moveros 4.0.8 → 4.1.0

package/src/workflows/walkthrough.md

@@ -1,523 +1,523 @@
----
-description: Guided interactive tour of Mover OS after setup. Explains system, folders, workflows, daily rhythm, and key features through conversation.
----
-
-# Mover OS Walkthrough (V4)
-
-**Trigger:** User runs `/walkthrough` (suggested after `/setup`).
-**Role:** The Guide (Onboarding Agent).
-**Task:** Give new users a comprehensive, interactive tour. Don't lecture — converse.
-**Risk Tier:** LOW
-**Duration:** 15-20 minutes.
-
----
-
-## 0. PRE-FLIGHT
-
-Global Rules are pre-loaded as CLAUDE.md / GEMINI.md. Output `[Rules Loaded - DATE]`.
-
-LOAD (Light — Active_Context is the cache):
-1. Active_Context.md (has Identity Snapshot S5 for personalization, Strategic Verdict S1 for goals, Active Patterns S5.5, Workflow State for predecessor detection)
-NOTE: Identity Snapshot S5 provides Archetype and Core Drive for personalizing examples. Strategic Verdict S1 provides current strategy for goal references. No need for full Identity/Strategy files — the tour needs context, not depth. If Active_Context is stale (>2 days), flag it and suggest /analyse-day first.
-
-PREDECESSOR DETECTION:
-Determine how the user got here. Check in order:
-1. **Conversation context:** Did `/setup` or `/update` run earlier in this session? (Check conversation history.)
-2. **Active_Context Workflow State:** Is `setup_date` today? → post-setup. Is there a recent `/update` entry in Active Sessions? → post-update.
-3. **Default:** Neither detected → standalone tour (user ran `/walkthrough` directly).
-
-Set PREDECESSOR to one of: `SETUP`, `UPDATE`, `STANDALONE`.
-
-**Customization by predecessor:**
-- **SETUP:** Full tour. User is brand new. Explain everything. Emphasize "this is all yours now."
-- **UPDATE:** Skip Sections 2-3 (folder structure + Engine files — they already know). Start at Section 4 (Daily Rhythm) but frame as "here's what's changed." Section 7 focuses on NEW features since their last version. Welcome: "Welcome back. Let me show you what's new."
-- **STANDALONE:** Full tour but acknowledge they may know some of this. "Stop me if you already know this."
-
-CONSTRAINTS:
-- Personalize all examples to user's Identity and Strategy
-- Keep explanations clear and non-technical (they are onboarding)
-- Do not overwhelm — explain what they need NOW, not everything at once
-- Grounded in loaded Engine files; no invented details about user
-- Windows: no && in commands
-- **INTERACTIVE:** This is a conversation, not a document dump. Wait at every STOP gate.
-
----
-
-## 1. WELCOME
-
-**If PREDECESSOR = SETUP:**
-"Your Mover OS is configured. Identity, Strategy, Dossier, Goals — all set.
-
-You've just installed something most productivity apps don't even attempt. This isn't a note-taking app. It's an execution engine that:
-- Holds you accountable to your own strategy
-- Audits what you actually did vs what you planned
-- Learns from your corrections and gets smarter over time
-
-Let me walk you through how everything works. I'll stop at key points so you can ask questions and explore your own files. About 15 minutes."
-
-**If PREDECESSOR = UPDATE:**
-"Welcome back. Mover OS has been updated. Your Engine files are untouched — Identity, Strategy, Dossier, all safe.
-
-Let me show you what's changed and what's new. This will be quick — you already know the basics. About 10 minutes."
-
-Skip directly to Section 4 (Daily Rhythm). Sections 2-3 are skipped for returning users.
-
-**If PREDECESSOR = STANDALONE:**
-"Let me give you a tour of how Mover OS works. I'll stop at key points so you can ask questions. Stop me if you already know something — about 15 minutes."
-
----
-
-## 2. YOUR FOLDER STRUCTURE
-
-```
-Your Vault/
-├── 01_Projects/              <- Active projects with timelines
-├── 02_Areas/                 <- Ongoing life areas
-│   └── Engine/               <- THE BRAIN (your core files)
-│       ├── Identity_Prime.md     <- WHO you are
-│       ├── Strategy.md           <- WHAT you're testing
-│       ├── Active_Context.md     <- WHERE you are NOW
-│       ├── Mover_Dossier.md      <- WHAT you have (assets)
-│       ├── Auto_Learnings.md     <- What AI has LEARNED about you
-│       ├── Goals.md              <- Your North Star targets
-│       ├── Metrics_Log.md        <- 30-day rolling performance data
-│       ├── Someday_Maybe.md      <- Ideas parked for later
-│       └── Dailies/              <- Daily notes (YYYY-MM/ subfolders)
-├── 03_Library/               <- Permanent knowledge base
-│   ├── MOCs/                 <- Maps of Content (indexes)
-│   ├── Principles/           <- Evergreen wisdom
-│   ├── Cheatsheets/          <- Quick reference
-│   ├── SOPs/                 <- Standard procedures
-│   ├── Entities/             <- People, Orgs, Decisions, Places
-│   ├── Inputs/               <- Articles, Books, Videos you consume
-│   └── Scripts/              <- Reusable code and utilities
-└── 04_Archives/              <- Completed/inactive projects
-```
-
-"Your `02_Areas/Engine/` folder is sacred. Those files are the brain of the system — I read them constantly to understand you and hold context across sessions."
-
----
-
-## 3. THE ENGINE FILES
-
-Let me explain each one. These are the most important files in your vault.
-
-### Identity_Prime.md — WHO You Are
-Your archetype, core drive, fatal flaw, Anti-Identity, non-negotiables. Three domains: Work, Vitality, and Faith/Purpose.
-
-"I read this before giving any advice. It's your psychological filter — it stops me from giving you generic suggestions that don't fit who you are."
-
-### Strategy.md — WHAT You're Testing
-Current business/life hypothesis, variables, metrics, sample sizes, and iteration protocol.
-
-"This changes as you learn. Think of it as your current bet. Weekly Review validates or kills hypotheses."
-
-### Active_Context.md — WHERE You Are NOW
-The live state of your life. Current phase, tactical state, waiting-for items, commitments, backlog, and session history.
-
-"This is ephemeral — it changes daily. Strategy is durable. Active_Context has sections for:
-- **Waiting For** — things you've delegated or that depend on someone else
-- **Commitments** — promises you've made to others
-- **Backlog** — tasks that surfaced mid-session but don't belong on today's plan
-- **My Customizations** — tracks your personal modifications so updates don't overwrite them
-- **Workflow State** — cross-workflow state bus (when you last ran each workflow)"
-
-### Mover_Dossier.md — WHAT You Have
-Assets (tangible + intangible), skills, survival capital, hidden assets.
-
-"I check this before suggesting strategy. No cold starts when you have ammunition."
-
-### Auto_Learnings.md — What AI Has LEARNED
-Behavioral patterns with confidence scores (1-5). AI corrections. Decays over time if not reinforced.
-
-"I track patterns and surface them when they matter. Correct me and I update my own rules."
-
-### Goals.md — Your North Star
-Three horizons: 90-day targets, 1-year vision, 10-year identity. Based on who you are, not temporary motivation.
-
-### Metrics_Log.md — Performance Data
-30-day rolling window of daily performance metrics. Fed by `/analyse-day`, consumed by `/review-week`.
-
-### Someday_Maybe.md — Idea Parking
-Ideas that are interesting but not right now. Reviewed during weekly review to see if any became relevant.
-
----
-
-**STOP GATE 1**
-
-"Those are your Engine files. Which ones are you most curious about? I can open any of them right now and show you what's inside yours."
-
-**STOP. Wait for user response. If they name a file, read it and walk them through their specific content. If they say "continue", proceed.**
-
----
-
-## 4. THE DAILY RHYTHM
-
-Here's how the system flows through a typical day. This is the core loop — everything else supports it.
-
-### Morning (Start of Session)
-**`/morning`** — 5-minute primer. Reads your Identity, checks your energy, surfaces relevant patterns from Auto_Learnings, and confirms your Single Test (the one thing that makes today a win).
-
-"You don't have to run this. But it prevents drift. It's the difference between starting intentional and starting reactive."
-
-### During Work
-**Just work.** The system runs silently in the background. When you finish a session:
-**`/log`** — Captures what happened to your Daily Note, syncs progress to plan.md, extracts lessons, detects commitments and waiting-for items.
-
-"You don't have to remember to update files. The system detects decisions, commitments, and actions from the conversation and writes them to the right places."
-
-### End of Day
-**`/analyse-day`** — Brutal audit. Compares what you actually did vs your Strategy. Detects avoidance, task drift, overrides. Updates Active_Context. Proposes corrections.
-
-**`/plan-tomorrow`** — Creates tomorrow's Battle Plan. Pulls from Strategy, backlogs, upcoming commitments, and today's analysis.
-
-### End of Week
-**`/review-week`** — The deep clean. Harvests wisdom, audits identity alignment, validates strategy, implements system improvements, decays stale patterns.
-
-"This is 30 minutes that saves hours of drift. Monthly and quarterly reviews trigger automatically when appropriate."
-
-### The Loop
-```
-/morning → [WORK] → /log → /analyse-day → /plan-tomorrow → [sleep] → /morning → ...
-```
-
-"What does your current workflow look like? Do you already have a daily rhythm, or are you starting from scratch?"
-
-**STOP. Wait for user response. Use their answer to personalize the next section — if they already have habits, connect workflows to their existing rhythm. If starting fresh, emphasize the minimum viable loop: /morning, /analyse-day, /plan-tomorrow.**
-
----
-
-## 5. ALL WORKFLOWS
-
-Here's every workflow available. You don't need to memorize them — the system suggests the next one when each completes.
-
-### Daily Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/morning` | Start of any work session | 5 min |
-| `/log` | End of work session | 5 min |
-| `/analyse-day` | Every evening (audit + context update) | 10-15 min |
-| `/plan-tomorrow` | Every evening (creates Battle Plan) | 5-10 min |
-
-### Weekly+ Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/review-week` | Weekly review (+ monthly/quarterly auto-detect) | 30 min |
-
-### Project Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/ignite` | Start a new project | 30-60 min |
-| `/ignite --feature` | Add a feature to an existing project | 20-30 min |
-| `/ignite --monetize` | Monetize a project (market validation) | 30-60 min |
-| `/overview` | Return to a project after time away | 5 min |
-| `/refactor-plan` | Update roadmap after a pivot | 10 min |
-
-### Strategic Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/pivot-strategy` | Formally change your strategy | 15 min |
-| `/debug-resistance` | When you're procrastinating or avoiding | 10 min |
-| `/reboot` | Coming back after 3+ days of inactivity | 10 min |
-
-### Knowledge Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/harvest` | Extract permanent knowledge to Library | 10 min |
-| `/debrief` | After a meeting, call, or conversation | 10-15 min |
-| `/history` | View how your Engine files evolved over time | 5 min |
-
-### System Workflows
-| Workflow | When to Use | Duration |
-|----------|-------------|----------|
-| `/setup` | First-time onboarding (you've already done this) | 60+ min |
-| `/walkthrough` | Guided tour (you're here now) | 15-20 min |
-| `/update` | After updating Mover OS to a new version | 5-10 min |
-
-**Deprecated:** `/init-plan` (merged into `/ignite`), `/project-notes` (merged into `/overview`), `/primer` (renamed to `/morning`).
-
-"Each workflow hands off to the next. You never have to wonder what's next."
-
----
-
-## 6. HOW WORKFLOWS CONNECT
-
-```
-/setup → /walkthrough → /plan-tomorrow → [WORK] → /log → /analyse-day → /plan-tomorrow
-                                                                   |
-                                                            [Review day?]
-                                                                   |
-                                                            /review-week
-
-NEW PROJECT:     /ignite → [WORK] → /log
-NEW FEATURE:     /ignite --feature → [WORK] → /log
-MONETIZE:        /ignite --monetize → Market Validation → [WORK]
-RECOVERY:        /reboot → /analyse-day
-STUCK:           /debug-resistance → [ACTION]
-PROJECT RETURN:  /overview → [WORK] → /log
-STRATEGY CHANGE: /pivot-strategy → /refactor-plan → [WORK]
-POST-MEETING:    /debrief → [WORK] (actions captured)
-SYSTEM UPDATE:   /update → /walkthrough (if major version)
-ENGINE HISTORY:  /history → (read-only, informational)
-```
-
-"You never have to wonder what's next. The system tells you."
-
-**STOP GATE 2 — Comprehension Check**
-
-"Does this flow make sense? Any questions about the daily rhythm or how workflows connect before we move on to the features that make this system different?"
-
-**STOP. Wait for user response before proceeding.**
-
----
-
-## 6.5. WHAT'S NEW (UPDATE predecessor only)
-
-**Skip this section if PREDECESSOR is SETUP or STANDALONE.**
-
-If PREDECESSOR = UPDATE, read the `/update` workflow's Section 1.5 (What's New) content that was already presented during the update process. Don't repeat it — instead summarize:
-
-"During the update, you saw the full changelog. The key things to know going forward:
-- [Top 2-3 capability changes relevant to daily use]
-- [Any workflow renames or deprecations]
-- [New workflows they should try]
-
-Want to dive into any of these, or shall I show you the features that make this system different?"
-
-**STOP. Wait for user response.**
-
----
-
-## 7. KEY FEATURES — What Makes This Different
-
-This isn't a to-do list or a note-taking app. Here's what makes Mover OS fundamentally different:
-
-### 7A. The System Challenges You
-
-**Escalating Friction** — When you drift from your plan, the system pushes back. Not walls — speed bumps:
-
-| Level | When | What happens |
-|-------|------|-------------|
-| 1 — Surface | Work doesn't match today's plan | "Your plan says X. You're working on Y. Intentional?" |
-| 2 — Justify | Work contradicts Strategy | "Why is this more important than your Single Test?" |
-| 3 — Time-box | You override Level 2 | "30 minutes on this, then back to the plan. Logging this as an override." |
-| 4 — Hard block | Destructive actions only | Refuses. Requires explicit confirmation. |
-
-You can always push through Levels 1-3. The system creates friction, not walls. But you have to do it consciously — no silent drift.
-
-**Override Logging** — Every time you override the system, it logs it. `/analyse-day` uses these to score whether your day was strategic or avoidant.
-
-### 7B. The System Learns From You
-
-**Correction Lifecycle** — When you correct the AI ("don't do that", "why did you do X?", "from now on do Y"):
-1. AI acknowledges the mistake
-2. Identifies which rule was violated and proposes a fix (in diff format)
-3. Waits for your approval
-4. If approved: implements the fix immediately and logs to Auto_Learnings.md
-5. If the same correction happens 3+ times: escalates to high confidence and flags as structural
-
-**Auto_Learnings** — Every pattern gets a confidence score (1-5). Below 3, the AI says "I don't know" instead of guessing. Patterns decay over time if not reinforced. `/review-week` reviews and cleans them.
-
-### 7C. The System Remembers
-
-**Entity Memory** — When you mention a person, organization, or place, the system can create a persistent note in `03_Library/Entities/`. Next session, it remembers who they are, your relationship, and last interaction. No more re-explaining context every time.
-
-**Passive Detection** — The system writes to files when you make decisions, commitments, or delegate tasks during conversation. You don't need to remember to update files — it detects and writes automatically:
-- "I did it" → marks task complete in plan.md
-- "I promised to send it by Friday" → writes to Active_Context Commitments
-- "Waiting on his reply" → writes to Active_Context Waiting For
-- "Scrap that idea" → writes to the Graveyard
-
-**Commitments + Waiting For** — Outbound promises and inbound dependencies are tracked in Active_Context.md. `/plan-tomorrow` checks for overdue items. Nothing falls through the cracks.
-
-### 7D. The System Runs Silently
-
-**6 Hooks (Claude Code):**
-- Session log reminder after 7+ exchanges
-- Engine file protection (confirms before modifying core files)
-- Git dirty tree guard
-- Context staleness detection
-- Transcript backup before context compaction
-- Git safety (prevents destructive commands)
-
-**Hard-Link Architecture** — Workflows exist in one place (`src/workflows/`) but are accessible from every editor simultaneously. Edit from Claude Code, Cursor, or Antigravity — same file.
-
-**13 Agents Supported** — Works best with Claude Code (full hook support, native commands, deep integration). Also supports Cursor, Cline, Windsurf, Gemini CLI, Copilot, Codex CLI, Codex Cloud, Antigravity, OpenClaw, Roo Code, Amp, and Aider.
-
----
-
-**STOP GATE 3**
-
-"Which of these four capabilities is most important to you?
-1. **Challenges me** — the friction and accountability system
-2. **Learns from me** — the correction lifecycle and pattern tracking
-3. **Remembers** — entity memory and passive detection
-4. **Runs silently** — hooks and infrastructure
-
-Pick one and I'll show you how it actually works in practice."
-
-**STOP. Wait for user response. Use their choice to select the live demo in Section 7.5.**
-
----
-
-## 7.5 LIVE DEMO
-
-Based on user's choice, walk through ONE feature interactively using their actual files:
-
-### If "Challenges me" (Escalating Friction)
-
-"Let's test it. Tell me you want to work on something that's NOT on your plan today."
-
-Wait for their response. When they name something off-plan:
-1. Apply Level 1 friction: "Your plan says [read today's Daily Note focus]. You want to work on [their thing]. Intentional?"
-2. If they push through, apply Level 2: "Why is this more important than your Single Test? One sentence."
-3. If they push through again: "Fine. 30 minutes on this, then we return to the plan. I'm logging this as an override."
-4. "That's how it works every day. Not blocking you — making you decide consciously."
-
-### If "Learns from me" (Correction Lifecycle)
-
-"Try it now. Correct me about something. Say 'from now on, always do X' or 'don't ever do Y again.'"
-
-Wait for their correction. When they provide one:
-1. Acknowledge: "I made a mistake."
-2. Identify a rule or behavior file and propose a diff fix
-3. "If you approve, I implement it immediately and log it to Auto_Learnings.md with Confidence 3. If you see me make the same mistake again, confidence goes up. At 5, it becomes a permanent structural fix."
-4. "That's the correction lifecycle. You train the system by using it."
-
-### If "Remembers" (Entity Memory + Passive Detection)
-
-"Tell me about someone you're working with or a decision you recently made."
-
-Wait for their response. When they mention a person or decision:
-1. "I'd normally create an entity note at `03_Library/Entities/People/[Name].md` with what I know about them. Next session, I'd load that context automatically."
-2. "And if you just said 'I promised [Name] I'd send the proposal by Friday' — that would automatically go into Active_Context.md under Commitments. `/plan-tomorrow` would surface it when Friday approaches."
-3. "You don't need to remember to update files. The system detects it from conversation."
-
-### If "Runs silently" (Hooks + Infrastructure)
-
-"Here's what's happening behind the scenes right now:
-1. **Session log reminder** — After 7+ exchanges, I'll remind you to run `/log` before ending
-2. **Engine protection** — If I try to edit Identity_Prime.md or Strategy.md, a hook fires and asks you to confirm
-3. **Transcript backup** — Before context compaction, the full transcript gets backed up so nothing is lost
-4. **Hard links** — The workflow files you're using right now exist in `src/workflows/` but are also accessible at `~/.claude/commands/`. Same file, two paths. Edit from either."
-
-After the demo:
-
-**STOP GATE 4**
-
-"Want to explore another feature, or are you ready to start using the system?"
-
-**STOP. Wait for response. If they want another demo, run a different one. If ready, proceed to Section 8.**
-
----
-
-## 8. TIPS FOR SUCCESS
-
-1. **Be brutally honest in logs.** The analysis is only as good as the input. "I avoided the hard task" is more useful than "productive day."
-2. **Run `/analyse-day` even on bad days.** Especially then. Bad days reveal patterns that good days hide.
-3. **Don't skip your weekly review.** 30 minutes that prevents hours of strategic drift.
-4. **Correct the AI.** Say "don't do that" or "from now on do X" — it proposes a fix, you approve it, and the system improves. This is how it gets smarter.
-5. **You don't need to update files manually.** The system detects decisions, commitments, and waiting-for items from conversation and writes them automatically. Trust the passive detection.
-6. **Update your Dossier.** New skill, connection, or asset? Add it. The AI uses it to avoid cold-start suggestions.
-7. **Garden your graph.** Connect notes with `[[Links]]`. Linked knowledge is wisdom. Unlinked knowledge is noise.
-8. **Start with 3 workflows.** `/morning`, `/analyse-day`, `/plan-tomorrow`. Master the core loop before exploring the rest.
-
----
-
-## 9. MULTI-AGENT SUPPORT
-
-Mover OS works with **13 AI coding agents.** The installer auto-detects which ones you have and configures each:
-
-| Agent | What Gets Installed |
-|-------|---------------------|
-| Claude Code | Rules + 18 commands + 38 skills + 6 hooks |
-| Cursor | Rules + 18 commands + 38 skills |
-| Cline | Rules + 38 skills |
-| Windsurf | Rules + 38 skills |
-| Gemini CLI | Rules + 18 workflows + 38 skills |
-| GitHub Copilot | Rules + 38 skills |
-| Codex CLI | Rules + 38 skills |
-| Codex (Cloud) | Rules + 38 skills |
-| Antigravity | Rules + 18 workflows |
-| OpenClaw | Rules + 38 skills |
-| Roo Code | Rules + 38 skills |
-| Amp | Rules + 38 skills |
-| Aider | Rules |
-
-"Claude Code gets the deepest integration — hooks, native slash commands, and the full correction lifecycle. Every other agent gets the core system: your rules, skills, and workflows. You can use multiple agents simultaneously — they all read the same Engine files."
-
----
-
-## 10. READY TO START
-
-"You now understand:
-- Your folder structure and where everything lives
-- The Engine files that power the system
-- The daily rhythm and how workflows connect
-- What makes this different — friction, learning, memory, and silent infrastructure
-- How to correct the AI to make it better
-
-There's one thing left to do: Start."
-
----
-
-## 11. SELF-VERIFICATION (CISC)
-
-Before completing the tour, self-check:
-1. All examples personalized to user's Identity and Strategy (not generic)? (90%+ accuracy)
-2. No invented details about user — all claims grounded in loaded files?
-3. User confirmed they understand the daily rhythm (Sections 4-6)?
-4. All STOP gates were respected (user responded before proceeding)?
-5. No banned phrases in output?
-
-If any section < 90% confident: flag it explicitly.
-
----
-
-## 12. ANSWER GATING
-
-Do not suggest handoff until:
-- User has confirmed they understand the daily rhythm (`/morning` -> work -> `/analyse-day` -> `/plan-tomorrow`)
-- User has no outstanding questions about any workflow
-- At least one STOP gate was properly respected (user responded)
-
-If pending: "You haven't confirmed you're clear on [X]. Want to clarify before we wrap up?"
-
----
-
-## 13. PROACTIVE SUGGESTIONS
-
-Based on observations during this walkthrough:
-1. If user seems overwhelmed by workflow count: "You only need 3 workflows to start: `/morning`, `/analyse-day`, `/plan-tomorrow`. Master those first." - Confidence: 3
-2. If Identity revealed a specific Anti-Identity trigger: "Your Anti-Identity ([X]) tends to emerge when [trigger]. `/debug-resistance` is specifically built for that." - Confidence: 2
-3. If Strategy shows a specific project phase: "You're in [phase]. The most relevant workflow right now is [workflow]. Want a quick demo?" - Confidence: 3
-4. If user has no active projects: "Run `/ignite` to set up your first project. That's where the system really comes alive." - Confidence: 3
-
-"Want me to address any of these?"
-
----
-
-## 14. SELF-IMPROVEMENT
-
-If user corrected AI during this workflow, execute S11 Trigger A:
-1. Acknowledge: "I made a mistake."
-2. Identify the rule violated AND the source file.
-3. Propose fix using S11 diff format with SAFETY CHECK.
-4. STOP. Wait for user approval.
-5. Log structured CORRECTION to Auto_Learnings.md.
-
----
-
-## HANDOFF
-
-Sources used: Active_Context.md (Identity Snapshot, Strategic Verdict, Active Patterns)
-Confidence range: Tour content is structural (system documentation) — personalization grounded in loaded Engine files
-
-**Next:**
-- If plan.md exists for any active project: "You already have active plans. Run `/morning` to prime your session, then get to work."
-- If no plan exists: "Run `/plan-tomorrow` to create your first Battle Plan. This is where it all begins."
-
-Key rules still active:
-- All claims must be grounded in loaded files
-- Confidence < 3 = ask, don't assert
-- Append-only: never delete history
+---
+description: Guided interactive tour of Mover OS after setup. Explains system, folders, workflows, daily rhythm, and key features through conversation.
+---
+
+# Mover OS Walkthrough (V4)
+
+**Trigger:** User runs `/walkthrough` (suggested after `/setup`).
+**Role:** The Guide (Onboarding Agent).
+**Task:** Give new users a comprehensive, interactive tour. Don't lecture — converse.
+**Risk Tier:** LOW
+**Duration:** 15-20 minutes.
+
+---
+
+## 0. PRE-FLIGHT
+
+Global Rules are pre-loaded as CLAUDE.md / GEMINI.md. Output `[Rules Loaded - DATE]`.
+
+LOAD (Light — Active_Context is the cache):
+1. Active_Context.md (has Identity Snapshot S5 for personalization, Strategic Verdict S1 for goals, Active Patterns S5.5, Workflow State for predecessor detection)
+NOTE: Identity Snapshot S5 provides Archetype and Core Drive for personalizing examples. Strategic Verdict S1 provides current strategy for goal references. No need for full Identity/Strategy files — the tour needs context, not depth. If Active_Context is stale (>2 days), flag it and suggest /analyse-day first.
+
+PREDECESSOR DETECTION:
+Determine how the user got here. Check in order:
+1. **Conversation context:** Did `/setup` or `/update` run earlier in this session? (Check conversation history.)
+2. **Active_Context Workflow State:** Is `setup_date` today? → post-setup. Is there a recent `/update` entry in Active Sessions? → post-update.
+3. **Default:** Neither detected → standalone tour (user ran `/walkthrough` directly).
+
+Set PREDECESSOR to one of: `SETUP`, `UPDATE`, `STANDALONE`.
+
+**Customization by predecessor:**
+- **SETUP:** Full tour. User is brand new. Explain everything. Emphasize "this is all yours now."
+- **UPDATE:** Skip Sections 2-3 (folder structure + Engine files — they already know). Start at Section 4 (Daily Rhythm) but frame as "here's what's changed." Section 7 focuses on NEW features since their last version. Welcome: "Welcome back. Let me show you what's new."
+- **STANDALONE:** Full tour but acknowledge they may know some of this. "Stop me if you already know this."
+
+CONSTRAINTS:
+- Personalize all examples to user's Identity and Strategy
+- Keep explanations clear and non-technical (they are onboarding)
+- Do not overwhelm — explain what they need NOW, not everything at once
+- Grounded in loaded Engine files; no invented details about user
+- Windows: no && in commands
+- **INTERACTIVE:** This is a conversation, not a document dump. Wait at every STOP gate.
+
+---
+
+## 1. WELCOME
+
+**If PREDECESSOR = SETUP:**
+"Your Mover OS is configured. Identity, Strategy, Dossier, Goals — all set.
+
+You've just installed something most productivity apps don't even attempt. This isn't a note-taking app. It's an execution engine that:
+- Holds you accountable to your own strategy
+- Audits what you actually did vs what you planned
+- Learns from your corrections and gets smarter over time
+
+Let me walk you through how everything works. I'll stop at key points so you can ask questions and explore your own files. About 15 minutes."
+
+**If PREDECESSOR = UPDATE:**
+"Welcome back. Mover OS has been updated. Your Engine files are untouched — Identity, Strategy, Dossier, all safe.
+
+Let me show you what's changed and what's new. This will be quick — you already know the basics. About 10 minutes."
+
+Skip directly to Section 4 (Daily Rhythm). Sections 2-3 are skipped for returning users.
+
+**If PREDECESSOR = STANDALONE:**
+"Let me give you a tour of how Mover OS works. I'll stop at key points so you can ask questions. Stop me if you already know something — about 15 minutes."
+
+---
+
+## 2. YOUR FOLDER STRUCTURE
+
+```
+Your Vault/
+├── 01_Projects/              <- Active projects with timelines
+├── 02_Areas/                 <- Ongoing life areas
+│   └── Engine/               <- THE BRAIN (your core files)
+│       ├── Identity_Prime.md     <- WHO you are
+│       ├── Strategy.md           <- WHAT you're testing
+│       ├── Active_Context.md     <- WHERE you are NOW
+│       ├── Mover_Dossier.md      <- WHAT you have (assets)
+│       ├── Auto_Learnings.md     <- What AI has LEARNED about you
+│       ├── Goals.md              <- Your North Star targets
+│       ├── Metrics_Log.md        <- 30-day rolling performance data
+│       ├── Someday_Maybe.md      <- Ideas parked for later
+│       └── Dailies/              <- Daily notes (YYYY-MM/ subfolders)
+├── 03_Library/               <- Permanent knowledge base
+│   ├── MOCs/                 <- Maps of Content (indexes)
+│   ├── Principles/           <- Evergreen wisdom
+│   ├── Cheatsheets/          <- Quick reference
+│   ├── SOPs/                 <- Standard procedures
+│   ├── Entities/             <- People, Orgs, Decisions, Places
+│   ├── Inputs/               <- Articles, Books, Videos you consume
+│   └── Scripts/              <- Reusable code and utilities
+└── 04_Archives/              <- Completed/inactive projects
+```
+
+"Your `02_Areas/Engine/` folder is sacred. Those files are the brain of the system — I read them constantly to understand you and hold context across sessions."
+
+---
+
+## 3. THE ENGINE FILES
+
+Let me explain each one. These are the most important files in your vault.
+
+### Identity_Prime.md — WHO You Are
+Your archetype, core drive, fatal flaw, Anti-Identity, non-negotiables. Three domains: Work, Vitality, and Faith/Purpose.
+
+"I read this before giving any advice. It's your psychological filter — it stops me from giving you generic suggestions that don't fit who you are."
+
+### Strategy.md — WHAT You're Testing
+Current business/life hypothesis, variables, metrics, sample sizes, and iteration protocol.
+
+"This changes as you learn. Think of it as your current bet. Weekly Review validates or kills hypotheses."
+
+### Active_Context.md — WHERE You Are NOW
+The live state of your life. Current phase, tactical state, waiting-for items, commitments, backlog, and session history.
+
+"This is ephemeral — it changes daily. Strategy is durable. Active_Context has sections for:
+- **Waiting For** — things you've delegated or that depend on someone else
+- **Commitments** — promises you've made to others
+- **Backlog** — tasks that surfaced mid-session but don't belong on today's plan
+- **My Customizations** — tracks your personal modifications so updates don't overwrite them
+- **Workflow State** — cross-workflow state bus (when you last ran each workflow)"
+
+### Mover_Dossier.md — WHAT You Have
+Assets (tangible + intangible), skills, survival capital, hidden assets.
+
+"I check this before suggesting strategy. No cold starts when you have ammunition."
+
+### Auto_Learnings.md — What AI Has LEARNED
+Behavioral patterns with confidence scores (1-5). AI corrections. Decays over time if not reinforced.
+
+"I track patterns and surface them when they matter. Correct me and I update my own rules."
+
+### Goals.md — Your North Star
+Three horizons: 90-day targets, 1-year vision, 10-year identity. Based on who you are, not temporary motivation.
+
+### Metrics_Log.md — Performance Data
+30-day rolling window of daily performance metrics. Fed by `/analyse-day`, consumed by `/review-week`.
+
+### Someday_Maybe.md — Idea Parking
+Ideas that are interesting but not right now. Reviewed during weekly review to see if any became relevant.
+
+---
+
+**STOP GATE 1**
+
+"Those are your Engine files. Which ones are you most curious about? I can open any of them right now and show you what's inside yours."
+
+**STOP. Wait for user response. If they name a file, read it and walk them through their specific content. If they say "continue", proceed.**
+
+---
+
+## 4. THE DAILY RHYTHM
+
+Here's how the system flows through a typical day. This is the core loop — everything else supports it.
+
+### Morning (Start of Session)
+**`/morning`** — 5-minute primer. Reads your Identity, checks your energy, surfaces relevant patterns from Auto_Learnings, and confirms your Single Test (the one thing that makes today a win).
+
+"You don't have to run this. But it prevents drift. It's the difference between starting intentional and starting reactive."
+
+### During Work
+**Just work.** The system runs silently in the background. When you finish a session:
+**`/log`** — Captures what happened to your Daily Note, syncs progress to plan.md, extracts lessons, detects commitments and waiting-for items.
+
+"You don't have to remember to update files. The system detects decisions, commitments, and actions from the conversation and writes them to the right places."
+
+### End of Day
+**`/analyse-day`** — Brutal audit. Compares what you actually did vs your Strategy. Detects avoidance, task drift, overrides. Updates Active_Context. Proposes corrections.
+
+**`/plan-tomorrow`** — Creates tomorrow's Battle Plan. Pulls from Strategy, backlogs, upcoming commitments, and today's analysis.
+
+### End of Week
+**`/review-week`** — The deep clean. Harvests wisdom, audits identity alignment, validates strategy, implements system improvements, decays stale patterns.
+
+"This is 30 minutes that saves hours of drift. Monthly and quarterly reviews trigger automatically when appropriate."
+
+### The Loop
+```
+/morning → [WORK] → /log → /analyse-day → /plan-tomorrow → [sleep] → /morning → ...
+```
+
+"What does your current workflow look like? Do you already have a daily rhythm, or are you starting from scratch?"
+
+**STOP. Wait for user response. Use their answer to personalize the next section — if they already have habits, connect workflows to their existing rhythm. If starting fresh, emphasize the minimum viable loop: /morning, /analyse-day, /plan-tomorrow.**
+
+---
+
+## 5. ALL WORKFLOWS
+
+Here's every workflow available. You don't need to memorize them — the system suggests the next one when each completes.
+
+### Daily Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/morning` | Start of any work session | 5 min |
+| `/log` | End of work session | 5 min |
+| `/analyse-day` | Every evening (audit + context update) | 10-15 min |
+| `/plan-tomorrow` | Every evening (creates Battle Plan) | 5-10 min |
+
+### Weekly+ Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/review-week` | Weekly review (+ monthly/quarterly auto-detect) | 30 min |
+
+### Project Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/ignite` | Start a new project | 30-60 min |
+| `/ignite --feature` | Add a feature to an existing project | 20-30 min |
+| `/ignite --monetize` | Monetize a project (market validation) | 30-60 min |
+| `/overview` | Return to a project after time away | 5 min |
+| `/refactor-plan` | Update roadmap after a pivot | 10 min |
+
+### Strategic Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/pivot-strategy` | Formally change your strategy | 15 min |
+| `/debug-resistance` | When you're procrastinating or avoiding | 10 min |
+| `/reboot` | Coming back after 3+ days of inactivity | 10 min |
+
+### Knowledge Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/harvest` | Extract permanent knowledge to Library | 10 min |
+| `/debrief` | After a meeting, call, or conversation | 10-15 min |
+| `/history` | View how your Engine files evolved over time | 5 min |
+
+### System Workflows
+| Workflow | When to Use | Duration |
+|----------|-------------|----------|
+| `/setup` | First-time onboarding (you've already done this) | 60+ min |
+| `/walkthrough` | Guided tour (you're here now) | 15-20 min |
+| `/update` | After updating Mover OS to a new version | 5-10 min |
+
+**Deprecated:** `/init-plan` (merged into `/ignite`), `/project-notes` (merged into `/overview`), `/primer` (renamed to `/morning`).
+
+"Each workflow hands off to the next. You never have to wonder what's next."
+
+---
+
+## 6. HOW WORKFLOWS CONNECT
+
+```
+/setup → /walkthrough → /plan-tomorrow → [WORK] → /log → /analyse-day → /plan-tomorrow
+                                                                   |
+                                                            [Review day?]
+                                                                   |
+                                                            /review-week
+
+NEW PROJECT:     /ignite → [WORK] → /log
+NEW FEATURE:     /ignite --feature → [WORK] → /log
+MONETIZE:        /ignite --monetize → Market Validation → [WORK]
+RECOVERY:        /reboot → /analyse-day
+STUCK:           /debug-resistance → [ACTION]
+PROJECT RETURN:  /overview → [WORK] → /log
+STRATEGY CHANGE: /pivot-strategy → /refactor-plan → [WORK]
+POST-MEETING:    /debrief → [WORK] (actions captured)
+SYSTEM UPDATE:   /update → /walkthrough (if major version)
+ENGINE HISTORY:  /history → (read-only, informational)
+```
+
+"You never have to wonder what's next. The system tells you."
+
+**STOP GATE 2 — Comprehension Check**
+
+"Does this flow make sense? Any questions about the daily rhythm or how workflows connect before we move on to the features that make this system different?"
+
+**STOP. Wait for user response before proceeding.**
+
+---
+
+## 6.5. WHAT'S NEW (UPDATE predecessor only)
+
+**Skip this section if PREDECESSOR is SETUP or STANDALONE.**
+
+If PREDECESSOR = UPDATE, read the `/update` workflow's Section 1.5 (What's New) content that was already presented during the update process. Don't repeat it — instead summarize:
+
+"During the update, you saw the full changelog. The key things to know going forward:
+- [Top 2-3 capability changes relevant to daily use]
+- [Any workflow renames or deprecations]
+- [New workflows they should try]
+
+Want to dive into any of these, or shall I show you the features that make this system different?"
+
+**STOP. Wait for user response.**
+
+---
+
+## 7. KEY FEATURES — What Makes This Different
+
+This isn't a to-do list or a note-taking app. Here's what makes Mover OS fundamentally different:
+
+### 7A. The System Challenges You
+
+**Escalating Friction** — When you drift from your plan, the system pushes back. Not walls — speed bumps:
+
+| Level | When | What happens |
+|-------|------|-------------|
+| 1 — Surface | Work doesn't match today's plan | "Your plan says X. You're working on Y. Intentional?" |
+| 2 — Justify | Work contradicts Strategy | "Why is this more important than your Single Test?" |
+| 3 — Time-box | You override Level 2 | "30 minutes on this, then back to the plan. Logging this as an override." |
+| 4 — Hard block | Destructive actions only | Refuses. Requires explicit confirmation. |
+
+You can always push through Levels 1-3. The system creates friction, not walls. But you have to do it consciously — no silent drift.
+
+**Override Logging** — Every time you override the system, it logs it. `/analyse-day` uses these to score whether your day was strategic or avoidant.
+
+### 7B. The System Learns From You
+
+**Correction Lifecycle** — When you correct the AI ("don't do that", "why did you do X?", "from now on do Y"):
+1. AI acknowledges the mistake
+2. Identifies which rule was violated and proposes a fix (in diff format)
+3. Waits for your approval
+4. If approved: implements the fix immediately and logs to Auto_Learnings.md
+5. If the same correction happens 3+ times: escalates to high confidence and flags as structural
+
+**Auto_Learnings** — Every pattern gets a confidence score (1-5). Below 3, the AI says "I don't know" instead of guessing. Patterns decay over time if not reinforced. `/review-week` reviews and cleans them.
+
+### 7C. The System Remembers
+
+**Entity Memory** — When you mention a person, organization, or place, the system can create a persistent note in `03_Library/Entities/`. Next session, it remembers who they are, your relationship, and last interaction. No more re-explaining context every time.
+
+**Passive Detection** — The system writes to files when you make decisions, commitments, or delegate tasks during conversation. You don't need to remember to update files — it detects and writes automatically:
+- "I did it" → marks task complete in plan.md
+- "I promised to send it by Friday" → writes to Active_Context Commitments
+- "Waiting on his reply" → writes to Active_Context Waiting For
+- "Scrap that idea" → writes to the Graveyard
+
+**Commitments + Waiting For** — Outbound promises and inbound dependencies are tracked in Active_Context.md. `/plan-tomorrow` checks for overdue items. Nothing falls through the cracks.
+
+### 7D. The System Runs Silently
+
+**6 Hooks (Claude Code):**
+- Session log reminder after 7+ exchanges
+- Engine file protection (confirms before modifying core files)
+- Git dirty tree guard
+- Context staleness detection
+- Transcript backup before context compaction
+- Git safety (prevents destructive commands)
+
+**Hard-Link Architecture** — Workflows exist in one place (`src/workflows/`) but are accessible from every editor simultaneously. Edit from Claude Code, Cursor, or Antigravity — same file.
+
+**13 Agents Supported** — Works best with Claude Code (full hook support, native commands, deep integration). Also supports Cursor, Cline, Windsurf, Gemini CLI, Copilot, Codex CLI, Codex Cloud, Antigravity, OpenClaw, Roo Code, Amp, and Aider.
+
+---
+
+**STOP GATE 3**
+
+"Which of these four capabilities is most important to you?
+1. **Challenges me** — the friction and accountability system
+2. **Learns from me** — the correction lifecycle and pattern tracking
+3. **Remembers** — entity memory and passive detection
+4. **Runs silently** — hooks and infrastructure
+
+Pick one and I'll show you how it actually works in practice."
+
+**STOP. Wait for user response. Use their choice to select the live demo in Section 7.5.**
+
+---
+
+## 7.5 LIVE DEMO
+
+Based on user's choice, walk through ONE feature interactively using their actual files:
+
+### If "Challenges me" (Escalating Friction)
+
+"Let's test it. Tell me you want to work on something that's NOT on your plan today."
+
+Wait for their response. When they name something off-plan:
+1. Apply Level 1 friction: "Your plan says [read today's Daily Note focus]. You want to work on [their thing]. Intentional?"
+2. If they push through, apply Level 2: "Why is this more important than your Single Test? One sentence."
+3. If they push through again: "Fine. 30 minutes on this, then we return to the plan. I'm logging this as an override."
+4. "That's how it works every day. Not blocking you — making you decide consciously."
+
+### If "Learns from me" (Correction Lifecycle)
+
+"Try it now. Correct me about something. Say 'from now on, always do X' or 'don't ever do Y again.'"
+
+Wait for their correction. When they provide one:
+1. Acknowledge: "I made a mistake."
+2. Identify a rule or behavior file and propose a diff fix
+3. "If you approve, I implement it immediately and log it to Auto_Learnings.md with Confidence 3. If you see me make the same mistake again, confidence goes up. At 5, it becomes a permanent structural fix."
+4. "That's the correction lifecycle. You train the system by using it."
+
+### If "Remembers" (Entity Memory + Passive Detection)
+
+"Tell me about someone you're working with or a decision you recently made."
+
+Wait for their response. When they mention a person or decision:
+1. "I'd normally create an entity note at `03_Library/Entities/People/[Name].md` with what I know about them. Next session, I'd load that context automatically."
+2. "And if you just said 'I promised [Name] I'd send the proposal by Friday' — that would automatically go into Active_Context.md under Commitments. `/plan-tomorrow` would surface it when Friday approaches."
+3. "You don't need to remember to update files. The system detects it from conversation."
+
+### If "Runs silently" (Hooks + Infrastructure)
+
+"Here's what's happening behind the scenes right now:
+1. **Session log reminder** — After 7+ exchanges, I'll remind you to run `/log` before ending
+2. **Engine protection** — If I try to edit Identity_Prime.md or Strategy.md, a hook fires and asks you to confirm
+3. **Transcript backup** — Before context compaction, the full transcript gets backed up so nothing is lost
+4. **Hard links** — The workflow files you're using right now exist in `src/workflows/` but are also accessible at `~/.claude/commands/`. Same file, two paths. Edit from either."
+
+After the demo:
+
+**STOP GATE 4**
+
+"Want to explore another feature, or are you ready to start using the system?"
+
+**STOP. Wait for response. If they want another demo, run a different one. If ready, proceed to Section 8.**
+
+---
+
+## 8. TIPS FOR SUCCESS
+
+1. **Be brutally honest in logs.** The analysis is only as good as the input. "I avoided the hard task" is more useful than "productive day."
+2. **Run `/analyse-day` even on bad days.** Especially then. Bad days reveal patterns that good days hide.
+3. **Don't skip your weekly review.** 30 minutes that prevents hours of strategic drift.
+4. **Correct the AI.** Say "don't do that" or "from now on do X" — it proposes a fix, you approve it, and the system improves. This is how it gets smarter.
+5. **You don't need to update files manually.** The system detects decisions, commitments, and waiting-for items from conversation and writes them automatically. Trust the passive detection.
+6. **Update your Dossier.** New skill, connection, or asset? Add it. The AI uses it to avoid cold-start suggestions.
+7. **Garden your graph.** Connect notes with `[[Links]]`. Linked knowledge is wisdom. Unlinked knowledge is noise.
+8. **Start with 3 workflows.** `/morning`, `/analyse-day`, `/plan-tomorrow`. Master the core loop before exploring the rest.
+
+---
+
+## 9. MULTI-AGENT SUPPORT
+
+Mover OS works with **13 AI coding agents.** The installer auto-detects which ones you have and configures each:
+
+| Agent | What Gets Installed |
+|-------|---------------------|
+| Claude Code | Rules + 18 commands + 38 skills + 6 hooks |
+| Cursor | Rules + 18 commands + 38 skills |
+| Cline | Rules + 38 skills |
+| Windsurf | Rules + 38 skills |
+| Gemini CLI | Rules + 18 workflows + 38 skills |
+| GitHub Copilot | Rules + 38 skills |
+| Codex CLI | Rules + 38 skills |
+| Codex (Cloud) | Rules + 38 skills |
+| Antigravity | Rules + 18 workflows |
+| OpenClaw | Rules + 38 skills |
+| Roo Code | Rules + 38 skills |
+| Amp | Rules + 38 skills |
+| Aider | Rules |
+
+"Claude Code gets the deepest integration — hooks, native slash commands, and the full correction lifecycle. Every other agent gets the core system: your rules, skills, and workflows. You can use multiple agents simultaneously — they all read the same Engine files."
+
+---
+
+## 10. READY TO START
+
+"You now understand:
+- Your folder structure and where everything lives
+- The Engine files that power the system
+- The daily rhythm and how workflows connect
+- What makes this different — friction, learning, memory, and silent infrastructure
+- How to correct the AI to make it better
+
+There's one thing left to do: Start."
+
+---
+
+## 11. SELF-VERIFICATION (CISC)
+
+Before completing the tour, self-check:
+1. All examples personalized to user's Identity and Strategy (not generic)? (90%+ accuracy)
+2. No invented details about user — all claims grounded in loaded files?
+3. User confirmed they understand the daily rhythm (Sections 4-6)?
+4. All STOP gates were respected (user responded before proceeding)?
+5. No banned phrases in output?
+
+If any section < 90% confident: flag it explicitly.
+
+---
+
+## 12. ANSWER GATING
+
+Do not suggest handoff until:
+- User has confirmed they understand the daily rhythm (`/morning` -> work -> `/analyse-day` -> `/plan-tomorrow`)
+- User has no outstanding questions about any workflow
+- At least one STOP gate was properly respected (user responded)
+
+If pending: "You haven't confirmed you're clear on [X]. Want to clarify before we wrap up?"
+
+---
+
+## 13. PROACTIVE SUGGESTIONS
+
+Based on observations during this walkthrough:
+1. If user seems overwhelmed by workflow count: "You only need 3 workflows to start: `/morning`, `/analyse-day`, `/plan-tomorrow`. Master those first." - Confidence: 3
+2. If Identity revealed a specific Anti-Identity trigger: "Your Anti-Identity ([X]) tends to emerge when [trigger]. `/debug-resistance` is specifically built for that." - Confidence: 2
+3. If Strategy shows a specific project phase: "You're in [phase]. The most relevant workflow right now is [workflow]. Want a quick demo?" - Confidence: 3
+4. If user has no active projects: "Run `/ignite` to set up your first project. That's where the system really comes alive." - Confidence: 3
+
+"Want me to address any of these?"
+
+---
+
+## 14. SELF-IMPROVEMENT
+
+If user corrected AI during this workflow, execute S11 Trigger A:
+1. Acknowledge: "I made a mistake."
+2. Identify the rule violated AND the source file.
+3. Propose fix using S11 diff format with SAFETY CHECK.
+4. STOP. Wait for user approval.
+5. Log structured CORRECTION to Auto_Learnings.md.
+
+---
+
+## HANDOFF
+
+Sources used: Active_Context.md (Identity Snapshot, Strategic Verdict, Active Patterns)
+Confidence range: Tour content is structural (system documentation) — personalization grounded in loaded Engine files
+
+**Next:**
+- If plan.md exists for any active project: "You already have active plans. Run `/morning` to prime your session, then get to work."
+- If no plan exists: "Run `/plan-tomorrow` to create your first Battle Plan. This is where it all begins."
+
+Key rules still active:
+- All claims must be grounded in loaded files
+- Confidence < 3 = ask, don't assert
+- Append-only: never delete history