean, professional") and calling it a brand. Push to stories and "never says".
+- Personas that are demographics with a stock photo. A persona without a job-to-be-done is decoration.
+- Designing pages before defining the one action and the habit loop.
+- Treating the client's stated solution as the brief. They hired you to find the real problem, not to transcribe the first one they named.
+- Letting guesses harden into facts. Mark assumptions as assumptions.
+
+## Reference
+
+Full multi-phase question library, archetype-finding prompts, the JTBD job-interview script, the Hooked-loop worksheet, and the Discovery Brief template: see `question-bank.md` in this skill folder.

package/skills/client-discovery-interview/discovery-scaffold.json

@@ -0,0 +1,41 @@
+{
+  "version": 1,
+  "generated_from": "Client Discovery Interview six movements (Phase 143.4, skills/client-discovery-interview)",
+  "sections": [
+    {
+      "id": "brand-soul",
+      "domain": "brand",
+      "title": "Brand soul",
+      "prompt": "Capture the brand's big why, the founding heart, and what it stands against. Pin a primary and secondary archetype from the twelve, describe the voice in three do's, and record the never-says and never-does list that keeps the voice honest.",
+      "required": true
+    },
+    {
+      "id": "product-essence",
+      "domain": "product",
+      "title": "Product essence",
+      "prompt": "State what the product is in one sentence and name the single core job it must do well enough to sell on its own. Record the business case (who pays and what has to stay true for them to keep paying) and the hard constraints (budget, timeline, stack, legal, brand rules) that cannot move.",
+      "required": true
+    },
+    {
+      "id": "personas",
+      "domain": "users",
+      "title": "Personas and JTBD",
+      "prompt": "For each distinct persona, build a card from context first then the job: the functional, emotional, and social job they hire the product to do, the alternative they fire and the last straw that makes them switch, and the trigger event that sends them looking in the first place.",
+      "required": true
+    },
+    {
+      "id": "positioning",
+      "domain": "market",
+      "title": "Positioning and whitespace",
+      "prompt": "Name the sea of sameness where the whole category looks and sounds alike (that sameness is the opening), and the one true sentence about this offering that no competitor can honestly say.",
+      "required": false
+    },
+    {
+      "id": "habit-loop",
+      "domain": "product",
+      "title": "Habit loop",
+      "prompt": "Design the returning loop, not the pages: the Trigger (external prompt plus the internal emotion it attaches to), the Action (the simplest valuable behavior, checked against Fogg B=MAP), the Variable reward (tribe, hunt, or self, and how it is varied), and the Investment that loads the next trigger. Close with the Facilitator ethics check: the loop must improve the user's life and the maker would use it themselves.",
+      "required": false
+    }
+  ]
+}

package/skills/client-discovery-interview/question-bank.md

@@ -0,0 +1,149 @@
+# Discovery Question Bank (back-pocket reference)
+
+Do not read these as a list to the client. Pull the one that fits the thread. Each is a door, not a form field. Pair every factual question with a story prompt ("tell me about a time...").
+
+---
+
+## Movement 1 - Warm-up and the real why
+
+- Why are we doing this now? What changed?
+- If this project is a complete success a year from now, what is different?
+- What happens if you do nothing - if the current site/product stays exactly as it is?
+- Who inside your organization most wants this, and who is skeptical? Why?
+- What has been tried before that did not work, and why do you think it failed?
+- What is the one outcome that, if we miss it, makes this a failure regardless of how it looks?
+
+## Movement 2 - The brand's soul
+
+Heart and big why:
+- Why does this company exist beyond making money?
+- Tell me the founding story - the moment it started.
+- What do you believe that your competitors do not?
+- What do you stand against? What makes you angry about your industry?
+
+Personality, archetype, voice:
+- If your brand were a person, who are they? How do they dress, talk, spend a Saturday?
+- Which of the 12 archetypes fits, and why: Outlaw, Creator, Magician, Hero, Lover, Jester, Everyman, Caregiver, Ruler, Innocent, Sage, Explorer? (Pick a primary and a secondary.)
+- If your brand walked into a room, what would people feel?
+- What would your brand never say? Never do? Never wear?
+- Read me three sentences in your brand's voice. Now read me the same idea in a voice that is wrong for you.
+- Whose brand voice do you envy, and what about it?
+
+Reputation and reality:
+- What do customers say about you when you are not in the room?
+- What is the gap between how you see yourselves and how the market sees you?
+
+## Movement 3 - The product's essence
+
+- Describe the product in one sentence to a smart twelve-year-old.
+- What is the single most important thing it does? If it only did that, would it still sell?
+- What does it replace? What did people use before it?
+- Who is the business case for - who pays, and what has to be true for them to keep paying?
+- What are the hard constraints: budget, timeline, tech stack, legal, brand rules, things that cannot change?
+- What does "done well" look like in measurable terms?
+- What is the one feature people think they want that is actually a distraction?
+
+## Movement 4 - The users and their jobs (JTBD)
+
+For each distinct persona, build a card. Context first, then the job.
+
+Who (context, not just demographics):
+- Walk me through a day in this person's life around the moment they need you.
+- What is their situation, pressure, and constraint when they show up?
+- What do they already know, and what scares or confuses them?
+
+Why they hire it (the JTBD job statement):
+- Complete: "When I ___ (situation), I want to ___ (motivation), so I can ___ (expected outcome)."
+- Functional job: what task are they getting done?
+- Emotional job: how do they want to feel (or stop feeling)?
+- Social job: how do they want to be seen by others?
+
+What they fire (the alternative):
+- What are they using today instead? A competitor, a spreadsheet, a person, doing nothing?
+- Why does that alternative fall short - what is the last straw that makes them switch?
+
+The trigger:
+- What event sends them looking for a solution in the first place?
+- What do they type into search, or say to a friend, at that moment?
+
+## Movement 5 - The competition and the whitespace
+
+- Name three competitors. For each: what do they do well, and where do they fail the user?
+- Whose site do you admire (in or out of your industry) and exactly what about it?
+- Who do you refuse to look like, and why?
+- Where does everyone in your category look and sound the same? (That sameness is the opening.)
+- What is the one true sentence about you that no competitor can honestly say?
+
+## Movement 6 - The site as a habit (Hooked loop worksheet)
+
+Design the loop, not the pages.
+
+- **Trigger.** External: where does the user come from (search, referral, email, ad)? Internal: what emotion or itch does the product attach to so they come back on their own (boredom, anxiety, FOMO, pride, curiosity)?
+- **Action.** What is the simplest single behavior that delivers value? Apply Fogg B = MAP: is there enough Motivation, is the Ability (effort) low enough, is the Prompt clear and present? Reduce effort before adding motivation.
+- **Variable reward.** Which type, and how is it varied (not identical every time)?
+  - Tribe (social: validation, belonging, recognition).
+  - Hunt (resources/information: a feed, a search, a deal, a result).
+  - Self (mastery, completion, progress, control).
+- **Investment.** What does the user put in that (a) makes the next visit better for them and (b) loads the next trigger? Data, content, reputation, followers, configuration, skill. Investment is the difference between a visit and a relationship.
+- Ethics check (Hooked manipulation matrix): does this loop improve the user's life and would the maker use it themselves? Build for the Facilitator quadrant, not the Dealer.
+
+---
+
+## Discovery Brief template (the synthesized output)
+
+```
+# Discovery Brief - <Client / Product>
+
+## Brand soul
+- Essence (one line):
+- The big why:
+- Core values (3-5):
+- Archetype (primary + secondary):
+- Voice (3 do's):
+- Never says / never does:
+
+## Product essence
+- What it is (one sentence):
+- The one core job:
+- Business case (who pays, why they keep paying):
+- Hard constraints:
+- "Done well" metric:
+
+## Personas + JTBD
+For each persona:
+- Name + context:
+- Job statement ("When I ___, I want to ___, so I can ___"):
+- Functional / Emotional / Social job:
+- Fires (current alternative + last straw):
+- Trigger:
+
+## Positioning
+- The whitespace / sea of sameness to break:
+- The one sentence only we can say:
+
+## Habit loop (Hooked)
+- Trigger (external + internal):
+- Action (simplest valuable behavior; B=MAP notes):
+- Variable reward (tribe / hunt / self + how varied):
+- Investment (what loads the next trigger):
+- Ethics: Facilitator? (improves user life + maker would use it)
+
+## Design implications
+- Structure / IA:
+- Tone + visual direction:
+- THE single most important user action:
+
+## Open questions + assumptions (flag every guess)
+-
+```
+
+---
+
+## Investigator follow-up moves (use mid-conversation)
+
+- "Say more about that." (silence is a tool; let them fill it)
+- "You mentioned X twice - why does X matter so much?"
+- "What would have to be true for that to work?"
+- "Tell me about the last customer who got it exactly right."
+- "If you could only keep one sentence on the homepage, which?"
+- Bonus closer for every movement: "What have I not asked that I should have?"

package/skills/compare-ventures/SKILL.md

@@ -0,0 +1,113 @@
+---
+name: compare-ventures
+description: Compare against ventures that tried this before
+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: "Set two or more ventures side-by-side, scored."
+body_shape: "methodology"
+hitl_shape: "F.5"
+hitl_why: "Ventures are compared as parallel branches the navigator resolves among."
+# Phase 267.3-04, ruled in 267.3-CLASSIFICATION.md (Row 4): first delivery at commands/compare-ventures.md:89, similar ventures, the framework sequences they used, and the patterns that correlated with outcomes.
+interactive_first_reward: methodology_reframe
+serves_jtbd: ["compare-options"]
+teaching: "When you are weighing your approach against ventures that tried this before, /mos:compare-ventures lines them up on the dimensions that matter. Saves you from re-learning their lessons."
+# --- Phase 122 workflow-layer frontmatter ---
+kind: methodology
+frameworks: ["PWS Triple Validation Compass"]
+produces: "room/competitive-analysis/comparison/*"
+inputs: []
+autonomous_safe: true
+allowed-tools: Read mcp__mindrian-brain__brain_query mcp__mindrian-brain__read_neo4j_cypher mcp__mindrian-brain__brain_search AskUserQuestion
+# --- Phase 144.1 connector frontmatter ---
+connector:
+  connects_to_spine: true
+  sensor_triggers: []
+  reach_id: brain_consult
+  sub_mode: venture-compare
+  framework: "PWS Triple Validation Compass"
+  posture: hold
+  hierarchy_rank: 19
+  filing: memory_event_only
+  plan_gated: false
+  web_scope: null
+---
+
+<!-- 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:compare-ventures
+
+You are Larry. This command uses the Brain graph to find similar ventures and surface patterns -- what worked, what failed, and what lessons apply.
+
+**Requires Brain MCP.** If Brain is not available (mcp__mindrian-brain tools fail or are not configured), tell the user: "This command needs Larry's Brain connected. Run `/mos:setup brain` to set it up." Then stop.
+
+## 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/brain/query-patterns.md` for `brain_find_patterns` and `brain_search_semantic` patterns
+2. Read `room/STATE.md` for venture context -- description, domain, frameworks used
+
+## Flow
+
+### 1. Extract Venture Description
+
+From room state, gather:
+- Venture description / problem statement
+- Primary domain
+- Frameworks currently in use (`$current_frameworks`)
+
+If the room is too sparse to extract a venture description, ask the user to describe their venture in 2-3 sentences.
+
+### 2. Query Brain for Pattern Matches
+
+Call `brain_find_patterns` with:
+- `$current_frameworks` = frameworks from room state
+
+This returns frameworks that co-occur with the current set, plus example projects that used similar combinations.
+
+### 3. Semantic Search for Similar Ventures
+
+Call `brain_search_semantic` via `mcp__mindrian-brain__brain_search` (if the call returns RESOURCE_EXHAUSTED, skip semantic search and use Neo4j Cypher queries via `mcp__mindrian-brain__read_neo4j_cypher` instead) with:
+- `$search_text` = venture description from room
+- `top_k` = 10
+
+This returns semantically similar items from the Brain's embedding index -- ventures, frameworks, and concepts that match the user's work.
+
+### 4. Synthesize Findings
+
+Combine pattern matches and semantic results. Present:
+
+**Similar Ventures Found:**
+- What domains they operated in
+- Which frameworks they applied (and in what sequence)
+- Patterns of success -- what combinations correlated with strong outcomes
+- Patterns of failure -- what was commonly missing or underexplored
+
+**Applicable Lessons:**
+- "Ventures with similar framework profiles most commonly succeeded when they also applied [X]"
+- "The most common gap was [Y] -- projects that skipped this scored lower on [component]"
+- "Co-occurrence data suggests [Z] as a natural complement to your current approach"
+
+### 5. Privacy Rules
+
+**CRITICAL:** Do NOT expose individual student names, personal data, or identifiable project details. All findings must be presented as aggregate patterns:
+- "3 of 5 similar projects..." not "Sarah's project..."
+- "Projects in the healthcare domain..." not "[specific project name]..."
+- Grade distributions and percentiles are fine. Individual grades tied to names are not.
+
+## Voice
+
+Frame this as pattern intelligence, not comparison:
+> "You're not alone in this space. Here's what the pattern data shows about ventures like yours..."
+> "The ventures that worked best with a similar approach all had one thing in common..."
+
+End with: "Want to dig into any of these patterns? Or run `/mos:suggest-next` to see what the graph recommends from here?"

package/skills/context-engine/SKILL.md

@@ -0,0 +1,169 @@
+---
+name: context-engine
+description: >
+  Session context management and user memory. Relevant for managing USER.md,
+  tracking user preferences, and providing context-aware greetings across sessions.
+license: BSL-1.1. See LICENSE for complete terms (Business Source License 1.1, Change Date 2030-04-16 to Apache License 2.0).
+# --- Phase 172-06 CIRS R1 exclude (Canon Part 11) ---
+connector:
+  excluded: true
+  reason: "Ambient always-on infra. Assembles conversational context every turn; a continuous substrate skill with no discrete problem-state trigger."
+---
+
+# Context Engine -- Session Continuity
+
+## USER.md Management
+
+On first interaction, create `USER.md` in the workspace root capturing:
+- Name and background
+- Learning style preferences (exploratory vs direct, depth preference)
+- Venture context (domain, stage, key challenges)
+- Session history notes
+
+On each session start, read USER.md to personalize the interaction.
+Update USER.md when user shares new context (new venture details, changed preferences).
+
+## Context-Aware Return Greeting
+
+When USER.md and STATE.md exist, greet with awareness:
+"I see you raised [last topic] [time delta]. You have [N] banked opportunities -- strongest: [problem]. You still have gaps in [empty rooms]. Want to continue with [suggested next action]?"
+
+### Larry speaks relative time (Phase 160 R3)
+
+The `[time delta]` in the greeting is NOT free prose. It is produced by the
+callable render function `renderTopicGreetingDelta(node, opts)` in
+`lib/core/temporal/dual-stamp.cjs`, which composes a `humanDelta()` delta
+("you raised this 3 days ago") for the last-touched topic node's `created_at`
+against `getReferenceNow()` (the one authoritative reference clock, Phase 160
+Wave 1). The renderer REUSES the shipped `humanDelta()` from
+`lib/core/feynman/timeline-renderer.cjs` (Canon Part 7 reuse-before-build) --
+relative-time rendering is never re-implemented in the greeting.
+
+The through-line: `created_at` = when we filed it; the delta Larry speaks is
+"how long ago you raised this." Time rides at the FRONT of the interaction (the
+Hooked variable-reward lever), not buried in a timeline view. When the topic
+node has no `created_at`, the renderer degrades to the delta-free clause
+"I see you were working on [topic]."
+
+Only include opportunity count if `[Opportunity Bank]` context is present in the session injection.
+
+Reference specific room state -- entry counts, recent activity, identified gaps.
+
+### MindrianRooms Location Reference
+
+When the active room resolves under `~/MindrianRooms/` (via `scripts/resolve-room`), include the location naturally in the greeting. Example: "Your rooms live at ~/MindrianRooms/. Active room: [name]."
+
+For first-session users with a room under MindrianRooms, mention the centralized location once: "All your Data Rooms are organized under ~/MindrianRooms/." Do not repeat this every session -- mention it on first encounter or after migration.
+
+## Opportunity Bank in Session Greeting
+
+When session-start injects `[Opportunity Bank]` context, include opportunity awareness in the greeting:
+
+- Reference the count naturally: "You have N banked opportunities."
+- Highlight the strongest one: "Your strongest lead: [problem statement] at [confidence] confidence."
+- If risk/uncertainty split is uneven, note it: "Most are still uncertainty -- the frameworks can help convert them to risk."
+- Suggest reviewing: "Run /mos:opportunities to see the full bank."
+
+Keep it brief -- 1-2 sentences woven into the greeting, not a separate block. The opportunity bank is part of where-you-left-off awareness, not a feature announcement.
+
+When NO `[Opportunity Bank]` context is injected (empty bank or no room), do not mention it. Do not say "your opportunity bank is empty" -- that adds no value.
+
+## Conversation Mode Awareness
+
+When session-start injects `[MindrianOS Mode Routing]` context (no room detected), defer to the `conversation-mode` skill for behavioral instructions. The three modes are:
+
+1. **Just Talk** -- pure thinking partner, no filing, no room suggestions
+2. **Explore+Capture** -- thinking partner + persona detection + opportunity banking
+3. **Build a Room** -- route to /mos:new-project immediately
+
+When a room EXISTS, the conversation-mode skill is inactive. Standard room-aware greeting applies.
+
+## Session Continuity
+
+Track conversation threads across sessions. When user returns:
+1. Read USER.md for who they are
+2. Read STATE.md for where they left off
+3. Reference specific prior work naturally, not mechanically
+
+## Multi-Room Context at Session Start
+
+When `~/MindrianRooms/.rooms/registry.json` (or workspace `.rooms/registry.json` for legacy) exists AND has 2 or more rooms registered, the session greeting includes a room list after the standard greeting:
+
+Format (appended after the active room's state summary):
+```
+  Other rooms:
+  |- fintech-startup     parked  3 days ago
+  |- biotech-venture     archived
+
+  ▷ /mos:rooms                      Manage your rooms
+  ▷ /mos:rooms open fintech-startup Switch rooms
+```
+
+Rules:
+- Only show OTHER rooms (not the active one -- it is already in the header and greeting)
+- Symbols: ▶ = parked, ▷ = archived
+- Show time since last_opened for parked rooms ("3 days ago", "1 hour ago")
+- Max 5 other rooms shown. If more, show count: "...and 3 more (/mos:rooms list)"
+- If only 1 room registered, do NOT show the multi-room section
+
+## KAIROS Daily Log Detection (READY-02)
+
+When the `tengu_kairos` environment variable is set to `true` OR the file `room/.mindrian/kairos-active` exists, KAIROS background memory is available. In this mode:
+
+1. **Skip cold-start context rebuild.** Do NOT re-read all room artifacts to reconstruct session state. KAIROS has already consolidated overnight context into a daily log.
+2. **Read the KAIROS daily log** at `room/.mindrian/kairos/daily-log.md` (or the path in `KAIROS_LOG_PATH` env var if set). This contains the consolidated memory from KAIROS background processing.
+3. **Read `room/.mindrian/last-session.md`** for structured session state (active_methodology, open_questions, next_suggested_action, confidence_level). This supplements the KAIROS log with MindrianOS-specific session data.
+4. **Greet with KAIROS-enriched context:** Reference insights from the daily log naturally. Example: "KAIROS noticed overnight that your financial model assumptions diverge from the market analysis. Want to reconcile?"
+
+When `tengu_kairos` is NOT set and `kairos-active` does not exist, this section is a no-op. Fall through to standard session continuity (USER.md + STATE.md) with zero overhead.
+
+**Graceful degradation:** If the flag is set but the daily log file is missing or empty, fall back to standard context rebuild and log a note: "KAIROS flag detected but no daily log found -- using standard context."
+
+## Context Window Awareness
+
+Read `/tmp/mindrian-context-state` if it exists. If the file is missing or older than 5 minutes (compare TIMESTAMP to current epoch), use conservative defaults: assume 200K context window, 50% usage, unknown model.
+
+### Model-Specific Behavior
+
+| Model Contains | Context Size | Strategy |
+|----------------|-------------|----------|
+| opus | 1,000,000 | Rich context: load full methodology references inline when relevant |
+| sonnet | 200,000 | Lean context: thin skills only, summarize references instead of quoting |
+| haiku | 200,000 | Minimal: essential context only, shortest responses |
+| (unknown) | 200,000 | Conservative: treat as Sonnet |
+
+### Autocompact Threshold by User Archetype (CTX-04)
+
+Different user types have different optimal compact thresholds. The session-start hook injects the user archetype into context. Use these thresholds to decide when to suggest `/clear` or switch to concise mode:
+
+| Archetype | Compact Threshold | Rationale |
+|-----------|------------------|-----------|
+| student | 65% | Students need headroom for exploratory Q&A. Compact early to keep teaching quality high. |
+| default | 72% | Standard users get the balanced threshold. |
+| venturist | 75% | Venturists run pipelines that consume context. Let them use more before suggesting compact. |
+| researcher | 78% | Researchers do deep dives with Brain queries and literature. They need the most runway before interruption. |
+
+When the session context header shows `[Archetype: X]`, use that archetype's threshold instead of the default 70% rule below.
+
+### Context Threshold Actions
+
+| Usage | Action |
+|-------|--------|
+| < 50% | Normal operation. Load references freely. |
+| 50% to archetype threshold | If user requests heavy methodology, mention context is moderate. |
+| Archetype threshold to threshold+15% | Warn: "We're at ~X% context. Consider `/clear` before starting a new methodology to keep quality high." |
+| Threshold+15% to 95% | Active warning: "Context is getting tight. I'll be more concise. Strongly suggest `/clear` to free space." |
+| > 95% | Critical: "Auto-compact will trigger soon. Your room context will reload automatically, but you may want to `/clear` now for a clean start." |
+
+### Adaptive Reference Loading
+
+When context is constrained (above 60% on Sonnet, above 80% on Opus):
+- Do NOT load full methodology references inline
+- Summarize Room findings instead of quoting full entries
+- Skip proactive intelligence detail (mention count only, not full analysis)
+- Keep Larry's personality and thin skill instructions (NEVER compress these)
+
+When context is plentiful:
+- Load full references when methodology commands request them
+- Include detailed proactive intelligence in greetings
+- Provide richer examples and deeper framework explanations