create-pmos 0.55.0 → 0.55.1

package/bin/create-pmos.js

@@ -292,9 +292,11 @@ function workspace(targetArg, opts) {
     console.log(forceReinstall
       ? 'Reinstalling the Terminal Workspace app (--reinstall)…'
       : `Upgrading the Terminal Workspace app (installed ${installed || 'unknown'} → bundled ${bundled})…`);
+    // Stop the running server FIRST. `npm install -g` deletes and rewrites the folder that server runs from; a server
+    // still running there afterwards keeps its old code and starts every new session in a deleted folder.
+    stopWorkspaceServer();
     const status = installWorkspaceGlobal();
     if (status === 0) {
-      stopWorkspaceServer();   // restart the persistent server on the new code
       console.log(`\n✅ pmos-workspace is now ${bundled}. Launching — if the UI still looks unchanged, hard-refresh the browser (Cmd/Ctrl+Shift+R) to drop the cached stylesheet.`);
     } else {
       console.log('\n⚠  Reinstall did not complete cleanly — launching the previously installed copy.');

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-pmos",
-  "version": "0.55.0",
+  "version": "0.55.1",
   "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": {

package/payload/KIT-VERSION

@@ -1,3 +1,3 @@
-kernel-source: pmos@058e723
+kernel-source: pmos@47defed
 assembled: 2026-10-06
 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: "2026-10-06T06:04:22Z"}
+generated: {by: process:create-pmos, at: "2026-10-06T07:02:27Z"}
 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: "2026-10-06T06:04:22Z"}
+generated: {by: process:create-pmos, at: "2026-10-06T07:02:27Z"}
 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: "2026-10-06T06:04:22Z"}
+generated: {by: process:create-pmos, at: "2026-10-06T07:02:27Z"}
 status: stable
 ---
 

package/workspace/guide/proactive-engine.md

@@ -32,6 +32,7 @@ The top of Today is the morning brief: a greeting, the day's calendar, the prior
 - Press **schedule the morning brief** in Today to make a daily schedule at 06:45. The schedule is an ordinary scheduled task, so it runs only while the workspace server runs. A sleeping computer gets no brief. A schedule that is made after its time of day for the first time runs within a minute, so pressing the button at 10:00 starts one brief session straight away.
 - Each source is a setting: calendar, wispr, mail, slack, jira, confluence, github, drive, figma and website. Calendar, wispr, slack, jira and confluence are on by default. Mail, github, drive, figma and website are off by default. The skill reads only subjects and senders of mail, never bodies. A source that is off cannot be reported as read.
 - **Scopes.** For Slack, Jira and Confluence you give each channel, project or space one scope. At most 50 names are kept for each service, and a name has at most 120 characters. **watch** means the session reads every message since the last read and writes a digest: the decisions, the open questions, and what touches a goal. **search** means the session reads nothing there and looks only for mentions of you and the words of your goals. **ignore** means the session does not read, search or report the place.
+- **Places.** The morning brief lists every Slack channel you are a member of, every Jira project you work in and every Confluence space you are in. It reads only their names and links, not their messages. Sources shows these places, so you can give each one a scope. Before the first brief that reads a service, Sources says that its places are listed after that read. At most 200 places are kept in a brief.
 - **Websites.** Under the website source you list at most 10 addresses. The session reads each one for changes that matter to a goal or a meeting.
 - **Mentions.** In every place that is not ignored, the skill finds the mentions of you: where, who, the link to the conversation, and a short paraphrase. It never copies a message.
 - **Meetings.** Each calendar event can carry a purpose and a prep note (at most 240 characters each), up to 6 steps before, during and after the meeting (each step says whether you or an agent does it), a suggested initiative with the reason, and up to 5 links.

package/workspace/guide/today.md

@@ -16,6 +16,8 @@ Above the cards, Today shows the day in two columns.
 
 Press **Prepare my day** to start the day. It runs from the brief to the delivery of each agent job, and you do not press anything else until a draft is ready for you. The page [Guide: Today and the proactive engine](proactive-engine.md) describes each step.
 
+If the morning-brief session stops before it submits a brief, the Read step says so and gives the reason. **Prepare my day** is then ready again, so press it to try once more. If the session says that its folder is not there, open the repository again from its current folder first.
+
 Each job has one of three kinds.
 
 | Kind | Who does it | What you do |

package/workspace/lib/brief.js

@@ -45,6 +45,7 @@ const WORTH_ASKING = 0.5;
 const MAX_ITEMS = 30;
 const MAX_EVENTS = 20;
 const MAX_DIGESTS = 30;
+const MAX_PLACES = 200;   // the Slack channels, Jira projects and Confluence spaces the PM is in: names only, so Sources can list them
 const MAX_RESOLVED = 200;
 const INI_RE = /^[a-z0-9][a-z0-9._-]*$/i;
 const GOAL_RE = /^g-[0-9a-f]{8}$/;
@@ -289,6 +290,24 @@ function validateBrief(raw, ctx = {}) {
     digests.push({ source: d.source, name, link: d.link || null, decisions: lines(d.decisions, 140), open: lines(d.open, 140), goals: lines(d.goals, 140) });
   }
 
+  // `places`: every channel, project and space the PM is in, by name, so the PM can set a level for each (G20). Names
+  // and links only: nothing of what is said there.
+  const pl = raw.places === undefined || raw.places === null ? [] : raw.places;
+  if (!Array.isArray(pl)) return { ok: false, error: 'places must be a list' };
+  if (pl.length > MAX_PLACES) return { ok: false, error: `places has ${pl.length} entries — at most ${MAX_PLACES}` };
+  const places = [];
+  for (const [i, x] of pl.entries()) {
+    const at = `place ${i + 1}`;
+    if (!x || typeof x !== 'object' || Array.isArray(x)) return { ok: false, error: `${at} is not an object` };
+    if (!DIGEST_SOURCES.includes(x.source)) return { ok: false, error: `${at}: source must be one of ${DIGEST_SOURCES.join(', ')}` };
+    if (sources[x.source] !== 'read') return { ok: false, error: `${at}: source "${x.source}" is not reported as read, so it cannot list places` };
+    if (typeof x.name !== 'string' || !x.name.trim()) return { ok: false, error: `${at} needs a name (the channel, project or space)` };
+    const err = optText(x.name, 120, `${at}: name`) || optLink(x.link, `${at}: link`);
+    if (err) return { ok: false, error: err };
+    const name = clip(x.name, 120);
+    if (!places.some((y) => y.source === x.source && y.name === name)) places.push({ source: x.source, name, link: x.link || null });
+  }
+
   // `resolved` is read by a merge only: the ids of stored items a check found done or answered.
   const rs = raw.resolved === undefined || raw.resolved === null ? [] : raw.resolved;
   if (!Array.isArray(rs) || rs.length > MAX_RESOLVED || rs.some((x) => typeof x !== 'string' || !ITEM_ID_RE.test(x))) return { ok: false, error: `resolved must be a list of at most ${MAX_RESOLVED} item ids` };
@@ -300,7 +319,7 @@ function validateBrief(raw, ctx = {}) {
     judgment = { used: raw.judgment.used };
   }
   return { ok: true, brief: { date, generated_at: new Date(now).toISOString(), sources, calendar, priorities: items.filter((x) => !x.worth_asking), worth_asking: items.filter((x) => x.worth_asking), judgment,
-    digests, resolved: [...new Set(rs)] } };
+    digests, places, resolved: [...new Set(rs)] } };
 }
 
 /**
@@ -310,12 +329,13 @@ function validateBrief(raw, ctx = {}) {
  *            An incoming item with a stored id replaces it and keeps its first_seen and new. More than 30 is refused.
  *   events   by id: the fields given overwrite, notes merge by kind (wispr, mail), new events are added.
  *   digests  replaced by (source, name).   sources  'read' wins, otherwise the stored state is kept.
+ *   places   a service's list is replaced when the incoming brief lists that service.
  * @returns {{ok:true, brief:object, added:number, removed:number}|{ok:false, error:string}}
  */
 function mergeBrief(stored, incoming, { now = Date.now() } = {}) {
   if (!incoming || typeof incoming !== 'object') return { ok: false, error: 'nothing to merge' };
   const inItems = [...(incoming.priorities || []), ...(incoming.worth_asking || [])];
-  if (!stored || typeof stored !== 'object') return { ok: true, brief: { ...incoming, digests: incoming.digests || [], resolved: [] }, added: inItems.length, removed: 0 };
+  if (!stored || typeof stored !== 'object') return { ok: true, brief: { ...incoming, digests: incoming.digests || [], places: incoming.places || [], resolved: [] }, added: inItems.length, removed: 0 };
   if (stored.date !== incoming.date) return { ok: false, error: `a merge is for the stored brief of the same day — the stored brief is for ${stored.date}` };
   const at = new Date(now).toISOString();
   const gone = new Set(incoming.resolved || []);
@@ -348,6 +368,10 @@ function mergeBrief(stored, incoming, { now = Date.now() } = {}) {
   for (const d of incoming.digests || []) dg.set(d.source + '|' + d.name, d);
   if (dg.size > MAX_DIGESTS) return { ok: false, error: `the brief would hold ${dg.size} digests — at most ${MAX_DIGESTS}` };
 
+  // places: a merge that lists places for a service replaces that service's list; a service it does not list keeps its own.
+  const listed = new Set((incoming.places || []).map((x) => x.source));
+  const places = [...(stored.places || []).filter((x) => !listed.has(x.source)), ...(incoming.places || [])];
+
   const sources = {};
   for (const k of SOURCES) {
     const a = (stored.sources || {})[k]; const b = (incoming.sources || {})[k];
@@ -359,9 +383,9 @@ function mergeBrief(stored, incoming, { now = Date.now() } = {}) {
     date: stored.date, generated_at: stored.generated_at || at, updated_at: at, sources, calendar,
     priorities: items.filter((x) => !x.worth_asking), worth_asking: items.filter((x) => x.worth_asking),
     judgment: { used: !!((stored.judgment && stored.judgment.used) || (incoming.judgment && incoming.judgment.used)) },
-    digests: [...dg.values()], resolved: [...new Set([...(stored.resolved || []), ...taken])].slice(-MAX_RESOLVED),
+    digests: [...dg.values()], places, resolved: [...new Set([...(stored.resolved || []), ...taken])].slice(-MAX_RESOLVED),
   } };
 }
 
-module.exports = { SOURCES, ITEM_SOURCES, SOURCE_STATES, OWNERS, EFFORTS, KINDS, DIGEST_SOURCES, WORTH_ASKING, MAX_ITEMS, MAX_EVENTS, MAX_DIGESTS,
+module.exports = { SOURCES, ITEM_SOURCES, SOURCE_STATES, OWNERS, EFFORTS, KINDS, DIGEST_SOURCES, WORTH_ASKING, MAX_ITEMS, MAX_EVENTS, MAX_DIGESTS, MAX_PLACES,
   isUrl, validateBrief, mergeBrief, greeting, localDate };

package/workspace/package-lock.json

@@ -1,6 +1,6 @@
 {
   "name": "pmos-workspace",
-  "version": "0.55.0",
+  "version": "0.55.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {

package/workspace/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pmos-workspace",
-  "version": "0.55.0",
+  "version": "0.55.1",
   "private": true,
   "description": "PMOS Terminal Workspace — local companion hosting real PTYs beside the library and gates (D62)",
   "license": "MIT",

package/workspace/public/pm/components.css

@@ -2172,6 +2172,7 @@
 }
 .td-day__hello { flex: 0 1 220px; min-width: 0; overflow: hidden; text-overflow: ellipsis; font-size: 14px; color: var(--foreground); white-space: nowrap; }
 .td-day__dim { color: var(--muted-foreground); }
+.td-day__warn { color: var(--warning); }
 .td-day__mono { font-family: var(--mono); font-variant-numeric: tabular-nums; font-size: 12px; }
 .td-day__steps { display: flex; flex-wrap: wrap; gap: 2px; list-style: none; margin: 0; padding: 0; }
 .td-day__ts { gap: 6px; padding: 0 6px; letter-spacing: 0.04em; color: var(--muted-foreground); }

package/workspace/public/pm/today-copy.js

@@ -104,6 +104,7 @@ export const COPY = {
     status: {
       idle: 'One press: the agent reads, makes the jobs and does them. You approve only what sends or decides.',
       reading: 'Reading your calendar and mentions…',
+      readFailed: (why) => `The read stopped: ${why}`,
       working: (w, n) => `${w} ${w === 1 ? 'agent' : 'agents'} working · ${n} ${n === 1 ? 'needs' : 'need'} you`,
       review: (n, d) => `${n} ${n === 1 ? 'needs' : 'need'} you · ${d} done`,
       done: (d) => `All jobs done · ${d} done`,
@@ -154,6 +155,7 @@ export const COPY = {
     place: { slack: 'channel', jira: 'project', confluence: 'space' },
     placeAdd: { slack: 'Add a channel', jira: 'Add a project', confluence: 'Add a space' },
     placeHint: 'Type a name and press Enter. It is watched.',
+    noPlaces: { slack: 'Your channels are listed after the first read of Slack. Press Prepare my day.', jira: 'Your projects are listed after the first read of Jira. Press Prepare my day.', confluence: 'Your spaces are listed after the first read of Confluence. Press Prepare my day.' },
     readNote: { slack: 'Mentions, DMs and your threads are always read.', jira: 'Issues assigned to you and mentions of you are always read.', confluence: 'Pages and comments that mention you are always read.' },
     decisions: 'Decisions',
     open: 'Open questions',
@@ -228,6 +230,7 @@ export const COPY = {
     stepsTitle: 'Steps found',
     readNone: 'Press Prepare my day. A session starts and reads your sources by itself.',
     readNow: 'The morning-brief session is reading your sources now.',
+    readFailed: (why) => `The morning-brief session did not finish: ${why} Press Prepare again to try again.`,
     readDone: (ev, it) => `Read ${ev} ${ev === 1 ? 'event' : 'events'} and ${it} ${it === 1 ? 'item' : 'items'}.`,
     notYet: 'Not started yet.',
     madeJobs: (n, you, agent) => `${n} ${n === 1 ? 'job' : 'jobs'}: ${you} for you, ${agent} for the agent.`,

package/workspace/public/pm/today-day.js

@@ -155,12 +155,14 @@ function model(d) {
   const scopes = (settings && settings.scopes) || {};
   const level = (svc, name) => (scopes[svc] && scopes[svc][name]) || 'search';
   const digests = ((brief && Array.isArray(brief.digests)) ? brief.digests : []).map((g) => ({ ...g, id: g.source + '|' + g.name, level: level(g.source, g.name) }));
+  const places = (brief && Array.isArray(brief.places)) ? brief.places : [];
   return {
-    d, brief, plan, goals, goalNo, run, tasks, jobs, events, now, settings, scopes, level, digests,
+    d, brief, plan, goals, goalNo, run, tasks, jobs, events, now, settings, scopes, level, digests, places,
     stage: run ? (d.stage || deriveStage(run, brief)) : null,
     mentions: jobs.filter((j) => j.item && j.item.mention),
     sources: (d && Array.isArray(d.sources)) ? d.sources : [],
     check: (d && d.check) || null,
+    readFailed: (d && d.read_failed) || null,
     free: plan && plan.time ? plan.time.free_minutes : null,
     maxGoals: (d && d.max_goals) || 7,
     needs: jobs.filter((j) => j.group === 'needs').length,
@@ -287,7 +289,7 @@ function topBar(ctx) {
       attrs: { class: 'td-day__ts', 'data-step-state': st, 'data-day-step': String(i), 'data-day-sel': 'step:' + i, 'aria-current': ctx.cur('step', String(i)), title: D.stepHint[i] } }));
   }));
   const working = m.tasks.filter((t) => t.status === 'working' || t.status === 'sending').length;
-  const status = !m.run ? D.status.idle : m.stage === 'reading' ? D.status.reading : m.stage === 'working' ? D.status.working(working, m.needs)
+  const status = !m.run ? D.status.idle : m.readFailed ? D.status.readFailed(m.readFailed.why) : m.stage === 'reading' ? D.status.reading : m.stage === 'working' ? D.status.working(working, m.needs)
     : m.stage === 'review' ? D.status.review(m.needs, m.done) : D.status.done(m.done);
   let check = null;
   if (m.run && m.check) {
@@ -300,7 +302,7 @@ function topBar(ctx) {
   const press = runs.get(ws);
   const busy = !!press && press.state === 'busy';
   const bs = m.run && m.run.brief_session;
-  const reading = !!bs && !bs.ended_at && m.stage === 'reading';
+  const reading = !!bs && !bs.ended_at && m.stage === 'reading' && !m.readFailed;
   const label = busy ? D.starting : reading ? D.stillReading : m.run ? D.again : D.prepare;
   return el('header', { class: 'td-day__bar' },
     (() => { const when = [dateLabel(m.now), m.free !== null && m.free !== undefined ? D.free(m.free) : null].filter(Boolean).join(' · ');
@@ -576,7 +578,8 @@ function stepDetail(ctx, i) {
   if (i === 0) {
     const bs = m.run.brief_session;
     const items = m.brief ? (m.brief.priorities || []).length + (m.brief.worth_asking || []).length : 0;
-    return [head, el('p', null, m.brief ? D.readDone(m.events.length, items) : D.readNow),
+    return [head, el('p', m.readFailed ? { class: 'td-day__warn', 'data-day': 'read-failed' } : null,
+      m.brief ? D.readDone(m.events.length, items) : m.readFailed ? D.readFailed(m.readFailed.why) : D.readNow),
       el('ul', { class: 'td-day__chk' }, ...m.sources.map((s) => el('li', { class: s.on && s.state === 'read' ? 'td-day__ok' : '' }, el('span', { class: s.on ? '' : 'td-day__dim' }, `${srcName(s.k)}: ${srcLine(s)}`)))),
       el('div', { class: 'td-day__acts' }, button({ label: D.manageSources, size: 'sm', onClick: () => select('sources', 'all') }),
         bs && bs.id ? button({ label: D.openSession, size: 'sm', variant: 'ghost', attrs: { 'data-day-act': 'session' }, onClick: () => openSession({ id: bs.id, name: 'morning-brief' }) }) : null)];
@@ -603,7 +606,9 @@ function srcLine(s) {
 
 function placeControls(ctx, svc) {
   const { m } = ctx;
-  const names = [...new Set([...Object.keys(m.scopes[svc] || {}), ...m.digests.filter((g) => g.source === svc).map((g) => g.name)])];
+  // every place the last read found you in, then any you set a level for or that has a digest
+  const names = [...new Set([...m.places.filter((x) => x.source === svc).map((x) => x.name), ...Object.keys(m.scopes[svc] || {}),
+    ...m.digests.filter((g) => g.source === svc).map((g) => g.name)])];
   const row = (name) => el('div', { class: 'td-day__ch', 'data-day-place': svc + ':' + name }, el('span', null, name), levelSeg(ctx, svc, name));
   const live = names.filter((x) => m.level(svc, x) !== 'ignore');
   const ign = names.filter((x) => m.level(svc, x) === 'ignore');
@@ -619,6 +624,7 @@ function placeControls(ctx, svc) {
   input.control.addEventListener('input', () => { ui.placeDraft[svc] = input.control.value; });
   return [
     note(D.readNote[svc]),
+    m.places.some((x) => x.source === svc) ? null : note(D.noPlaces[svc]),
     ...live.map(row),
     ign.length ? button({ label: ui.showIgn[svc] ? D.hideIgnored : D.showIgnored(ign.length), variant: 'ghost', size: 'sm',
       attrs: { 'data-day-ignored': svc, 'aria-expanded': ui.showIgn[svc] ? 'true' : 'false' }, onClick: () => { ui.showIgn[svc] = !ui.showIgn[svc]; rerender(); } }) : null,

package/workspace/server.js

@@ -156,6 +156,13 @@ const EMBEDDED_SELF = !INSTALLED_COPY && SELF_ROOT === PARENT_DIR &&
 // launcher named a real repo via PMOS_SELF_ROOT. An installed copy launched WITHOUT that env
 // (unsupported direct run) has no home — the client must open a repo to do anything.
 const HOME_ENABLED = EMBEDDED_SELF || (!!process.env.PMOS_SELF_ROOT && dirExists(SELF_ROOT));
+// An installed copy runs from the global npm folder, and the launcher starts it there. An upgrade (`npm install -g`)
+// deletes and rewrites that folder, so a process that inherits it as its working directory is left in a deleted
+// directory: the tmux daemon this server starts, and every agent CLI that daemon or a PTY starts, inherit it, and an
+// agent CLI that reads its working directory at start exits ("the current working directory was deleted"). Every path
+// in this server is absolute (__dirname, STATE, a workspace root), so an installed copy moves to the home folder,
+// which no upgrade removes.
+if (INSTALLED_COPY) { try { process.chdir(os.homedir()); } catch { /* keep the launch directory */ } }
 const PUB = path.join(__dirname, 'public');
 const PM_DIR = path.join(PUB, 'pm');   // the four-panel front end — served as a directory, guarded below
 const XROOT = path.join(__dirname, 'node_modules', '@xterm');
@@ -2849,7 +2856,10 @@ function writeTmuxConf() {
 // `env` is an OPT-IN override and only `new-session` ever passes one (profiles R3). has-session /
 // kill-session / list-sessions / display-message have no session context, so handing them a
 // profile's overlay would apply it to whichever pane the command happened to touch — a silent bug.
-function tmuxCtl(args, env) { return spawnSync(TMUX, ['-L', TMUX_SOCK, ...args], { env: env || sessionEnv(), encoding: 'utf8' }); }
+// The tmux daemon is born from the first of these calls and keeps that working directory for its whole life, across
+// workspace upgrades. It is started in the home folder, never in this server's install folder (see the chdir above).
+function tmuxCtl(args, env) { return spawnSync(TMUX, ['-L', TMUX_SOCK, ...args], { env: env || sessionEnv(), encoding: 'utf8', cwd: tmuxCwd() }); }
+function tmuxCwd() { const h = os.homedir(); return dirExists(h) ? h : undefined; }
 function tmuxHasSession(name) { return !!TMUX && tmuxCtl(['has-session', '-t', name]).status === 0; }
 function tmuxKillSession(name) { if (TMUX && name) tmuxCtl(['kill-session', '-t', name]); }
 // Create the tmux session a FRESH spawn runs in. It REFUSES a name that already exists rather
@@ -3109,6 +3119,11 @@ function spawnSession({ name, agent, cwd, ws, origin, initialInput, profile, ini
   const cmd = isShell ? (commandExists(SHELL) ? SHELL : 'sh') : agent;
   // a session's working directory is confined to its OWN workspace root (R3)
   const workdir = cwd && insideRoot(root, path.resolve(root, cwd)) ? path.resolve(root, cwd) : root;
+  // A repository folder that was moved or deleted after it was registered: tmux would start the agent in its own folder
+  // instead, and the agent would run in the wrong place or not at all. Refuse here and name the folder.
+  if (!dirExists(workdir)) return exitedSession({ id, ws: w ? w.id : wsId(SELF_ROOT), root, name: (name && String(name).trim()) || agent, agent,
+    origin: tag, profile: null, profileRefusal: null, plan_item: planItem || null, token: sessionToken },
+  workdir, `the folder ${workdir} is not there. Open the repository again from its current folder, then try again.`);
   // Resolved once, here, so every path below spawns under the SAME environment the record claims.
   // `pres.id` is null whenever the overlay was not applied — including when a profile was asked for
   // and could not be read — so the pane line can never name a profile that did nothing (R3/R9).
@@ -7122,7 +7137,8 @@ async function handleRequest(req, res) {
       const list = readSchedules(); list.unshift(sched); writeSchedules(list); seeded = true;
     }
     const result = fireSchedule(sched);
-    if (result.ok && result.sessionId) run.brief_session = { id: result.sessionId, started_at: new Date(pnow).toISOString(), ended_at: null };
+    // A session that could not start is recorded as already ended, so the page says why instead of waiting on a read.
+    if (result.sessionId) run.brief_session = { id: result.sessionId, started_at: new Date(pnow).toISOString(), ended_at: result.ok ? null : new Date(pnow).toISOString() };
     writeDayRun(run);
     return json(res, { ok: !!result.ok, started: !!result.ok, session: result.sessionId || null, run, seeded, schedule: { id: sched.id, time: sched.time }, result, goals: open.length,
       ...(result.ok ? {} : { error: result.reason || 'the morning-brief session did not start' }) }, result.ok ? 200 : 500);
@@ -7154,7 +7170,13 @@ async function handleRequest(req, res) {
     const due = dayRunLib.checkDue(run, { settings: st, now, events: brief ? brief.calendar || [] : [] });
     const items = brief ? [...(brief.priorities || []), ...(brief.worth_asking || [])] : [];
     const sources = proactiveLib.BRIEF_SOURCES.map((k) => ({ k, on: st.briefSources[k] === true, state: brief ? brief.sources[k] || null : null, got: items.filter((x) => x.source === k).length }));
-    return json(res, { ok: true, ws: w ? w.id : null, today: date, now: new Date(now).toISOString(), run, stage: dayRunLib.stage(run, brief), brief, plan, cards_error: cardsError,
+    // The brief session ended and no brief arrived: say so, with the session's own reason, so the page does not wait on a
+    // read that will never finish (it reads as "stuck on Read" otherwise).
+    const bs = run && run.brief_session; const bsess = bs ? sessions.get(bs.id) : null;
+    const readFailed = !run || brief ? null
+      : !bs ? { session: null, why: 'the morning-brief session did not start.' }
+      : (!!bs.ended_at || !bsess || bsess.state === 'exited') ? { session: bs.id, why: (bsess && bsess.note) || 'the session ended without submitting a brief.' } : null;
+    return json(res, { ok: true, ws: w ? w.id : null, today: date, now: new Date(now).toISOString(), run, stage: dayRunLib.stage(run, brief), brief, plan, cards_error: cardsError, read_failed: readFailed,
       settings: { briefSources: st.briefSources, scopes: st.scopes, dayCheck: st.dayCheck, websites: st.websites }, sources,
       check: { every: st.dayCheck.every, last: run ? run.checks.last : null, next: due.next, running: !!(run && run.checks.session_id), log: run ? run.checks.log : [] } });
   }

package/workspace/skills/morning-brief.skill

@@ -34,6 +34,7 @@ Procedure:
    - Drive, if on: files shared with the PM or changed since yesterday that a meeting or a goal names. Keep the name, the owner and the link.
    - Figma, if on: files the PM owns or was asked to look at, with new comments since yesterday. Keep the file, the commenter and a short paraphrase of the comment.
    - Website, if on: read each address in the websites list once, only to see what changed that matters to a goal or a meeting. Report what changed in one line. Do not follow links to other sites.
+   - LIST THE PLACES THE PM IS IN, for each of Slack, Jira and Confluence that you read: the Slack channels the PM is a member of (public and private, not direct messages), the Jira projects the PM works in (projects with issues assigned to, reported by or watched by the PM), and the Confluence spaces the PM is a member of or has written in. Keep only the name (for example "#pm-team", "PRICE", "Product") and the link. Report them under "places" (see step 4). This list is how the PM sets a scope for each place, so list every place, not only the busy ones. Read only the names: do not read messages to make this list.
    - Slack, Jira and Confluence take a scope for each name. Read the scopes from `pmos proactive settings`. A name with no scope is "search".
      - watch: read EVERY message or comment since the last read (yesterday morning if there is none). Then write one digest (see step 4) with the decisions made, the open questions, and what touches a goal. Write the digest, not the messages.
      - search: do not read the place. Look only for MENTIONS of the PM and for the words of the PM's goals and initiatives.
@@ -70,9 +71,10 @@ Procedure:
     "items": [{"title": "…", "why_now": "…", "source": "calendar|wispr|mail|slack|jira|confluence|github|drive|figma|website|derived", "initiative_id": "<id or omit>", "goal": "<goal id or omit>", "mention": false, "owner": "pm|agent", "effort": "minutes|an-hour|half-day|day", "confidence": 0.0, "evidence": "…",
               "link": "https://…", "where": "#pm-team", "who": "…", "about": "…", "ask": "…", "reply_needed": true, "how": "…", "kind": "you|approve|auto"}],
     "digests": [{"source": "slack|jira|confluence", "name": "#pm-team", "link": "https://…", "decisions": ["…"], "open": ["…"], "goals": ["…"]}],
+    "places": [{"source": "slack|jira|confluence", "name": "#pm-team", "link": "https://…"}],
     "judgment": {"used": false}}
    ```
-   A digest has at most 5 decisions, 5 open questions and 5 goal lines, each at most 140 characters, and at most 30 digests. Write a digest only for a place with scope watch, and only for a source that is "read". Only report calendar events if calendar is "read". Only give an item a source that is "read" (or "derived"). Only use a goal id that `pmos proactive goals` listed today: the workspace refuses any other. Leave out a field you cannot fill: an empty string is worse than no field.
+   "places" holds at most 200 entries, each a name of at most 120 characters, only for a source that is "read". A digest has at most 5 decisions, 5 open questions and 5 goal lines, each at most 140 characters, and at most 30 digests. Write a digest only for a place with scope watch, and only for a source that is "read". Only report calendar events if calendar is "read". Only give an item a source that is "read" (or "derived"). Only use a goal id that `pmos proactive goals` listed today: the workspace refuses any other. Leave out a field you cannot fill: an empty string is worse than no field.
 5. Run `pmos proactive brief submit --file <that file>`. If it is refused, read the reason, fix the file and submit once more. Do not work around a refusal.
 6. Finish by saying in three lines how many goals have a step and which have none, how many priorities and worth-asking items you submitted, and which sources were not reached.