create-pmos 0.523.0 → 0.524.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-pmos",
-  "version": "0.523.0",
+  "version": "0.524.0",
   "description": "Scaffold the PMOS portable kit — run the PMOS product-management workflow (initiative → lane-fit spec → contract → build → three gates, with work-type lanes for features, research, maintenance, incidents, and content) entirely inside your own repo, with all state as git-versioned files. Includes `workflow-sweep`, which audits the state every board derives from and separates real defects from work you simply have not done. `create-pmos workspace` also installs + launches the local Terminal Workspace GUI. No database, no network, no telemetry.",
   "license": "MIT",
   "bin": {
@@ -32,6 +32,49 @@
     "audit"
   ],
   "pmosChanges": [
+    {
+      "v": "0.524.0",
+      "at": "2026-11-04T08:57:48.087Z",
+      "f": 4,
+      "prs": [
+        {
+          "n": 1224,
+          "t": "Setup as one conversation, with the fundamentals settled (onboarding B1)",
+          "s": "Setup as one conversation, with the fundamentals settled (onboarding B1)",
+          "k": "new"
+        },
+        {
+          "n": 1223,
+          "t": "A feature appears when its moment comes (onboarding B3, gradual exposure)",
+          "s": "A feature appears when its moment comes (onboarding B3, gradual exposure)",
+          "k": "new"
+        },
+        {
+          "n": 1222,
+          "t": "After setup, Home shows what it did, the product portrait and the drafted first initiative (onboarding B2)",
+          "s": "After setup, Home shows what it did, the product portrait and the drafted first initiative (onboarding B2)",
+          "k": "new"
+        },
+        {
+          "n": 1227,
+          "t": "Idle lab screens: ✦ cards, compare view and Intelligence › Lab (build U)",
+          "s": "Idle lab screens: ✦ cards, compare view and Intelligence › Lab (build U)",
+          "k": "new"
+        },
+        {
+          "n": 1234,
+          "t": "Good morning reads the quarter — goal states, the PM's line, snapshot, goal tags, Yes effect (GM1)",
+          "s": "Good morning reads the quarter — goal states, the PM's line, snapshot, goal tags, Yes effect (GM1)",
+          "k": "new"
+        },
+        {
+          "n": 1228,
+          "t": "One queue for all background work (idle lab build Q)",
+          "s": "One queue for all background work (idle lab build Q)",
+          "k": "new"
+        }
+      ]
+    },
     {
       "v": "0.523.0",
       "at": "2026-11-04T08:06:51.586Z",
@@ -371,21 +414,6 @@
           "d": "A recording could only be finished from the tab that started it. Now any tab that shows it can finish it."
         }
       ]
-    },
-    {
-      "v": "0.504.0",
-      "at": "2026-11-03T06:02:59.187Z",
-      "f": 4,
-      "prs": [
-        {
-          "n": 1189,
-          "t": "The Working card ends when Claude's answer ends",
-          "s": "The Working card ends when Claude's answer ends",
-          "k": "fix",
-          "h": "The Working card ends on time",
-          "d": "The Working card could keep counting after Claude had answered. Now it goes away when the answer ends."
-        }
-      ]
     }
   ]
 }

package/payload/KIT-VERSION

@@ -1,3 +1,3 @@
-kernel-source: pmos@45be2110
+kernel-source: pmos@fbb375e0
 assembled: 2026-11-04
 mode: vendor

package/payload/planning/evals/eval-calibration.md

@@ -2,7 +2,7 @@
 type: Reference
 title: "Eval calibration log"
 description: "Append-only record of every PM override of an evaluator verdict."
-generated: {by: process:create-pmos, at: "2026-11-04T08:06:47Z"}
+generated: {by: process:create-pmos, at: "2026-11-04T08:57:44Z"}
 status: stable
 ---
 

package/payload/planning/okrs/health-budgets.md

@@ -2,7 +2,7 @@
 type: Reference
 title: "Health budgets"
 description: "Closed registry of health-budget anchors (D59), edited only by the PM."
-generated: {by: process:create-pmos, at: "2026-11-04T08:06:47Z"}
+generated: {by: process:create-pmos, at: "2026-11-04T08:57:44Z"}
 status: stable
 ---
 

package/payload/state/discoveries.md

@@ -2,7 +2,7 @@
 type: Reference
 title: "Discoveries and friction log"
 description: "Append-only record of what runs discovered and where the process rubbed."
-generated: {by: process:create-pmos, at: "2026-11-04T08:06:47Z"}
+generated: {by: process:create-pmos, at: "2026-11-04T08:57:44Z"}
 status: stable
 ---
 

package/workspace/changelog.json

@@ -1,4 +1,47 @@
 [
+ {
+  "v": "0.524.0",
+  "at": "2026-11-04T08:57:48.087Z",
+  "f": 4,
+  "prs": [
+   {
+    "n": 1224,
+    "t": "Setup as one conversation, with the fundamentals settled (onboarding B1)",
+    "s": "Setup as one conversation, with the fundamentals settled (onboarding B1)",
+    "k": "new"
+   },
+   {
+    "n": 1223,
+    "t": "A feature appears when its moment comes (onboarding B3, gradual exposure)",
+    "s": "A feature appears when its moment comes (onboarding B3, gradual exposure)",
+    "k": "new"
+   },
+   {
+    "n": 1222,
+    "t": "After setup, Home shows what it did, the product portrait and the drafted first initiative (onboarding B2)",
+    "s": "After setup, Home shows what it did, the product portrait and the drafted first initiative (onboarding B2)",
+    "k": "new"
+   },
+   {
+    "n": 1227,
+    "t": "Idle lab screens: ✦ cards, compare view and Intelligence › Lab (build U)",
+    "s": "Idle lab screens: ✦ cards, compare view and Intelligence › Lab (build U)",
+    "k": "new"
+   },
+   {
+    "n": 1234,
+    "t": "Good morning reads the quarter — goal states, the PM's line, snapshot, goal tags, Yes effect (GM1)",
+    "s": "Good morning reads the quarter — goal states, the PM's line, snapshot, goal tags, Yes effect (GM1)",
+    "k": "new"
+   },
+   {
+    "n": 1228,
+    "t": "One queue for all background work (idle lab build Q)",
+    "s": "One queue for all background work (idle lab build Q)",
+    "k": "new"
+   }
+  ]
+ },
  {
   "v": "0.523.0",
   "at": "2026-11-04T08:06:51.586Z",

package/workspace/guide/agent-capabilities.md

@@ -2,7 +2,7 @@
 type: Guide
 title: "What Claude does behind the scenes"
 description: "The runbook of the workspace agent: each job Claude runs by itself, when it runs, what it reads, where its result shows, how to stop it, what it fixes by itself, the rules every job follows, and how to see what ran."
-timestamp: 2026-11-03
+timestamp: 2026-11-04
 ---
 
 # What Claude does behind the scenes
@@ -18,6 +18,9 @@ This page lists every job, so you know what runs and why. Each row links to the
 - **Few at a time.** A run shows at most 3 proposals. In a quiet week it shows nothing.
 - **Tested first.** A new version of a job shows nothing until its replay passes: its own known cases, and the guard's own cases (a made-up quote, a missing source, a path outside the repository, a write key). Each of those must be dropped.
 - **Your login, your connections.** A job runs as a background Claude session with your own login, model, profile and connectors. The workspace holds no password or token. `PMOS_HEADLESS=off` sends every job back to a terminal pane.
+- **One queue.** Every background job waits for its turn in one queue. Each minute the jobs go in this order: what is due at a set time (schedules, the morning brief and day check, the data refresh, report schedules), then the Initiative Driver, then the deep drafts (core documents, Your impact, metric drafts and reads, grill, proposals), then the idle work (the Horizon scan). What you ask for, such as Scan now or Prepare my day, starts at once and never waits in the queue.
+- **One run for each initiative.** A background run never starts on an initiative that another background run is working on. It tries again a minute later.
+- **Idle work waits for you to step away.** The idle work starts only when you have not used PMOS for 30 minutes, or in the last hour of your working day if you never stepped away. It also waits when 3 background runs are live, inside quiet hours, and at the spending limit.
 - **Your hours.** Scheduled jobs run inside your working hours and never inside your quiet hours.
 - **A spending limit.** Claude's checks spend at most the monthly credit you set in Settings › Claude's checks ($5, $10, $20 or $50). At the limit Claude pauses. It never spends beyond the limit.
 - **Every run is logged.** Each job writes its own ledger in the runtime folder on this computer. Only you see it, and it never goes into a repository.
@@ -34,7 +37,7 @@ This page lists every job, so you know what runs and why. Each row links to the
 | [Initiative descriptions](descriptions.md) | Once a day. At most 3 drafts a day for each product. | The PRD's first Problem or Why paragraph, or the record's own first paragraph | "Description" under the initiative's ✦ | `pmos ai off descriptions`. `PMOS_DESCRIPTIONS=off` |
 | [Expected outcomes](outcomes.md) | Once a day. At most 3 drafts a day for each product. | The PRD's success criteria, or the anchored key result | "Expected outcome" under the initiative's ✦. Today when a check is due. | `pmos ai off outcomes`. `PMOS_OUTCOMES=off` |
 | [Metric drafts](metric-drafts.md) | A product with no North Star, the first 14 days of a quarter, a program with no goal, and an initiative that gets a PRD | The last 8 weeks of real readings | The page of that product, program or initiative | `PMOS_METRIC_DRAFT=off` |
-| [Horizon scan](horizon-scan.md) | Daily by default, at most once a day. Scan now: at most 3 a day. | The public web for your kept watch list, and your repository | Report documents and proposals you approve | Cadence: Off. `PMOS_SCAN=off` |
+| [Horizon scan](horizon-scan.md) | Daily by default, at most once a day, the first time you step away after 09:00. Scan now: at most 3 a day. | The public web for your kept watch list, and your repository | Report documents and proposals you approve | Cadence: Off. `PMOS_SCAN=off` |
 | [What you might have missed](what-you-missed.md) | Each Monday, for the week before. It uses no model. | Records only: tasks, dates, initiatives and Jira status | One ✦ in Home's chat, at most 3 items | `pmos ai off drift`. `PMOS_DRIFT=off` |
 | [Your impact](your-impact.md) | Claude reads once a day. Home shows the week after it ends. | Slack, mail, Jira, your meeting notes, your documents and your key results | Adler's sprout on Home, and your impact log | Settings › General › Your impact. `PMOS_IMPACT=off` |
 | [Morning brief and day check](proactive-engine.md) | The morning brief at 06:45, once you schedule it. The day check every 60 minutes by default (30, 60 or off). | Calendar, Wispr Flow, Slack, Jira and Confluence are on by default. Mail, GitHub, Drive, Figma and websites are off by default. Mail: subjects and senders only. | Today | Each source is a setting. `pmos proactive settings` |
@@ -71,4 +74,4 @@ When you ask an agent session to do a piece of work, it can finish it, including
 - `pmos ai log [<job>]`: the last audit lines. Each dropped proposal shows why.
 - `pmos ai off [<job>]` and `pmos ai on [<job>]`: turn every AI job, or one job, off or on. `PMOS_AI=off` in the environment wins over the setting.
 
-The AI jobs are `past-repeat`, `decision-memory`, `outcomes`, `descriptions`, `drift`, `rehearse`, `next-prompt` and `sharpen-prompt`. A background run cannot change the switch.
+The AI jobs are `past-repeat`, `decision-memory`, `outcomes`, `descriptions`, `drift`, `rehearse`, `next-prompt` and `sharpen-prompt`. The switch also covers each job in the queue by its queue name: `schedules`, `proactive`, `prepare`, `day`, `refresh`, `reports`, `driver`, `finder`, `coredocs`, `milestones`, `propose`, `grill`, `metric-drafts`, `impact`, `metric-reads`, `forecast`, `report-design`, `report-structure` and `scan`. `pmos ai off` with no job stops all of them. Git sync uses no model, so the switch does not stop it. A background run cannot change the switch.

package/workspace/guide/home.md

@@ -195,7 +195,7 @@ There is no My work page. The list on Home is your list:
 - **While Claude works.** A line at the top of the chat shows what Claude is trying to do. If Claude keeps a to-do list, the line shows the to-do in progress and how many are done, for example **2 of 5 done**. Tap the line to see the full list. If Claude has no to-do list, the line shows Claude's newest one-line status and the time since your message. The line never shows raw commands. To see the commands, tap the terminal icon.
 - **A session's terminal.** **Watch its terminal** in **+**, or a session row such as **Claude sign-in**, opens the session's terminal in the Viewer. While the session runs, it is a real terminal: type in it to answer Claude or paste a code. It fills the Viewer and fits its lines again when you resize the Viewer or the window. The chat goes on showing every answer while the terminal is open and after you close it: the terminal is a second view of the same session. When the session has ended, the Viewer shows it read only and says so.
 - **Setting up your first project opens by itself** until a registered folder holds an initiative or setup has finished. The `PMOS` folder that the one-line installer makes holds no initiative yet, so a new user who opens PMOS from the Start menu or the Dock still gets the setup conversation, not an empty Home.
-- **Signing in to Claude when you set up your first project.** If the login you choose is signed in already, PMOS does not open a terminal. It tests the login at once. Otherwise PMOS runs `claude auth login`: Claude opens your browser to sign in, and the terminal under the question shows only Claude's request for a code, if it needs one. Claude's theme and trust-this-folder screens do not appear. When the sign-in ends, PMOS tests the login by itself. **The Claude login on this computer** is the login your usual terminal uses (`~/.claude.json`); PMOS no longer points it at a different file. You choose between two logins only: **The Claude login on this computer** and **A separate login just for PMOS**. The test passes for a login that works in your terminal. Claude updates its own files when it runs, and the test does not count that as a changed setup.
+- **Signing in to Claude when you set up your first project.** Setup tests the Claude login on this computer by itself and asks only when the test fails ([First run](welcome.md)). If the login you choose is signed in already, PMOS does not open a terminal. It tests the login at once. Otherwise PMOS runs `claude auth login`: Claude opens your browser to sign in, and the terminal under the question shows only Claude's request for a code, if it needs one. Claude's theme and trust-this-folder screens do not appear. When the sign-in ends, PMOS tests the login by itself. **The Claude login on this computer** is the login your usual terminal uses (`~/.claude.json`); PMOS no longer points it at a different file. You choose between two logins only: **The Claude login on this computer** and **A separate login just for PMOS**. The test passes for a login that works in your terminal. Claude updates its own files when it runs, and the test does not count that as a changed setup.
 - **Documents not written yet.** A document that does not exist yet shows **Draft it**, in its stage or ahead of it. The agent writes a draft. Nothing is issued until you read it.
 - **Meeting notes.** A decision, to-do or question that the notes caught is not a fact until you confirm it in the meeting's Replay. Home, Your day and the evening recap do not count unconfirmed points.
 

package/workspace/guide/horizon-scan.md

@@ -2,7 +2,7 @@
 type: Guide
 title: "The Horizon scan: reports and proposals for the whole portfolio"
 description: "What the portfolio Horizon scan does, its settings and schedule, Scan now and its limits, the report documents, the proposals you approve, what it can and cannot do, and the pmos scan commands."
-timestamp: 2026-10-28
+timestamp: 2026-11-04
 ---
 
 # The Horizon scan: reports and proposals for the whole portfolio
@@ -19,7 +19,7 @@ The settings belong to the workspace, not to a board:
 - **Repository.** On or off. When off, a scan reads the web only.
 - **Report types.** Benchmark, competitive, insight and productivity. All are on by default.
 
-A scheduled scan runs at most once a day, inside your working hours, never inside your quiet hours. Only one scan runs at a time. `PMOS_DRIVER=off` or `PMOS_SCAN=off` starts no scan at all. Board settings only show that the scan runs for the whole portfolio, with a link to the setup.
+A scheduled scan runs at most once a day, inside your working hours, never inside your quiet hours. After 09:00 it waits until you step away from PMOS for 30 minutes. If you do not step away, it runs in the last hour of your working day. Only one scan runs at a time. `PMOS_DRIVER=off` or `PMOS_SCAN=off` starts no scan at all. Board settings only show that the scan runs for the whole portfolio, with a link to the setup.
 
 ## The watch list
 

package/workspace/guide/idle-lab.md

@@ -0,0 +1,46 @@
+---
+type: Guide
+title: "The idle lab: prototypes and scans made while you are away"
+description: "Where the idle lab's results wait (on the initiative, or in Intelligence › Lab), how you keep or drop them, the compare view, and what the Lab shows each week."
+timestamp: 2026-11-04
+---
+
+# The idle lab: prototypes and scans made while you are away
+
+When you step away from PMOS, Claude can use the idle time to make prototypes and to scan the web for your initiatives. Each result is checked before you see it: a prototype must open and render, and a scan claim must cite a page Claude read. Nothing is written to a product repository before you press ✓.
+
+## Where a result waits
+
+A result for one initiative waits under that initiative's blue ✦, in one of three groups:
+
+- **Claude prototyped it.** A first round of 2 or 3 different ways to do it, as thumbnails. Press a thumbnail to open it in the compare view.
+- **Claude improved it.** A new round of one prototype. It says what changed, usually one thing, and where the change came from, for example a meeting note.
+- **Claude scanned for it.** A finding to attach as evidence, with the pages Claude read.
+
+✓ keeps it as one of the initiative's documents. ✕ asks why first. Claude learns from the reason, and you can leave it empty. Press ✕ again to drop it. Undo in the toast takes a ✓ back.
+
+A result that fits no initiative waits in **Intelligence › Lab**. One kind proposes a new initiative from a scan finding: ✓ starts it on its board.
+
+## The compare view
+
+The compare view shows a prototype beside what it changes:
+
+- **Production**: the screen as it is on the product repository's main branch.
+- **Design**: the design for the screen, from Figma or a Claude canvas.
+- **Prototype**: the round you opened.
+
+A column with no source is left out, with one line that says why. Claude is still proving that it can render production and read the design, so for now those two columns are left out and the prototype shows alone. **Where they differ** lists the differences as numbered pins. **History** lists every round: press a round to show it in the Prototype column. ✓ and ✕ in the header keep or drop the round shown.
+
+Each prototype opens in a sandboxed frame. It cannot reach PMOS or your data.
+
+## Intelligence › Lab
+
+The Lab shows only once there is something to show. Top to bottom:
+
+- One ✦ line for the week: the idle time Claude used, how many results you kept, and how many findings fit no initiative.
+- The findings that fit no initiative, each with ✓ and ✕.
+- **What Claude did in the background**: every background job this week, not only the idle lab. A result that waits on an initiative is a link there. It is not a second copy.
+- **You kept**: how many prototypes and scans you kept.
+- **Next time you step away**: what Claude plans to do next. Claude picks these by itself. To skip one, say so in chat.
+
+See [What Claude does behind the scenes](agent-capabilities.md) for every background job and how to stop it.

package/workspace/guide/index.md

@@ -40,6 +40,7 @@ PMOS has three places: **Home**, **Board** and **Intelligence**. The round menu
 
 - [What Claude does behind the scenes](agent-capabilities.md): every job Claude runs by itself, when it runs, what it reads, where you see it, how to stop it, and what it fixes by itself.
 - [First run](welcome.md): test your Claude login, describe the project, choose its folder, search your tools read-only, and set up PMOS from Home.
+- [Features appear when their moment comes](moments.md): why a new PMOS starts quiet, what opens the Board, Intelligence and reports, the one ✦ a day, and the dot on the menu.
 - [Layout](layout.md): the three places, the menu button, the Viewer and its trail, docking panels, the keys, and the phone.
 - [Start the workspace without a terminal](launcher.md): the PMOS app, the login item and `pmos launcher`.
 - [Settings](settings.md): what each section of Settings holds.
@@ -90,6 +91,7 @@ PMOS has three places: **Home**, **Board** and **Intelligence**. The round menu
 - [Forecasts](forecast.md): how PMOS forecasts a goal, a delivery, a program's milestones and a health budget, and `pmos forecast`.
 - [Claude drafts metrics without being asked](metric-drafts.md): when Claude drafts the metrics of a product, a program or an initiative, keep, drop, Undo, and `pmos metrics draft`.
 - [The Horizon scan](horizon-scan.md): how often it runs, the watch list, where findings land, what it cannot do, and `pmos scan`.
+- [The idle lab](idle-lab.md): prototypes and scans Claude makes while you are away, where they wait, the compare view and Intelligence › Lab.
 - [Set up a report by asking](reports-set-up.md), [Shape what a report says](report-structure.md), [Design how a report looks](report-design.md) and [Report slides](report-slides.md): a report from setup to slides.
 - [Draft and issue a cadence report](reports.md): `pmos report`, the notes kept when you draft again, and Issue.
 

package/workspace/guide/moments.md

@@ -0,0 +1,43 @@
+---
+type: Guide
+title: "Features appear when their moment comes"
+description: "Why a new PMOS starts quiet: the facts that open the Board, Intelligence and reports, the one ✦ a day in Home's chat, the soft dot on the menu, and what ✓ and ✕ do."
+timestamp: 2026-11-04
+---
+
+# Features appear when their moment comes
+
+A new PMOS does not show everything on the first day. Each place stays quiet until there is something real for it to show. When that happens, Claude tells you once, in Home's chat.
+
+## What opens each place
+
+A place opens on a fact PMOS can check in your records. It never opens because a number of days passed.
+
+| Place | It opens when | Before that it says |
+|---|---|---|
+| Board | You have three open initiatives, or a program has a date (its target or a milestone). | "The Board starts at three initiatives, or a date." |
+| Intelligence | Your OKR file has a goal with at least one key result. | "Intelligence starts when you keep a goal." |
+| Reports | There is a week of movement: your oldest initiative is at least 7 days old, and a record changed in the last 7 days. | "The first report comes after a week of movement." |
+
+Your first initiative is the first moment. The card that sets it up comes right after setup, so this list starts after it.
+
+## One ✦ a day in Home's chat
+
+When a place opens, Home's own chat shows one ✦ for it, with a link to its page in this guide (for example "How the Board works").
+
+- **✓** opens the place.
+- **✕** drops it. That ✦ does not come back.
+
+Claude shows at most one of these a day. Your impact and What you might have missed come first, so Home's chat shows one ✦ at a time.
+
+## The dot on the menu
+
+The first time a place opens, its row in the menu gets one soft blue dot, and so does the menu button. The dot goes away when you open that place. The reports dot is on the Intelligence row, because reports are in Intelligence under **⋯**.
+
+## A PMOS that was already in use
+
+This applies only to a product you set up with PMOS's first run. A product that was in use before keeps every place open, with no ✦ and no dot. A page shows its quiet line only when every product on this computer is new and none has reached that place.
+
+## Where it is kept
+
+Which moments were shown, kept or dropped, and which places opened, are kept with the product's setup progress, in PMOS's own folder on this computer. Nothing is written to your repository, and nothing leaves your computer.

package/workspace/guide/welcome.md

@@ -1,7 +1,7 @@
 ---
 type: Guide
 title: "First run: set up a project in Home"
-timestamp: 2026-11-01
+timestamp: 2026-11-04
 ---
 
 # First run
@@ -18,57 +18,66 @@ The desktop installers set up Node.js, Git, Claude Code and the PMOS launcher. P
 
 Setup needs a Claude Code login on this computer. Signing in on the Claude website alone does not sign in Claude Code. Jira, Slack, Google Drive and other tools need their own connections under that same login.
 
-## 1. Connect Claude and test the login
+In every step, **✓** says yes and **✕** asks Claude to change something. You type the change in the chat box. The chat box is open from the first screen.
 
-Choose **Set up a separate PMOS login**, or select an existing Claude setup. A profile name does not prove that the account is signed in.
+## 1. The Claude login
 
-Complete Claude's sign-in in the terminal that opens. Type `/login` there if needed. PMOS does not copy your password or credentials.
+When setup opens, PMOS tests the Claude login on this computer by itself. The test makes one small Claude request with no tools, no hooks and no connectors. It may use a little of your account usage.
 
-Choose **Test this login**. The test makes one small Claude request and may use your account allowance. It runs with no tools, no hooks and no connectors, and it does not read your files.
+If the test passes, setup shows one line: **Claude Code is signed in on this computer, and I just tested it**. To use another login, choose **change in Settings**.
 
-A failed test shows the reason: not signed in, Claude Code missing or too old, no network, or no account usage. A failed test does not switch to a different account, and setup does not continue until a test passes.
+If the test fails, setup shows one ✦ **Sign in to Claude Code** with the reason: not signed in, Claude Code missing or too old, no network, or no account usage. ✓ opens Claude's sign-in, and PMOS tests the login again when the sign-in ends. ✕ or **Use another login** shows the two logins: **The Claude login on this computer** and **A separate login just for PMOS**. PMOS never switches to a different account by itself. PMOS does not copy your password or credentials.
 
-If you change the login after a passing test, setup asks you to test again. Results from the earlier login are dropped.
+## 2. Name the product
 
-## 2. Describe the project
+The first question is the product or project. A name is enough. The team, a Jira key or a link help Claude find the right sources. For example: "Checkout redesign, payments team. Jira key PAY." To use a folder you already have, include its path.
 
-Type the project name and a few words about it. For example: "Checkout redesign, payments team. Jira key PAY." To use a folder you already have, include its path.
+If you answer before the login test passes, Claude keeps your answer and starts when the test passes.
 
-## 3. Choose the folder
+## 3. ✦ Where it lives, and what Claude reads
 
-Setup proposes a new folder under `PMOS` in your home folder, named after the project. Nothing is created until you confirm. Setup never proposes a folder that already holds files. If the name is taken, it adds a number.
+Claude proposes one ✦ with three parts:
 
-To use a folder you already have, choose **Choose an existing folder…**, or type its path. Setup then shows what is already there: a Git repository, a README, or an existing PMOS install. In an existing folder, setup adds only what is missing; your files and history stay as they are. **Put the new folder somewhere else…** changes where the new folder goes.
+- **Folder.** A folder named after the product, or the `PMOS` folder that the installer made while it is still empty. **Use another folder** opens the folder chooser. In an existing folder, setup adds only what is missing; your files and history stay as they are.
+- **Search, read-only.** The connected tools of the tested login, and the folder's README and Git history for an existing folder. The search can only use tools whose names say they read, search, list or get. Every other tool is refused before it runs. Searching uses a little of your account usage.
+- **Needs signing in.** A tool that needs a sign-in inside Claude is left out, with a **sign in** link.
 
-The search step can also read the folder's README and Git history.
+Nothing is created and nothing is searched before ✓. ✕ asks what to change. Type a folder path, the name of a tool to leave out, or "no search".
 
-## 4. Search your connected tools
+## 4. ✦ Set up PMOS with this context
 
-Choose **Show what I can read**. Setup lists the connectors of the tested login. This step makes no Claude request. A connector that needs sign-in is shown but cannot be selected; sign in to it in Claude first.
+After the search, Claude shows what it found as up to three tiles, one for each tool. A tool that could not be read adds nothing, and Claude says why.
 
-Select the sources, then choose **Search, read-only**. Searching uses some of your account usage. The search can only use tools whose names say they read, search, list or get. Every other tool is refused before it runs. Setup asks for no permission and sends no message.
+Then Claude proposes ✦ **Set up PMOS in** the folder **with this context**. It lists Purpose, Team, Timeline, Weekly, Owner and Goals, each with the source it came from. A line that no source supports says **Not found. Tell me, or leave blank**. Setup does not invent an owner, a goal or a Jira project. Press the pencil to edit a line, or type it in the chat box, for example "owner is Dana". If more than one Jira project matched, choose one first.
 
-Each source shows one result: what it found, nothing found, or why it could not read. A source that could not be read adds nothing to the review.
+✓ creates the folder if it is new, sets up Git in a new folder, installs PMOS (or reuses the PMOS already there), adds one line to `CLAUDE.md`, and saves the context to `pmos/planning/project-context.md`. Each step is recorded. If setup stops part of the way, ✓ continues from the step that did not finish.
 
-To skip the search, choose **Skip, use my description**.
+## 5. The basics, settled for you
 
-## 5. Review what was found
+After setup, one card lists what PMOS settled from `project-context.md`. Each line names its source and has one change:
 
-Each proposed field names the source it came from, with a link. A field is empty when no source supports it. Setup does not invent an owner, a goal or a Jira project.
+- **Profile.** A document product, unless the folder holds code. **it has code** changes it.
+- **Workflow.** Product feature, for every new initiative. If the purpose names a tender, vendor onboarding, a campaign, hiring or an event, **draft a … workflow** opens New workflow.
+- **Jira.** Only when the search reached Jira: the key is read, read-only, every weekday morning, and your tickets show in My work. **turn off** stops it.
+- **Calendar.** Only when the weekly came from a calendar that answered: the calendar is read. **turn off** stops it.
+- **About.** The purpose, with its source. **edit** opens Home to change it.
+- **Goal file.** One key result drafted from the purpose, marked PROPOSED in `pmos/planning/okrs/okrs-draft.md`. It claims no number. It counts only when you keep it in Intelligence. **remove** deletes it while it is still the text setup wrote. A product that already has a key result gets no draft.
+- **Launcher.** PMOS opens from Applications and starts at login (`pmos launcher on`). **turn off** removes it.
 
-Edit any field, and clear the findings you do not want. If more than one Jira project matched, choose one before you continue. Treat what a connector returns as information to check, not as instructions.
+Then ✦ **Draft the first initiative from the purpose, with its expected outcome**. ✓ opens Home with the draft. ✕ opens Home without one. An agent does not start automatically.
 
-## 6. Set up PMOS
+## 6. See your product, and keep the first initiative
 
-Choose **Set up PMOS in** the folder. Setup creates the folder if it is new, sets up Git in a new folder, installs PMOS, adds one line to `CLAUDE.md`, and saves the project context to `pmos/planning/project-context.md`.
+The project opens in Home. The chat shows four things, in this order. If you pressed ✕ on the draft, the third one is left out.
 
-If the folder already has PMOS, setup reuses it as it is and does not reinstall it. The button then says **Open PMOS in**. If a different project context file is already there, setup saves a second one beside it.
+- **What setup did.** One line for each finished step, with the time it took from the start of setup.
+- **Your product as PMOS sees it.** A card with the name, the team, the workflow and the timeline, and up to three cells: **The goal** (now and target), **The time** (the months, today, your weekly meeting and the check date) and **The work** (open tickets and the tickets that are yours). A cell shows only when a source gave its data. Each cell names its source. If a line is wrong, say so in the chat, or press the pencil.
+- **The first initiative, drafted.** The title comes from the purpose. The card shows the expected outcome, the check date and the owner, each with its source. Press ✓ to keep it, or ✕ to drop it. Nothing is written before you press ✓.
+- **What the notes did not say**, such as the owner or the goals. Say them in the chat at any time.
 
-Each step is recorded. If setup stops part of the way, run it again and it continues from the step that did not finish.
+When you press ✓, PMOS creates the initiative in Triage with its expected outcome and check date. Then it opens the initiative's chat and asks Claude to draft the BRD from the same sources, with the open questions marked. If the product has no goal yet, the expected outcome becomes its first goal (`pmos/planning/okrs/`). The card tells you this before you press ✓. If the notes give no expected outcome, type one on the card.
 
-## 7. Open the project
-
-The project opens in My Work. The last turn lists what setup did and offers next steps, such as checking the project context or showing what is open in Jira. An agent does not start automatically.
+The list on the left shows your weekly meeting, the draft and your tickets from the start. The Board, Intelligence and reports appear when they have something to show.
 
 ## Repair the login later