Remove redundant work-detection from HEARTBEAT.md (now in AGENTS.md)

Add wind-down rules to AGENTS.md + first 19:00 nudge to HEARTBEAT.md
Add Jellyfin CLI + update wind-down instructions with concrete suggestions
2026-02-07 19:20:12 +00:00 · 2026-02-07 19:18:34 +00:00 · 2026-02-07 16:09:01 +00:00 · 2026-02-07 15:54:48 +00:00 · 2026-02-07 15:37:00 +00:00 · 2026-02-07 15:15:32 +00:00
15 changed files with 950 additions and 360 deletions
--- a/AGENTS.md
+++ b/AGENTS.md
@ -48,6 +48,26 @@ Capture what matters. Decisions, context, things to remember. Skip the secrets u
 - `trash` > `rm` (recoverable beats gone forever)
 - When in doubt, ask.

+### 🔑 Credentials
+- **Never read credential files.** Not even to "verify" or "check" them.
+- When setting up a new integration, create `.credentials/service.env` with **placeholder values** and let the human fill them in.
+- Scripts source credentials at runtime — you don't need to see them.
+- Example placeholder file:
+  ```
+  SERVICE_URL=https://example.com
+  SERVICE_API_KEY=FILL_IN
+  SERVICE_USER=FILL_IN
+  ```
+
+## Evening Wind-Down (19:00-23:00 Vienna)
+
+After 19:00 Vienna time, you're in wind-down mode. This applies in **all conversations**, not just heartbeats:
+
+- **If chatting about work/tinkering**: Actively look for stopping points. When a subtask completes or there's a natural pause, suggest wrapping up. Don't just passively help them keep going.
+- **If they want to start something new after 20:00**: Gently suggest postponing to tomorrow.
+- **Concrete suggestions**: Use `bin/jellyfin` and `bin/audiobooks` to suggest specific shows/audiobooks instead of vague "maybe wind down."
+- **Evening reminders**: When they're transitioning to wind-down, remind about evening recurring tasks (e.g., nose shower) before they get too comfortable.
+
 ## External vs Internal

 **Safe to do freely:**
--- a/HEARTBEAT.md
+++ b/HEARTBEAT.md
@ -8,19 +8,17 @@ Check the following and notify only once per event (track in `memory/heartbeat-s

 3. **Evening wind-down guidance (19:00-22:00 Vienna)**:
   
-   **Active work detection**: If user is chatting with me about work tasks, I KNOW they're working. Don't just passively help — actively look for stopping points.
-   
-   **During work conversations after 19:00**:
-   - Track how long the current task has been going
-   - When a subtask completes or there's a natural pause, suggest: "This might be a good stopping point for tonight"
-   - If a task is dragging on (30+ min with no end in sight), gently note: "This is getting complex — want to pick it up tomorrow with fresh eyes?"
-   - Don't wait for them to ask — proactively identify when a task can be paused
+   **First nudge at 19:00**: On the first heartbeat at/after 19:00 Vienna, always check in — ask what they're doing if not in active conversation, or nudge toward winding down if already chatting about work/tinkering.
   
   **If NOT in active conversation**: Send a brief WhatsApp asking what they're doing (working? winding down? about to start something new?). Log responses in `memory/wind-down-log.json`.
   
-   **If they're about to start something new after 20:00**: Gently suggest postponing to tomorrow.
+   **Concrete wind-down suggestions**: Don't just say "maybe wind down" — use real data:
+   - `bin/jellyfin resume` → suggest continuing what they're watching
+   - `bin/jellyfin recent` → suggest something newly added
+   - `bin/audiobooks current` → suggest picking up their audiobook
+   - Make it specific: "You're 44% through Die zweite Legion" or "New movie added: The Rip (2026)"
   
-   **Work-end reminders**: When they indicate they're wrapping up or transitioning to wind-down, check `memory/brain-dump.json` for `recurring` items with `when: "evening"` and remind them (e.g., nose shower) before they get too deep into relaxation mode.
+   **Work-end reminders**: When they indicate they're wrapping up or transitioning to wind-down, check `memory/tasks.json` for `recurring` items with `when: "evening"` and remind them (e.g., nose shower) before they get too deep into relaxation mode.

   **📝 AUTO-LOG EVERYTHING 19:00→SLEEP** — Log to `memory/wind-down-log.json` as events happen:
   - What they're doing (working, watching TV, chatting, browsing, etc.)
@ -44,10 +42,18 @@ Check the following and notify only once per event (track in `memory/heartbeat-s

 5. **Steam Machine / Steam Frame**: Check for **official Valve news only** about pricing or release date. Notify once when official price or specific release date is announced. Baseline: "early 2026" window, no price yet.

-6. **Brain dump nudging**: During daytime hours (10:00-18:00 Vienna), occasionally check `memory/brain-dump.json` for:
-   - Items with `remindOn` date matching today
-   - Items older than 2 days that haven't been addressed (weekdays only — Mon-Fri)
-   On weekends: only surface items with `remindOn` = today.
+6. **OpenClaw + Claude Code updates**: Check once daily (daytime) for:
+   - OpenClaw releases (especially for Opus 4.6 support via PR #9863)
+   - Claude Code new versions
+   Track in `memory/heartbeat-state.json` watchlist. Notify when new versions ship.
+
+7. **Task nudging**: During daytime hours (10:00-18:00 Vienna), occasionally check `memory/tasks.json` for:
+   - `now` priority tasks — nudge daily
+   - `soon` priority tasks — nudge every 2-3 days (weekdays only)
+   - `someday` tasks — only mention during weekly reviews or when contextually relevant
+   On weekends: only nudge `now` priority items.
   Surface 1-2 items max, conversationally. NOT in evening — that's wind-down time.
+   **Auto-capture:** When user mentions tasks naturally in chat ("I should...", "I need to...", "Don't forget..."), add to tasks.json and confirm with a short acknowledgment.
+   **Done = delete.** Remove completed tasks immediately. History lives in daily memory logs.

 If nothing new to report, reply HEARTBEAT_OK.
--- a/TOOLS.md
+++ b/TOOLS.md
@ -67,6 +67,47 @@ derstandard raw [max]            # Full RSS XML
 2. Pick interesting ones, optionally fetch full content with `articles`
 3. Next briefing: only shows articles published since last check

+## Audiobookshelf
+
+Helper script: `~/clawd/bin/audiobooks`
+
+```bash
+audiobooks current              # Currently listening / in-progress books
+audiobooks recent [limit]       # Recently active books (default: 5)
+audiobooks finished [limit]     # Finished books (default: 10)
+audiobooks stats                # Listening stats overview
+audiobooks libraries            # List libraries
+```
+
+- Credentials: `.credentials/audiobookshelf.env`
+- URL: `https://audiobooks.cloonar.com`
+- Output is tab-separated for minimal tokens
+- Use during wind-down to suggest continuing audiobook
+
+---
+
+## Jellyfin
+
+Helper script: `~/clawd/bin/jellyfin`
+
+```bash
+jellyfin resume [limit]         # Continue watching (in-progress items)
+jellyfin recent [limit]         # Recently added to library
+jellyfin watched [limit]        # Watch history (last played)
+jellyfin shows                  # All series with watch status
+jellyfin movies                 # All movies with watch status
+jellyfin search <query>         # Search library
+jellyfin stats                  # Watch statistics overview
+jellyfin libraries              # List libraries
+```
+
+- Credentials: `.credentials/jellyfin.env` (user-scoped token for `tv`)
+- URL: `https://jellyfin.cloonar.com`
+- Output is tab-separated for minimal tokens
+- Use during wind-down to suggest specific shows/movies
+
+---
+
 ## Forgejo Git Access

 Helper script: `~/bin/forgejo`
@ -137,28 +178,30 @@ ainews reset                # Clear seen history

 ## Brain Dump CLI

-Helper script: `~/clawd/bin/brain-dump`
+Helper script: `~/clawd/bin/tasks`

 ```bash
-brain-dump list [--priority now,soon] [--due] [--limit N]
-brain-dump add --text "..." --priority soon [--context "..."]
-brain-dump edit <id> [--text "..."] [--priority ...] [--context "..."]
-brain-dump done <id>
-brain-dump show <id>
-brain-dump nudged <id1>,<id2>,...
-brain-dump recurring [--when evening]
+tasks list [--priority now,soon] [--due] [--limit N]
+tasks add --text "..." --priority soon [--context "..."]
+tasks add --recurring --text "..." --frequency daily [--when evening] [--context "..."]
+tasks edit <id> [--text "..."] [--priority|--frequency|--when|--context "..."]
+tasks done <id>
+tasks show <id>
+tasks nudged <id1>,<id2>,...
+tasks recurring
 ```

- Data: `memory/brain-dump.json`
+- Data: `memory/tasks.json`
 - `--due` filters by nudge interval: now=1d, soon=3d, someday=7d
 - `nudged` marks tasks as just-nudged (resets due timer)
- `recurring` lists recurring reminders (nose shower etc.)
+- `recurring` lists all recurring items with full context (note, when, frequency)
+- `add --recurring` / `edit` / `done` work for both tasks and recurring items
 - Output is tab-separated for minimal tokens

 **Heartbeat workflow:**
-1. `brain-dump list --due --limit 2` → get tasks needing a nudge
+1. `tasks list --due --limit 2` → get tasks needing a nudge
 2. Mention them conversationally
-3. `brain-dump nudged <id1>,<id2>` → mark as nudged
+3. `tasks nudged <id1>,<id2>` → mark as nudged

 ---

--- a/bin/audiobooks
+++ b/bin/audiobooks
@ -0,0 +1,169 @@
+#!/usr/bin/env node
+
+const fs = require('fs');
+const path = require('path');
+const https = require('https');
+
+const CRED_PATH = path.join(__dirname, '..', '.credentials', 'audiobookshelf.env');
+
+function loadCreds() {
+  const raw = fs.readFileSync(CRED_PATH, 'utf8');
+  const env = {};
+  for (const line of raw.split('\n')) {
+    const m = line.match(/^(\w+)=(.+)$/);
+    if (m) env[m[1]] = m[2];
+  }
+  return env;
+}
+
+function api(endpoint) {
+  const { AUDIOBOOKSHELF_URL, AUDIOBOOKSHELF_TOKEN } = loadCreds();
+  const url = `${AUDIOBOOKSHELF_URL}${endpoint}`;
+  return new Promise((resolve, reject) => {
+    const req = https.get(url, { headers: { 'Authorization': `Bearer ${AUDIOBOOKSHELF_TOKEN}` } }, (res) => {
+      let data = '';
+      res.on('data', c => data += c);
+      res.on('end', () => {
+        try { resolve(JSON.parse(data)); }
+        catch { reject(new Error(`Invalid JSON from ${endpoint}`)); }
+      });
+    });
+    req.on('error', reject);
+    req.setTimeout(10000, () => { req.destroy(); reject(new Error('Timeout')); });
+  });
+}
+
+function formatDuration(secs) {
+  const h = Math.floor(secs / 3600);
+  const m = Math.floor((secs % 3600) / 60);
+  if (h > 0) return `${h}h ${m}m`;
+  return `${m}m`;
+}
+
+function formatProgress(progress) {
+  return `${Math.round(progress * 100)}%`;
+}
+
+async function cmdCurrent() {
+  const data = await api('/api/me/items-in-progress');
+  const items = data.libraryItems || [];
+  if (items.length === 0) {
+    console.log('Nothing in progress.');
+    return;
+  }
+  for (const item of items) {
+    const meta = item.media?.metadata || {};
+    const title = meta.title || 'Unknown';
+    const author = meta.authorName || 'Unknown';
+    const narrator = meta.narratorName || '-';
+    const series = meta.seriesName || '-';
+    const duration = formatDuration(item.media?.duration || 0);
+    
+    // Get progress from user's mediaProgress
+    const me = await api('/api/me');
+    const prog = (me.mediaProgress || []).find(p => p.libraryItemId === item.id);
+    const progress = prog ? formatProgress(prog.progress) : '?';
+    const currentTime = prog ? formatDuration(prog.currentTime) : '?';
+    const remaining = prog ? formatDuration((item.media?.duration || 0) - (prog.currentTime || 0)) : '?';
+    
+    console.log(`${title}\t${author}\t${series}\t${progress}\t${currentTime}/${duration}\tremaining:${remaining}\tnarrator:${narrator}`);
+  }
+}
+
+async function cmdLibraries() {
+  const data = await api('/api/libraries');
+  for (const lib of (data.libraries || [])) {
+    console.log(`${lib.id}\t${lib.name}\t${lib.mediaType}`);
+  }
+}
+
+async function cmdRecent(args) {
+  const limit = args[0] || '5';
+  const me = await api('/api/me');
+  const progress = (me.mediaProgress || [])
+    .filter(p => !p.isFinished)
+    .sort((a, b) => (b.lastUpdate || 0) - (a.lastUpdate || 0))
+    .slice(0, parseInt(limit, 10));
+  
+  if (progress.length === 0) {
+    console.log('No recent listening activity.');
+    return;
+  }
+  
+  for (const p of progress) {
+    try {
+      const item = await api(`/api/items/${p.libraryItemId}?expanded=1`);
+      const meta = item.media?.metadata || {};
+      const title = meta.title || 'Unknown';
+      const author = meta.authorName || 'Unknown';
+      const series = meta.seriesName || '-';
+      const pct = formatProgress(p.progress);
+      const lastUpdate = new Date(p.lastUpdate).toISOString().slice(0, 10);
+      console.log(`${title}\t${author}\t${series}\t${pct}\tlast:${lastUpdate}`);
+    } catch {
+      // Item might have been removed
+    }
+  }
+}
+
+async function cmdFinished(args) {
+  const limit = args[0] || '10';
+  const me = await api('/api/me');
+  const finished = (me.mediaProgress || [])
+    .filter(p => p.isFinished)
+    .sort((a, b) => (b.finishedAt || 0) - (a.finishedAt || 0))
+    .slice(0, parseInt(limit, 10));
+  
+  if (finished.length === 0) {
+    console.log('No finished books.');
+    return;
+  }
+  
+  for (const p of finished) {
+    try {
+      const item = await api(`/api/items/${p.libraryItemId}?expanded=1`);
+      const meta = item.media?.metadata || {};
+      const title = meta.title || 'Unknown';
+      const author = meta.authorName || 'Unknown';
+      const finishedDate = p.finishedAt ? new Date(p.finishedAt).toISOString().slice(0, 10) : '?';
+      console.log(`${title}\t${author}\tfinished:${finishedDate}`);
+    } catch {
+      // Item might have been removed
+    }
+  }
+}
+
+async function cmdStats() {
+  const me = await api('/api/me');
+  const progress = me.mediaProgress || [];
+  const inProgress = progress.filter(p => !p.isFinished && !p.hideFromContinueListening).length;
+  const finished = progress.filter(p => p.isFinished).length;
+  const totalListened = progress.reduce((sum, p) => sum + (p.currentTime || 0), 0);
+  
+  console.log(`in-progress:${inProgress}\tfinished:${finished}\ttotal-listened:${formatDuration(totalListened)}`);
+}
+
+// --- Main ---
+
+const [cmd, ...args] = process.argv.slice(2);
+
+const run = async () => {
+  switch (cmd) {
+    case 'current': await cmdCurrent(); break;
+    case 'recent': await cmdRecent(args); break;
+    case 'finished': await cmdFinished(args); break;
+    case 'stats': await cmdStats(); break;
+    case 'libraries': await cmdLibraries(); break;
+    default:
+      console.log(`Usage: audiobooks <command>
+
+Commands:
+  current              Currently listening / in-progress books
+  recent [limit]       Recently active books (default: 5)
+  finished [limit]     Finished books (default: 10)
+  stats                Listening stats overview
+  libraries            List libraries`);
+  }
+};
+
+run().catch(err => { console.error(`Error: ${err.message}`); process.exit(1); });
--- a/bin/jellyfin
+++ b/bin/jellyfin
@ -0,0 +1,150 @@
+#!/usr/bin/env node
+// Jellyfin CLI — token-efficient output for assistant use
+const https = require('https');
+const fs = require('fs');
+const path = require('path');
+
+// Load credentials
+const envFile = fs.readFileSync(path.join(__dirname, '..', '.credentials', 'jellyfin.env'), 'utf8');
+const env = {};
+envFile.split('\n').forEach(l => { const [k, ...v] = l.split('='); if (k && v.length) env[k.trim()] = v.join('=').trim(); });
+
+const BASE = env.JELLYFIN_URL;
+const API_KEY = env.JELLYFIN_API_KEY;
+const USER_ID = env.JELLYFIN_USER_ID;
+
+function api(endpoint) {
+  const sep = endpoint.includes('?') ? '&' : '?';
+  const url = `${BASE}${endpoint}${sep}api_key=${API_KEY}`;
+  return new Promise((resolve, reject) => {
+    https.get(url, res => {
+      let data = '';
+      res.on('data', c => data += c);
+      res.on('end', () => {
+        try { resolve(JSON.parse(data)); } catch { resolve(data); }
+      });
+    }).on('error', reject);
+  });
+}
+
+function getUserId() { return USER_ID; }
+
+function dur(ticks) {
+  if (!ticks) return '';
+  const min = Math.round(ticks / 600000000);
+  if (min < 60) return `${min}m`;
+  return `${Math.floor(min/60)}h${min%60 ? ' '+min%60+'m' : ''}`;
+}
+
+function pct(userData, runTimeTicks) {
+  if (!userData) return '';
+  if (userData.Played) return '✓';
+  if (userData.PlayedPercentage) return Math.round(userData.PlayedPercentage) + '%';
+  if (userData.PlaybackPositionTicks && runTimeTicks) return Math.round(userData.PlaybackPositionTicks / runTimeTicks * 100) + '%';
+  return '';
+}
+
+const commands = {
+  async resume(limit = 10) {
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items/Resume?Limit=${limit}&Fields=Overview,RunTimeTicks&MediaTypes=Video`);
+    if (!r.Items.length) return console.log('Nothing in progress');
+    console.log('TITLE\tEPISODE\tPROGRESS\tDURATION');
+    r.Items.forEach(i => {
+      const show = i.SeriesName || '';
+      const name = show ? `${i.Name}` : i.Name;
+      console.log([show || i.Name, show ? i.Name : '', pct(i.UserData, i.RunTimeTicks), dur(i.RunTimeTicks)].join('\t'));
+    });
+  },
+
+  async recent(limit = 10) {
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items/Latest?Limit=${limit}&Fields=RunTimeTicks`);
+    if (!r.length) return console.log('No recent items');
+    console.log('TYPE\tTITLE\tNAME\tYEAR\tDURATION');
+    r.forEach(i => {
+      console.log([i.Type === 'Episode' ? 'EP' : 'MOV', i.SeriesName || i.Name, i.SeriesName ? i.Name : '', i.ProductionYear || '', dur(i.RunTimeTicks)].join('\t'));
+    });
+  },
+
+  async watched(limit = 20) {
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items?Limit=${limit}&SortBy=DatePlayed&SortOrder=Descending&Filters=IsPlayed&Recursive=true&IncludeItemTypes=Movie,Episode&Fields=RunTimeTicks,DateLastMediaAdded`);
+    if (!r.Items.length) return console.log('No watched items');
+    console.log('TYPE\tTITLE\tEPISODE\tYEAR');
+    r.Items.forEach(i => {
+      console.log([i.Type === 'Episode' ? 'EP' : 'MOV', i.SeriesName || i.Name, i.SeriesName ? i.Name : '', i.ProductionYear || ''].join('\t'));
+    });
+  },
+
+  async shows() {
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items?IncludeItemTypes=Series&Recursive=true&SortBy=SortName&Fields=RunTimeTicks`);
+    if (!r.Items.length) return console.log('No shows');
+    console.log('TITLE\tYEAR\tSTATUS');
+    r.Items.forEach(i => {
+      console.log([i.Name, i.ProductionYear || '', pct(i.UserData, i.RunTimeTicks) || 'unwatched'].join('\t'));
+    });
+  },
+
+  async movies() {
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items?IncludeItemTypes=Movie&Recursive=true&SortBy=SortName&Fields=RunTimeTicks`);
+    if (!r.Items.length) return console.log('No movies');
+    console.log('TITLE\tYEAR\tDURATION\tSTATUS');
+    r.Items.forEach(i => {
+      console.log([i.Name, i.ProductionYear || '', dur(i.RunTimeTicks), pct(i.UserData, i.RunTimeTicks) || '-'].join('\t'));
+    });
+  },
+
+  async search(query) {
+    if (!query) { console.error('Usage: jellyfin search <query>'); process.exit(1); }
+    const uid = await getUserId();
+    const r = await api(`/Users/${uid}/Items?SearchTerm=${encodeURIComponent(query)}&Recursive=true&IncludeItemTypes=Movie,Series,Episode&Limit=10&Fields=RunTimeTicks`);
+    if (!r.Items.length) return console.log('No results');
+    console.log('TYPE\tTITLE\tNAME\tYEAR');
+    r.Items.forEach(i => {
+      console.log([i.Type, i.SeriesName || i.Name, i.SeriesName ? i.Name : '', i.ProductionYear || ''].join('\t'));
+    });
+  },
+
+  async stats() {
+    const uid = await getUserId();
+    const [movies, shows, episodes] = await Promise.all([
+      api(`/Users/${uid}/Items?IncludeItemTypes=Movie&Recursive=true&Filters=IsPlayed`),
+      api(`/Users/${uid}/Items?IncludeItemTypes=Series&Recursive=true`),
+      api(`/Users/${uid}/Items?IncludeItemTypes=Episode&Recursive=true&Filters=IsPlayed`),
+    ]);
+    console.log(`Movies watched:\t${movies.TotalRecordCount}`);
+    console.log(`Shows in library:\t${shows.TotalRecordCount}`);
+    console.log(`Episodes watched:\t${episodes.TotalRecordCount}`);
+  },
+
+  async libraries() {
+    const uid = getUserId();
+    const r = await api(`/Users/${uid}/Views`);
+    console.log('NAME\tTYPE\tID');
+    (r.Items || []).forEach(l => {
+      console.log([l.Name, l.CollectionType || 'mixed', l.Id].join('\t'));
+    });
+  }
+};
+
+const cmd = process.argv[2];
+const args = process.argv.slice(3);
+
+if (!cmd || !commands[cmd]) {
+  console.log('Usage: jellyfin <command> [args]\n');
+  console.log('Commands:');
+  console.log('  resume [limit]     Continue watching');
+  console.log('  recent [limit]     Recently added');
+  console.log('  watched [limit]    Watch history');
+  console.log('  shows              All series');
+  console.log('  movies             All movies');
+  console.log('  search <query>     Search library');
+  console.log('  stats              Watch statistics');
+  console.log('  libraries          List libraries');
+  process.exit(0);
+}
+
+commands[cmd](...args).catch(e => { console.error(e.message); process.exit(1); });
--- a/bin/brain-dump
+++ b/bin/brain-dump
@ -4,7 +4,7 @@ const fs = require('fs');
 const path = require('path');
 const crypto = require('crypto');

-const DUMP_PATH = path.join(__dirname, '..', 'memory', 'brain-dump.json');
+const DUMP_PATH = path.join(__dirname, '..', 'memory', 'tasks.json');

 const NUDGE_INTERVALS = {
  now: 24 * 60 * 60 * 1000,       // 1 day
@ -41,29 +41,31 @@ function formatTask(t) {
  return line;
 }

+function formatRecurring(r) {
+  let line = `${r.id}\t${r.frequency}\t${r.when || '-'}\t${r.text}`;
+  if (r.context) line += `\t${r.context}`;
+  return line;
+}
+
 // --- Commands ---

 function cmdList(args) {
  const data = load();
  let tasks = data.tasks || [];

-  // Filter by priority
  const prioArg = getFlag(args, '--priority');
  if (prioArg) {
    const prios = prioArg.split(',').map(s => s.trim());
    tasks = tasks.filter(t => prios.includes(t.priority));
  }

-  // Filter by due
  if (args.includes('--due')) {
    tasks = tasks.filter(isDue);
  }

-  // Sort: now > soon > someday
  const prioOrder = { now: 0, soon: 1, someday: 2 };
  tasks.sort((a, b) => (prioOrder[a.priority] ?? 3) - (prioOrder[b.priority] ?? 3));

-  // Limit
  const limitArg = getFlag(args, '--limit');
  if (limitArg) {
    tasks = tasks.slice(0, parseInt(limitArg, 10));
@ -80,72 +82,112 @@ function cmdList(args) {
 }

 function cmdAdd(args) {
+  const isRecurring = args.includes('--recurring');
  const text = getFlag(args, '--text');
  if (!text) { console.error('--text required'); process.exit(1); }
-  const priority = getFlag(args, '--priority') || 'soon';
  const context = getFlag(args, '--context') || undefined;

  const data = load();
-  const task = {
-    id: shortId(),
-    added: new Date().toISOString().slice(0, 10),
-    text,
-    priority,
-  };
-  if (context) task.context = context;
-  data.tasks = data.tasks || [];
-  data.tasks.push(task);
-  save(data);
-  console.log(`Added: ${task.id}\t${priority}\t${text}`);
+
+  if (isRecurring) {
+    const frequency = getFlag(args, '--frequency') || 'daily';
+    const when = getFlag(args, '--when') || undefined;
+    const item = { id: shortId(), text, frequency };
+    if (when) item.when = when;
+    if (context) item.context = context;
+    data.recurring = data.recurring || [];
+    data.recurring.push(item);
+    save(data);
+    console.log(`Added recurring: ${formatRecurring(item)}`);
+  } else {
+    const priority = getFlag(args, '--priority') || 'soon';
+    const task = { id: shortId(), added: new Date().toISOString().slice(0, 10), text, priority };
+    if (context) task.context = context;
+    data.tasks = data.tasks || [];
+    data.tasks.push(task);
+    save(data);
+    console.log(`Added: ${task.id}\t${priority}\t${text}`);
+  }
 }

 function cmdEdit(args) {
  const id = args[0];
-  if (!id) { console.error('Usage: brain-dump edit <id> [--text ...] [--priority ...] [--context ...]'); process.exit(1); }
+  if (!id) { console.error('Usage: tasks edit <id> [--text ...] [--priority ...] [--context ...] [--frequency ...] [--when ...]'); process.exit(1); }

  const data = load();
-  const task = (data.tasks || []).find(t => t.id === id);
-  if (!task) { console.error(`Task ${id} not found.`); process.exit(1); }
+
+  // Check tasks first, then recurring
+  let item = (data.tasks || []).find(t => t.id === id);
+  let isRecurring = false;
+  if (!item) {
+    item = (data.recurring || []).find(r => r.id === id);
+    isRecurring = true;
+  }
+  if (!item) { console.error(`Item ${id} not found.`); process.exit(1); }

  const text = getFlag(args, '--text');
-  const priority = getFlag(args, '--priority');
  const context = getFlag(args, '--context');
+  if (text) item.text = text;
+  if (context !== null && context !== undefined) item.context = context;

-  if (text) task.text = text;
-  if (priority) task.priority = priority;
-  if (context !== null && context !== undefined) task.context = context;
-
-  save(data);
-  console.log(`Updated: ${formatTask(task)}`);
+  if (isRecurring) {
+    const frequency = getFlag(args, '--frequency');
+    const when = getFlag(args, '--when');
+    if (frequency) item.frequency = frequency;
+    if (when) item.when = when;
+    save(data);
+    console.log(`Updated recurring: ${formatRecurring(item)}`);
+  } else {
+    const priority = getFlag(args, '--priority');
+    if (priority) item.priority = priority;
+    save(data);
+    console.log(`Updated: ${formatTask(item)}`);
+  }
 }

 function cmdDone(args) {
  const id = args[0];
-  if (!id) { console.error('Usage: brain-dump done <id>'); process.exit(1); }
+  if (!id) { console.error('Usage: tasks done <id>'); process.exit(1); }

  const data = load();
-  const idx = (data.tasks || []).findIndex(t => t.id === id);
-  if (idx === -1) { console.error(`Task ${id} not found.`); process.exit(1); }

-  const removed = data.tasks.splice(idx, 1)[0];
-  save(data);
-  console.log(`Done: ${removed.text}`);
+  // Check tasks
+  let idx = (data.tasks || []).findIndex(t => t.id === id);
+  if (idx !== -1) {
+    const removed = data.tasks.splice(idx, 1)[0];
+    save(data);
+    console.log(`Done: ${removed.text}`);
+    return;
+  }
+
+  // Check recurring
+  idx = (data.recurring || []).findIndex(r => r.id === id);
+  if (idx !== -1) {
+    const removed = data.recurring.splice(idx, 1)[0];
+    save(data);
+    console.log(`Removed recurring: ${removed.text}`);
+    return;
+  }
+
+  console.error(`Item ${id} not found.`);
+  process.exit(1);
 }

 function cmdShow(args) {
  const id = args[0];
-  if (!id) { console.error('Usage: brain-dump show <id>'); process.exit(1); }
+  if (!id) { console.error('Usage: tasks show <id>'); process.exit(1); }

  const data = load();
-  const task = (data.tasks || []).find(t => t.id === id);
-  if (!task) { console.error(`Task ${id} not found.`); process.exit(1); }
+  const item = (data.tasks || []).find(t => t.id === id)
+    || (data.recurring || []).find(r => r.id === id);
+  if (!item) { console.error(`Item ${id} not found.`); process.exit(1); }

-  console.log(JSON.stringify(task, null, 2));
+  console.log(JSON.stringify(item, null, 2));
 }

 function cmdNudged(args) {
  const ids = (args[0] || '').split(',').map(s => s.trim()).filter(Boolean);
-  if (ids.length === 0) { console.error('Usage: brain-dump nudged <id1>,<id2>,...'); process.exit(1); }
+  if (ids.length === 0) { console.error('Usage: tasks nudged <id1>,<id2>,...'); process.exit(1); }

  const data = load();
  const now = new Date().toISOString();
@ -161,19 +203,15 @@ function cmdNudged(args) {
  console.log(`Marked ${count} task(s) as nudged.`);
 }

-function cmdRecurring(args) {
+function cmdRecurring() {
  const data = load();
-  const when = getFlag(args, '--when');
-  let items = data.recurring || [];
-  if (when) {
-    items = items.filter(r => r.when === when);
-  }
+  const items = data.recurring || [];
  if (items.length === 0) {
    console.log('No recurring items.');
    return;
  }
  for (const r of items) {
-    console.log(`${r.id}\t${r.frequency}\t${r.when || '-'}\t${r.text}`);
+    console.log(formatRecurring(r));
  }
 }

@ -196,16 +234,17 @@ switch (cmd) {
  case 'done': cmdDone(args); break;
  case 'show': cmdShow(args); break;
  case 'nudged': cmdNudged(args); break;
-  case 'recurring': cmdRecurring(args); break;
+  case 'recurring': cmdRecurring(); break;
  default:
-    console.log(`Usage: brain-dump <command>
+    console.log(`Usage: tasks <command>

 Commands:
  list [--priority now,soon] [--due] [--limit N]
  add --text "..." --priority soon [--context "..."]
-  edit <id> [--text "..."] [--priority ...] [--context "..."]
+  add --recurring --text "..." --frequency daily [--when evening] [--context "..."]
+  edit <id> [--text "..."] [--priority|--frequency|--when|--context "..."]
  done <id>
  show <id>
  nudged <id1>,<id2>,...
-  recurring [--when evening]`);
+  recurring`);
 }
--- a/memory/2026-02-06.md
+++ b/memory/2026-02-06.md
@ -0,0 +1,23 @@
+# 2026-02-06 (Friday)
+
+## Key Events
+- **Opus 4.6 upgrade**: User pointed out OpenClaw update included Opus 4.6 support. Config was still on 4.5 — patched config to set Opus 4.6 as primary with 4.5 fallback. Updated TOOLS.md.
+- **Cron jobs fixed**: News cron jobs were broken all day — root cause was `wakeMode: "next-heartbeat"` causing jobs to be skipped when heartbeat timing didn't align. Fixed by switching all 8 news jobs to `wakeMode: "now"`. Tested successfully at 21:35 and 21:39 UTC.
+- **Cron delivery note**: What appears truncated in system logs is actually the full message on WhatsApp — I only see a summary of cron agent output.
+- **Task system revamp**: Restructured brain-dump.json — removed status field (done = delete), added priority system (now/soon/someday), updated HEARTBEAT.md with auto-capture and smarter nudging rules.
+- **AirPods listed on Willhaben**: Helped write title/description for AirPods Pro 2 USB-C listing. Task completed and removed from brain-dump.
+- **Bazzite 42→43 upgrade**: Helped user rebase GPD Win 4 from Bazzite 42 to 43. Used `ostree-unverified-registry` to bypass stuck ref. Upgrade successful.
+- **FSR4 setup**: Guided user through FSR4 emulated RDNA3 on GPD Win 4 HX 370. ProtonPlus → Proton GE → `DXIL_SPIRV_CONFIG=wmma_rdna3_workaround FSR4_UPGRADE=1`. Set up global env vars.
+- **GPD Win 4 notes**: HX 370 processor, BIOS 0.10 (latest for this model), 1080p display. 15W TDP = ~28W total system draw (normal). Clair Obscur needs ~20W TDP for 30fps at 1080p FSR performance.
+- **SVS + Watzek Petschinka reminders**: Set for Monday 9 AM Vienna.
+- **Codex CLI**: Explained plan mode (Shift+Tab to cycle modes: Ask/Plan/Auto).
+
+## Wind-down
+- 19:00-22:00 window: Failed to check in at 19:00 — user called me out. Must proactively reach out when wind-down window starts.
+- User worked until ~23:00 on Friday (hook issue → NixOS config → cron testing). Classic scope creep pattern.
+- Nose shower done at 23:01.
+
+## Lessons
+- **Always check in at 19:00 Vienna** when wind-down window starts, even if user seems quiet.
+- **Cron jobs**: Use `wakeMode: "now"` for isolated jobs that need precise timing. `next-heartbeat` is unreliable for scheduled tasks.
+- **Cron delivery**: I only see summaries of cron agent output, not the full message. Don't assume truncation on WhatsApp.
--- a/memory/2026-02-07.md
+++ b/memory/2026-02-07.md
@ -0,0 +1,71 @@
+# 2026-02-07 (Saturday)
+
+## Cron Bug Fix 🐛
+- **Root cause**: `recomputeNextRuns()` in `/app/src/cron/service/store.ts` (line ~409 in ensureLoaded) recomputes `nextRunAtMs` for ALL jobs before `runDueJobs()` runs. For recurring schedules (cron/every), `computeNextRunAtMs()` always returns the NEXT occurrence after `now`, advancing the schedule past the current time. `runDueJobs()` then finds no due jobs.
+- One-shot `at` jobs were unaffected because `computeJobNextRunAtMs` returns the original timestamp for them.
+- **Fix**: Patched `/app/dist/gateway-cli-D6nsc3s0.js` — in `recomputeNextRuns`, added check to skip recomputing if `nextRunAtMs <= now` (job is due, let it execute first).
+- **Impact**: All 8 news cron jobs (Der Standard + AI News × 4 times/day) had NEVER fired since creation on Feb 6.
+- Hard-killed gateway (PID 2) to force container restart and load patched JS bundle.
+- Verified: recurring `every` job fired correctly after patch. Real news jobs scheduled for 14:00 Vienna.
+- **TODO**: File upstream bug report on OpenClaw GitHub (issue in `recomputeNextRuns` race condition).
+
+## GPD Win 4 Optimization
+- User working on improving performance/watt on GPD Win 4 2025 (HX 370, Bazzite 43, FSR 4 enabled)
+- Discussed: TDP tuning (15-20W sweet spot), core parking via PowerTools Decky plugin, resolution/FSR upscaling, fan curves, frame limiting
+- Core parking: disable SMT, reduce threads to 4 for GPU-bound games
+- BIOS: UMA frame buffer option not available on HX 370 model (others report same)
+- Recommended removing ReShade since FSR 4 handles upscaling
+- ProtonGE: update every 2-4 weeks or when a game has issues
+
+## Android → Snapcast Streaming (Reference)
+- No native AirPlay-like solution for Android → Snapcast
+- **Best option:** DLNA/UPnP — gmrender-resurrect piped into Snapcast server + BubbleUPnP app on phone
+- **Alternatives:** Bluetooth A2DP sink (range limited), TCP source (no good Android app yet), AirPlay via Android app (hacky)
+- User interested but not urgent — revisit when asked
+
+## Cron Bug — Already Fixed Upstream
+- Bug fixed in OpenClaw **2026.2.6** (we're on 2026.2.4)
+- Changelog: "prevent recomputeNextRuns from skipping due jobs when timer fires late by reordering onTimer flow"
+- PR #10776 by @tyler6204
+- User wants to update OpenClaw (asked, not done yet)
+
+## Cron Jobs Consolidated
+- 8 news jobs → 2 jobs using multi-time cron expressions: `0 10,14,18,22 * * *`
+- Der Standard: job ID `917a96d5-14e7-4057-96e5-d83f2729ed9b`
+- AI News: job ID `a8991f6b-a31a-4997-bc92-b286573241d4`
+- Both confirmed working after fix (14:00 Vienna delivery successful)
+
+## Tasks CLI (`bin/tasks`)
+- Renamed from `brain-dump` to `tasks`
+- Data file: `memory/tasks.json` (was `memory/brain-dump.json`)
+- Commands: list, add, edit, done, show, nudged, recurring
+- Supports both regular tasks and recurring items
+- `--due` flag with nudge intervals: now=1d, soon=3d, someday=7d
+- Merged `--note` and `--context` into single `--context` field
+- Design decision: tasks tool is for actionable items ONLY, not knowledge storage. Reference info goes in memory files where `memory_search` finds it.
+
+## Audiobookshelf Integration
+- Built `bin/audiobooks` CLI tool
+- URL: https://audiobooks.cloonar.com
+- Credentials stored in `.credentials/audiobookshelf.env`
+- Commands: current, recent, finished, stats, libraries
+- Currently listening: "Die zweite Legion" (Askir Band 2) by Richard Schwartz — 44% done, 8h 46m remaining
+- Purpose: wind-down suggestions ("pick up your audiobook")
+
+## Infuse + Jellyfin
+- User uses Infuse app on Apple TV with Jellyfin server
+- Asked about syncing watch state from Infuse → Jellyfin
+- Infuse has native Jellyfin integration that auto-syncs watch state
+- InfuseSync plugin also available (github.com/firecore/InfuseSync)
+- Need to confirm: is user using Jellyfin connection or direct file share?
+- Task added: `47eaa129` — look into Jellyfin playing stats with Intune app on iOS
+
+## Improvement Ideas Discussed
+- Better wind-down: concrete suggestions using audiobook/TV data instead of vague nagging
+- Proactive calendar prep: weather, travel time, context from previous visits
+- Email triage: decided not worth it (user still uses mail client)
+- Workout: not needed (has working app already)
+- Decided: keep wind-down in main session (not separate agent) for better context
+
+## Calendar
+- Marie besuchen 13:00-17:00 (reminder sent at 11:30 Vienna)
--- a/memory/ainews-seen.txt
+++ b/memory/ainews-seen.txt
@ -1,15 +1,83 @@
+https://simonwillison.net/2026/Feb/6/an-update-on-heroku/#atom-everything
+https://simonwillison.net/2026/Feb/6/karel-doosterlinck/#atom-everything
+https://simonwillison.net/2026/Feb/5/ai-adoption-journey/#atom-everything
+https://simonwillison.net/2026/Feb/5/two-new-models/#atom-everything
+https://simonwillison.net/2026/Feb/5/the-world-factbook/#atom-everything
+https://simonwillison.net/2026/Feb/4/voxtral-2/#atom-everything
+https://simonwillison.net/2026/Feb/4/distributing-go-binaries/#atom-everything
+https://simonwillison.net/2026/Feb/3/introducing-deno-sandbox/#atom-everything
 https://simonwillison.net/2026/Feb/3/january/#atom-everything
 https://simonwillison.net/2026/Feb/3/brandon-sanderson/#atom-everything
 https://simonwillison.net/2026/Feb/2/introducing-the-codex-app/#atom-everything
 https://simonwillison.net/2026/Feb/2/no-humans-allowed/#atom-everything
 https://simonwillison.net/2026/Feb/1/openclaw-in-docker/#atom-everything
+https://simonwillison.net/2026/Jan/31/andrej-karpathy/#atom-everything
+https://simonwillison.net/2026/Jan/31/collective-efficacy/#atom-everything
+https://simonwillison.net/2026/Jan/30/steve-yegge/#atom-everything
+https://simonwillison.net/2026/Jan/30/moltbook/#atom-everything
+https://simonwillison.net/2026/Jan/30/a-programming-tool-for-the-arts/#atom-everything
+https://simonwillison.net/2026/Jan/29/datasette-10a24/#atom-everything
+https://simonwillison.net/2026/Jan/28/dynamic-features-static-site/#atom-everything
+https://simonwillison.net/2026/Jan/28/the-five-levels/#atom-everything
+https://simonwillison.net/2026/Jan/27/one-human-one-agent-one-browser/#atom-everything
+https://simonwillison.net/2026/Jan/27/kimi-k25/#atom-everything
+https://simonwillison.net/2026/Jan/26/tests/#atom-everything
+https://simonwillison.net/2026/Jan/26/chatgpt-containers/#atom-everything
+https://simonwillison.net/2026/Jan/25/the-browser-is-the-sandbox/#atom-everything
+https://simonwillison.net/2026/Jan/25/kakapo-cam/#atom-everything
+https://simonwillison.net/2026/Jan/24/dont-trust-the-process/#atom-everything
+https://simonwillison.net/2026/Jan/24/jasmine-sun/#atom-everything
+https://simonwillison.net/2026/Jan/23/fastrender/#atom-everything
+https://openai.com/policies/kr-privacy-policy
+https://openai.com/index/our-approach-to-localization
+https://openai.com/index/gpt-5-lowers-protein-synthesis-cost
+https://openai.com/index/trusted-access-for-cyber
+https://openai.com/index/introducing-openai-frontier
+https://openai.com/index/navigating-health-questions
+https://openai.com/index/introducing-gpt-5-3-codex
+https://openai.com/index/gpt-5-3-codex-system-card
+https://openai.com/index/unlocking-the-codex-harness
+https://openai.com/index/vfl-wolfsburg
 https://openai.com/index/sora-feed-philosophy
 https://openai.com/index/snowflake-partnership
 https://openai.com/index/introducing-the-codex-app
 https://openai.com/index/inside-our-in-house-data-agent
 https://openai.com/index/retiring-gpt-4o-and-older-models
+https://openai.com/index/taisei
+https://openai.com/index/emea-youth-and-wellbeing-grant
+https://openai.com/index/the-next-chapter-for-ai-in-the-eu
+https://openai.com/index/ai-agent-link-safety
+https://openai.com/index/pvh-future-of-fashion
+https://openai.com/index/trustbank
+https://openai.com/index/introducing-prism
+https://openai.com/index/indeed-maggie-hulce
+https://openai.com/index/unrolling-the-codex-agent-loop
+https://openai.com/index/scaling-postgresql
+https://openai.com/index/praktika
+https://openai.com/business/guides-and-resources/chatgpt-usage-and-adoption-patterns-at-work
+https://openai.com/index/higgsfield
+https://openai.com/index/edu-for-countries
+https://openai.com/index/how-countries-can-end-the-capability-overhang
 https://magazine.sebastianraschka.com/p/categories-of-inference-time-scaling
 https://magazine.sebastianraschka.com/p/state-of-llms-2025
 https://magazine.sebastianraschka.com/p/llm-research-papers-2025-part2
 https://magazine.sebastianraschka.com/p/technical-deepseek
 https://magazine.sebastianraschka.com/p/beyond-standard-llms
+https://magazine.sebastianraschka.com/p/llm-evaluation-4-approaches
+https://magazine.sebastianraschka.com/p/qwen3-from-scratch
+https://magazine.sebastianraschka.com/p/from-gpt-2-to-gpt-oss-analyzing-the
+https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison
+https://magazine.sebastianraschka.com/p/llm-research-papers-2025-list-one
+https://magazine.sebastianraschka.com/p/coding-the-kv-cache-in-llms
+https://magazine.sebastianraschka.com/p/coding-llms-from-the-ground-up
+https://magazine.sebastianraschka.com/p/the-state-of-llm-reasoning-model-training
+https://magazine.sebastianraschka.com/p/first-look-at-reasoning-from-scratch
+https://magazine.sebastianraschka.com/p/state-of-llm-reasoning-and-inference-scaling
+https://magazine.sebastianraschka.com/p/understanding-reasoning-llms
+https://magazine.sebastianraschka.com/p/ai-research-papers-2024-part-2
+https://magazine.sebastianraschka.com/p/ai-research-papers-2024-part-1
+https://magazine.sebastianraschka.com/p/llm-research-papers-the-2024-list
+https://magazine.sebastianraschka.com/p/understanding-multimodal-llms
+https://simonwillison.net/2026/Feb/6/tom-dale/#atom-everything
+https://simonwillison.net/2026/Feb/6/pydantic-monty/#atom-everything
+https://simonwillison.net/2026/Feb/7/software-factory/#atom-everything
--- a/memory/calendar-reminders-sent.json
+++ b/memory/calendar-reminders-sent.json
@ -1,3 +1,5 @@
 {
-  "sent": []
+  "sent": [
+    {"event": "Marie besuchen", "date": "2026-02-07", "remindedAt": "2026-02-07T09:31:00Z"}
+  ]
 }
--- a/memory/derstandard-seen.txt
+++ b/memory/derstandard-seen.txt
@ -1,49 +1,87 @@
-https://www.derstandard.at/story/3000000306979/trump-always-chickens-out-macht-taco-tatsaechlich-immer-einen-rueckzieher?ref=rss
-https://www.derstandard.at/story/3000000307063/usa-schiessen-iranische-drohne-nahe-flugzeugtraeger-im-arabischen-meer-ab?ref=rss
-https://www.derstandard.at/story/3000000306672/neue-epstein-files-werfen-fragen-an-adelige-diplomaten-und-eva-dichand-auf?ref=rss
-https://www.derstandard.at/story/3000000306965/wodurch-fast-vier-von-zehn-krebsfaellen-vermeidbar-waeren?ref=rss
-https://www.derstandard.at/story/3000000306903/etappensieg-fuer-saudischen-staatsfonds-mubadala-im-streit-mit-signa?ref=rss
-https://www.derstandard.at/story/3000000307018/das-gehalt-ueberzeugt-beschaeftigte-nicht-dafuer-die-kollegen?ref=rss
-https://www.derstandard.at/story/3000000307007/waymo-will-robotaxis-global-anbieten-und-sammelt-dafuer-neue-milliarden-ein?ref=rss
-https://www.derstandard.at/story/3000000307041/babys-erkennen-objekte-schon-im-alter-von-zwei-monaten?ref=rss
-https://www.derstandard.at/story/3000000306987/welche-bedrohung-durch-russland-ist-realistisch?ref=rss
-https://www.derstandard.at/story/3000000306974/minister-wiederkehr-warum-weniger-latein-im-gymnasium-mehr-humanismus-bedeutet?ref=rss
-https://www.derstandard.at/story/3000000306758/olympia-2026-zerstueckelte-winterspiele-sind-die-letzte-chance?ref=rss
-https://www.derstandard.at/story/3000000303908/warum-online-bewerbungsgespraeche-so-schwierig-sind-und-wie-man-dennoch-punktet?ref=rss
-https://www.derstandard.at/story/3000000306921/psychiaterin-kastner-ueber-kindstoetung-das-groesste-risiko-ist-eine-unbehandelte-depression?ref=rss
-https://www.derstandard.at/story/3000000307052/trotz-kreuzbandrisses-vonn-haelt-an-olympia-start-fest?ref=rss
-https://www.derstandard.at/story/3000000306703/vermoegen-sinnvoll-anlegen-eigentum-miete-oder-alternative-geldanlage?ref=rss
-https://www.derstandard.at/story/3000000306735/krankenstand-wegen-skiunfalls-wann-darf-der-arbeitgeber-das-gehalt-verweigern?ref=rss
-https://www.derstandard.at/story/3000000306990/maja-t-droht-in-budapester-antifa-prozess-ein-hartes-urteil?ref=rss
-https://www.derstandard.at/story/3000000306855/kunstsammler-leon-black-war-jeffrey-epsteins-haupteinnahmequelle?ref=rss
-https://www.derstandard.at/story/3000000306890/epstein-aff228re-ehepaar-clinton-zu-aussage-vor-dem-kongress-bereit?ref=rss
-https://www.derstandard.at/story/3000000306960/konflikt-bei-den-salzburger-festspielen?ref=rss
-https://www.derstandard.at/story/3000000307071/burgenlands-ex-landeshauptmann-niessl-will-f252r-hofburg-kandidieren?ref=rss
-https://www.derstandard.at/story/3000000307072/gaddafi-sohn-seif-al-islam-gestorben?ref=rss
-https://www.derstandard.at/story/3000000306899/prozess-wegen-sexuellen-missbrauchs-gegen-ex-politiker-stronach-beginnt-in-kanada?ref=rss
-https://www.derstandard.at/story/3000000307067/anklage-will-wahlausschluss-f252r-le-pen-aber-nicht-sofort?ref=rss
-https://www.derstandard.at/story/3000000307066/auf-kuba-hat-es-null-grad?ref=rss
-https://www.derstandard.at/story/3000000307006/trump-fordert-von-elite-uni-harvard-eine-milliarde-dollar-schadenersatz?ref=rss
-https://www.derstandard.at/story/3000000306975/beschuss-bei-17-grad-menschen-suchen-schutz-in-kyjiws-u-bahn-stationen?ref=rss
-https://www.derstandard.at/story/3000000306996/tod-eines-haeftlings-expertenkommission-soll-strafvollzug-verbessern?ref=rss
-https://www.derstandard.at/story/3000000306831/skelettierte-frauenleiche-in-wien-favoriten-entdeckt?ref=rss
-https://www.derstandard.at/story/3000000306971/von-grammys-bis-kennedy-center-der-maga-kulturkampf-folgt-einer-strategie?ref=rss
-https://www.derstandard.at/story/3000000306968/nicht-die-ki-ist-ausser-kontrolle-sondern-der-mensch?ref=rss
-https://www.derstandard.at/story/3000000306772/latein-kuerzen-lieber-eine-bildungsreform-die-wir-brauchen?ref=rss
-https://www.derstandard.at/story/3000000306775/weniger-latein-humanistische-bildung-ist-kein-ballast?ref=rss
-https://www.derstandard.at/story/3000000305513/wunderheilung-im-winter?ref=rss
-https://www.derstandard.at/story/3000000306811/machtkampf-in-china-mit-unberechenbaren-folgen?ref=rss
-https://www.derstandard.at/story/3000000306670/eine-regierung-die-gestalten-will-sollte-mehr-ideen-haben-als-eine-volksbefragung?ref=rss
+https://www.derstandard.at/story/3000000306839/sex-geld-macht-was-war-das-geheimnis-des-jeffrey-epstein?ref=rss
+https://www.derstandard.at/story/3000000306835/stocker-europa-wird-aus-vielen-richtungen-bedroht?ref=rss
+https://www.derstandard.at/story/3000000307558/kritik-an-rassistischem-trump-post-ueber-barack-und-michelle-obama?ref=rss
+https://www.derstandard.at/story/3000000307468/der-bitcoin-crasht-die-szene-laechelt-muede?ref=rss
+https://www.derstandard.at/story/3000000307239/batterie-ohne-lithium-warum-china-jetzt-auf-druckluftspeicher-setzt?ref=rss
+https://www.derstandard.at/story/3000000307485/so-gross-wie-ein-schulbus-phantomqualle-in-atlantik-gefilmt?ref=rss
+https://www.derstandard.at/story/3000000307236/jetzt-weiss-jeder-wo-nuuk-ist-wie-sich-trump-auf-den-groenland-tourismus-auswirkt?ref=rss
+https://www.derstandard.at/story/3000000307471/die-babler-demontage-ist-verrueckt?ref=rss
+https://www.derstandard.at/story/3000000307488/die-epstein-files-sind-eine-gewollte-ueberforderung-der-gesellschaft?ref=rss
+https://www.derstandard.at/story/3000000307363/latein-und-ein-aufstand-der-eliten?ref=rss
+https://www.derstandard.at/story/3000000306982/darin-gleicht-donald-trump-fast-allen-diktatoren?ref=rss
+https://www.derstandard.at/story/3000000307480/nervoes-sam-altman-zuckt-wegen-spott-von-anthropic-aus?ref=rss
+https://www.derstandard.at/story/3000000307472/der-herr-der-kameras-orf-regisseur-koegler-inszeniert-die-olympia-abfahrt?ref=rss
+https://www.derstandard.at/story/3000000307321/quiz-zehn-fragen-um-die-welt?ref=rss
+https://www.derstandard.at/story/3000000307578/polens-parlamentspraesident-gegen-nobelpreis-fuer-trump-us-botschafter-empoert?ref=rss
+https://www.derstandard.at/story/3000000307466/die-eu-kommission-testet-alternativen-zu-microsoft-teams?ref=rss
+https://www.derstandard.at/story/3000000300279/oesterreichs-fussball-bundesliga-ist-so-schlecht-wie-seit-jahrzehnten-nicht?ref=rss
+https://www.derstandard.at/story/3000000305885/als-ai-weiwei-seinen-mittelfinger-dem-tiananmen-platz-entgegenstreckte?ref=rss
+https://www.derstandard.at/story/3000000307479/baeume-in-den-dolomiten-haben-wohl-doch-keine-sonnenfinsternis-vorhergesehen?ref=rss
+https://www.derstandard.at/story/3000000306805/der-standard-preismonitor-diese-produkte-wurden-am-teuersten?ref=rss
+https://www.derstandard.at/story/3000000307527/attentat-in-moskau-auf-stellvertretenden-leiter-des-militaergeheimdienstes?ref=rss
+https://www.derstandard.at/story/3000000307544/abfahrt-mit-kreuzbandriss-trainer-svindal-glaubt-an-lindsey-vonns-medaillenchance?ref=rss
+https://www.derstandard.at/story/3000000307177/worauf-wir-uns-bei-olympia-am-meisten-freuen?ref=rss
+https://www.derstandard.at/story/3000000307391/zweite-runde-mit-dichtem-programm-woeginger-muss-ab-mittwoch-zurueck-auf-die-anklagebank?ref=rss
+https://www.derstandard.at/story/3000000307505/-frau-in-wien-nach-mutma223lichem-sexualdelikt-in-lebensgefahr?ref=rss
+https://www.derstandard.at/story/3000000307556/abermals-kritik-vor-erster-phase-des-kopftuchverbots-an-schulen?ref=rss
+https://www.derstandard.at/story/3000000307559/viele-tote-bei-anschlag-auf-moschee-in-islamabad?ref=rss
+https://www.derstandard.at/story/3000000307557/ermittlungen-gegen-norwegischen-ex-regierungschef-wegen-epstein-affaere?ref=rss
+https://www.derstandard.at/story/3000000307515/washington-post-redaktion-protestiert-gegen-kahlschlag?ref=rss
+https://www.derstandard.at/story/3000000307486/stocker-fuehrt-politisches-armdruecken-in-der-koalition-ein-zu-seinem-schaden?ref=rss
+https://www.derstandard.at/story/3000000307554/schon-zu-viel-trump?ref=rss
+https://www.derstandard.at/story/3000000307489/der-wert-von-lebensmitteln-sollte-nicht-am-preis-haengen?ref=rss
+https://www.derstandard.at/story/3000000307357/schriftsteller-thomas-raab-latein-als-humanistische-eintrittskarte?ref=rss
+https://www.derstandard.at/story/3000000307333/die-russische-epstein-connection?ref=rss
+https://www.derstandard.at/story/3000000307336/am-niedergang-der-washington-post-ist-nicht-nur-jeff-bezos-schuld?ref=rss
+https://www.derstandard.at/story/3000000307138/wie-lange-noch-bleiben-sozialunternehmen-die-unsichtbaren-optimisten?ref=rss
 https://www.derstandard.at/story/3000000253356/finde-den-fehler?ref=rss
 https://www.derstandard.at/story/3000000306814/cartoons-februar?ref=rss
-https://www.derstandard.at/story/3000000306730/was-ist-euer-lieblings-club-wiens?ref=rss
-https://www.derstandard.at/story/3000000306462/wenn-der-letzte-wille-nicht-dem-zufall-ueberlassen-werden-soll?ref=rss
-https://www.derstandard.at/story/3000000306728/seid-ihr-auch-von-gestiegenen-svs-beitraegen-betroffen?ref=rss
-https://www.derstandard.at/story/3000000306902/teilauszahlung-des-gehalts-in-bitcoin-eure-erfahrungen?ref=rss
-https://www.derstandard.at/story/3000000306593/ki-statt-latein-7-kritische-fragen-zur-reform?ref=rss
-https://www.derstandard.at/story/3000000306299/neuigkeiten-im-gluecksspielrecht-lootboxen-und-geschaeftsfuehrerhaftung?ref=rss
-https://www.derstandard.at/story/3000000306729/sollten-geschaefte-regulaer-sonntags-oeffnen?ref=rss
-https://www.derstandard.at/story/3000000306767/wie-oft-schreibt-ihr-noch-mit-der-hand?ref=rss
-https://www.derstandard.at/story/3000000306139/was-kostet-ein-verlorenes-tier-rechtlich-gesehen?ref=rss
-https://www.derstandard.at/story/3000000305258/das-virtuelle-labor-im-weltraum?ref=rss
-https://www.derstandard.at/story/3000000306940/befreiungsschlag-als-bumerang-stockers-volksabstimmung-verstimmt-auch-in-der-oevp?ref=rss
+https://www.derstandard.at/story/3000000307458/nutzt-ihr-amazon-fuer-die-taeglichen-verbrauchsartikel?ref=rss
+https://www.derstandard.at/story/3000000307065/profitieren-wirklich-alle-aktive-fonds-und-lebensversicherungen?ref=rss
+https://www.derstandard.at/story/3000000306520/kurzzeitvermietung-in-wien-das-ende-von-airbnb-als-geschaeftsmodell?ref=rss
+https://www.derstandard.at/story/3000000307531/ist-latein-mehr-als-nur-ein-unterrichtsfach?ref=rss
+https://www.derstandard.at/story/3000000307305/was-sind-ihre-erfahrungen-als-zugbegleiterinnen?ref=rss
+https://www.derstandard.at/story/3000000306657/lueger-ehrenmal-ein-monument-der-gescheiterten-entnazifizierung?ref=rss
+https://www.derstandard.at/story/3000000307296/thema-grundwehrdienstverlaengerung-was-sind-eure-erfahrungen?ref=rss
+https://www.derstandard.at/story/3000000306485/doppelte-botanik-pflanzen-mit-geschichten-und-zum-geniessen?ref=rss
+https://www.derstandard.at/story/3000000307370/was-sind-eure-erinnerungen-an-die-olympischen-winterspiele?ref=rss
+https://www.derstandard.at/story/3000000303147/zwischen-information-und-image-was-aerztliche-werbung-darf?ref=rss
+https://www.derstandard.at/story/3000000307416/usa-und-eu-kooperieren-bei-seltenen-erden-warum-sie-keine-andere-wahl-haben?ref=rss
+https://www.derstandard.at/story/3000000307497/schwaerzungen-und-kontrollverlust-warum-epsteins-opfer-die-veroeffentlichungen-kritisieren?ref=rss
+https://www.derstandard.at/story/3000000307598/wehrpflicht-neos-gegen-volksbefragung?ref=rss
+https://www.derstandard.at/story/3000000307127/cora-liess-sich-mit-23-sterilisieren-so-krassen-hate-habe-ich-nie-bekommen?ref=rss
+https://www.derstandard.at/story/3000000307156/fuer-die-eiserne-reserve-der-benkos-wird-es-eng-was-steckt-in-der-laura-privatstiftung?ref=rss
+https://www.derstandard.at/story/3000000307588/grossvater-schoss-in-oberoesterreich-auf-30-jaehrige-enkelin?ref=rss
+https://www.derstandard.at/story/3000000307564/epsteins-prominente-verstrickungen-in-die-kunstwelt-haben-ein-nachspiel?ref=rss
+https://www.derstandard.at/story/3000000307568/minnesotas-ice-watchers-how-tactics-of-1960s-radicals-went-mainstream?ref=rss
+https://www.derstandard.at/story/3000000307500/fleisch-die-empoerung-im-schnitzel-land?ref=rss
+https://www.derstandard.at/story/3000000307322/haeppchenweise-tuerkisch-im-neuen-kuebey?ref=rss
+https://www.derstandard.at/story/3000000307110/goldverkauf-leicht-gemacht?ref=rss
+https://www.derstandard.at/story/3000000307561/geeetech-m1s-im-test-kompetenter-3d-drucker-fuer-kids?ref=rss
+https://www.derstandard.at/story/3000000306765/abschied-vom-schlechten-kaffee-in-wien?ref=rss
+https://www.derstandard.at/story/3000000307596/zwischen-finger-am-abzug-und-gutem-start-gemischte-reaktionen-im-iran-auf-verhandlungen?ref=rss
+https://www.derstandard.at/story/3000000307331/nasa-astronaut-bowen-in-wien-der-mond-ist-nur-der-anfang?ref=rss
+https://www.derstandard.at/story/3000000307225/neue-ausstellung-belegt-kunst-braucht-kommerz?ref=rss
+https://www.derstandard.at/story/3000000307539/bitcoin-crash-wie-es-dazu-kam-und-wie-weit-es-noch-abwaerts-gehen-kann?ref=rss
+https://www.derstandard.at/story/3000000307584/ich-habe-keinen-fehler-gemacht-trump-distanziert-sich-von-rassistische-posting-ueber-die-obamas?ref=rss
+https://www.derstandard.at/story/3000000307604/nach-ofarims-aussagen-zum-davidstern-verfahren-reagiert-sein-anwalt?ref=rss
+https://www.derstandard.at/story/3000000307605/bundestag-verweigerte-afd-personal-laut-parteiangaben-ausstellung-von-hausausweisen?ref=rss
+https://www.derstandard.at/story/3000000307533/sahel-usa-ruecken-wieder-naeher-europa-ringt-derweil-um-einen-kurs?ref=rss
+https://www.derstandard.at/story/3000000307601/von-allmen-holt-olympiagold-in-der-abfahrt-oesterreicher-verpassen-medaille?ref=rss
+https://www.derstandard.at/story/3000000307595/-zur-prostitution-gezwungen-und-entf252hrt-junge-frau-in-n214-befreit?ref=rss
+https://www.derstandard.at/story/3000000307120/braucht-es-2026-ueberhaupt-noch-feminismus?ref=rss
+https://www.derstandard.at/story/3000000307420/wie-gross-sollte-ein-kinderzimmer-sein?ref=rss
+https://www.derstandard.at/story/3000000307605/usa-wollen-laut-selenskyj-kriegsende-noch-vor-dem-sommer?ref=rss
+https://www.derstandard.at/story/3000000306765/profitieren-wirklich-alle-aktive-fonds-und-lebensversicherungen?ref=rss
+https://www.derstandard.at/story/3000000306828/in-den-usa-haben-nazi-vergleiche-hochkonjunktur?ref=rss
+https://www.derstandard.at/story/3000000306837/moltbook-apokalypse-und-shitposting?ref=rss
+https://www.derstandard.at/story/3000000306928/vom-ich-zum-du-100-tage-als-personenbetreuer-beduerftiger-menschen?ref=rss
+https://www.derstandard.at/story/3000000307507/warum-thailand-bei-der-cannabis-legalisierung-zurueckgerudert-ist?ref=rss
+https://www.derstandard.at/story/3000000306092/zukunftsforscher-gen-z-boomer-millennials-sind-keine-guten-kategorien?ref=rss
+https://www.derstandard.at/story/3000000307298/lady-gaga-adrien-brody-friends-hoehepunkte-der-super-bowl-spots?ref=rss
+https://www.derstandard.at/story/3000000306983/gesuender-essen-im-alltag-11-expertentipps-die-zeigen-wie-es-gelingen-kann?ref=rss
+https://www.derstandard.at/story/3000000307609/buhrufe-fuer-vance-bei-olympia-eroeffnung-waren-im-us-tv-nicht-zu-hoeren?ref=rss
+https://www.derstandard.at/story/3000000306833/olympia-biathletin-slettemark-niemand-in-groenland-will-zum-amerikaner-werden?ref=rss
+https://www.derstandard.at/story/3000000303500/kunstduenger-ruiniert-millionen-alte-prozesse-im-boden?ref=rss
+https://www.derstandard.at/story/3000000307033/wo-sind-die-games-fuer-einen-einzelnen-abend?ref=rss
+https://www.derstandard.at/story/3000000307298/warum-thailand-bei-der-cannabis-legalisierung-zurueckgerudert-ist?ref=rss
--- a/memory/heartbeat-state.json
+++ b/memory/heartbeat-state.json
@ -3,14 +3,45 @@
    "usa_iran_attack": true,
    "rheinmetall_above_1950": false,
    "steam_hardware_price": false,
-    "steam_hardware_release_date": false
+    "steam_hardware_release_date": false,
+    "hamr_40tb_release": false
  },
  "lastChecks": {
    "news": "2026-01-30T08:17:00Z",
-    "rheinmetall": "2026-02-03T22:00:00Z",
-    "rheinmetall_price": 1773.50,
-    "calendar": "2026-02-03T22:00:00Z",
+    "rheinmetall": "2026-02-07T08:41:00Z",
+    "rheinmetall_price": 1587.00,
+    "calendar": "2026-02-04T04:25:00Z",
    "steam_hardware": "2026-02-03T22:00:00Z",
-    "notes": "RHM: €1,773.50, below threshold. Calendar Feb 4: 'nyc' meeting at 12:00 (video call, no action). Steam hardware: still 'early 2026', no official price/date."
+    "hamr_40tb": "2026-02-04T21:10:00Z",
+    "notes": "RHM: €1,902.00, below €1,950 threshold. Steam hardware: still 'early 2026', no official price/date. HAMR 40TB: expected mid-2026, no release date yet."
+  },
+  "watchlist": {
+    "hamr_40tb": {
+      "description": "Seagate 40TB HAMR drives (IronWolf Pro / Exos)",
+      "expectedRelease": "mid-2026",
+      "notifyOn": "release with price",
+      "lastCheck": "2026-02-04",
+      "status": "not yet released"
+    },
+    "steam_hardware": {
+      "description": "Steam Machine / Steam Frame",
+      "expectedRelease": "early 2026",
+      "notifyOn": "official price or release date",
+      "status": "announced, no price/date"
+    },
+    "openclaw_opus46": {
+      "description": "OpenClaw support for Claude Opus 4.6",
+      "pr": "#9863",
+      "notifyOn": "PR merged or new release with opus-4-6 support",
+      "lastCheck": "2026-02-05",
+      "status": "PR open, not merged"
+    },
+    "claude_code_update": {
+      "description": "Claude Code updates",
+      "notifyOn": "new version released",
+      "lastCheck": "2026-02-05",
+      "lastKnownVersion": null,
+      "status": "watching"
+    }
  }
 }
--- a/memory/brain-dump.json
+++ b/memory/brain-dump.json
@ -2,35 +2,35 @@
  "description": "Task tracking — auto-captured from chat or manually added. Done tasks get deleted.",
  "recurring": [
    {
-      "id": "nose-shower",
+      "id": "843efabf",
      "text": "Nose shower 👃🚿",
      "frequency": "daily",
      "when": "evening",
-      "note": "Remind in evening, whether work just wrapped up or already relaxing"
+      "context": "Remind in evening, whether work just wrapped up or already relaxing"
    }
  ],
  "tasks": [
    {
-      "id": "forgejo-script",
-      "added": "2026-01-31",
+      "id": "b9d1d89e",
+      "added": "2026-02-07",
      "text": "Create script to auto-update workflows to Forgejo monorepo actions",
-      "context": "Monorepo is done, using GitHub actions as fallback for now",
-      "priority": "someday"
+      "priority": "someday",
+      "context": "Monorepo is done, using GitHub actions as fallback for now"
    },
    {
-      "id": "forgejo-mcp-server",
-      "added": "2026-01-31",
+      "id": "af198211",
+      "added": "2026-02-07",
      "text": "Set up MCP server for Forgejo instance",
-      "context": "Phase 1: MCP server for interactive Claude web sessions (fork + PR workflow). Phase 2: Automated issue-solving pipeline with clarification comments.",
      "priority": "soon",
-      "lastNudged": "2026-02-07T13:59:36.313Z"
+      "context": "Phase 1: MCP server for interactive Claude web sessions (fork + PR workflow). Phase 2: Automated issue-solving pipeline with clarification comments.",
+      "lastNudged": "2026-02-07T14:11:07.284Z"
    },
    {
-      "id": "typo3-v13-gbv",
-      "added": "2026-02-01",
+      "id": "0902aaba",
+      "added": "2026-02-07",
      "text": "TYPO3 v13 upgrade for GBV",
      "priority": "soon",
-      "lastNudged": "2026-02-07T13:59:36.313Z"
+      "lastNudged": "2026-02-07T14:11:07.284Z"
    }
  ]
 }
--- a/memory/tv-shows.json
+++ b/memory/tv-shows.json
@ -0,0 +1,145 @@
+{
+  "description": "TV shows tracking for recommendations",
+  "shows": [
+    {
+      "title": "Monarch: Legacy of Monsters",
+      "status": "watching",
+      "startedWatching": "2026-02-04",
+      "currentEpisode": 2,
+      "platform": "Apple TV+",
+      "genre": ["Sci-Fi", "Fantasy", "Drama", "Action & Adventure"],
+      "universe": "MonsterVerse (Godzilla)",
+      "rating": null,
+      "notes": "Features Kurt Russell & Wyatt Russell. Ask when finished."
+    },
+    {
+      "title": "Fallout",
+      "status": "watching",
+      "season": 2,
+      "platform": "Prime Video",
+      "genre": ["Sci-Fi", "Post-apocalyptic", "Drama"],
+      "notes": "Season 2 finale ready to watch as of Feb 4"
+    },
+    {
+      "title": "Prodigal Son",
+      "status": "watched",
+      "genre": ["Crime", "Drama", "Thriller"],
+      "notes": "Was watching Feb 3"
+    },
+    {
+      "title": "Hijack",
+      "status": "watching",
+      "season": 2,
+      "platform": "Apple TV+",
+      "genre": ["Thriller", "Drama"],
+      "notes": "S2 finale drops March 4, 2026"
+    }
+  ],
+  "history": [
+    {
+      "title": "Wheel of Time",
+      "rating": "liked",
+      "genre": ["Fantasy", "Adventure"]
+    },
+    {
+      "title": "Stargate SG-1",
+      "rating": "liked",
+      "genre": ["Sci-Fi", "Adventure", "Military"]
+    },
+    {
+      "title": "Stargate Atlantis",
+      "rating": "liked",
+      "genre": ["Sci-Fi", "Adventure", "Military"]
+    },
+    {
+      "title": "Doctor Who",
+      "rating": "liked",
+      "genre": ["Sci-Fi", "Adventure", "Time travel"]
+    },
+    {
+      "title": "Chuck",
+      "rating": "liked",
+      "genre": ["Action", "Comedy", "Spy"],
+      "notes": "Light-hearted spy action"
+    },
+    {
+      "title": "Sherlock",
+      "rating": "liked",
+      "genre": ["Crime", "Mystery", "Drama"]
+    },
+    {
+      "title": "Psych",
+      "rating": "liked",
+      "genre": ["Comedy", "Crime", "Mystery"],
+      "notes": "Light-hearted crime procedural"
+    },
+    {
+      "title": "Big Bang Theory",
+      "rating": "liked",
+      "genre": ["Sitcom", "Comedy"]
+    },
+    {
+      "title": "King of Queens",
+      "rating": "liked",
+      "genre": ["Sitcom", "Comedy"]
+    },
+    {
+      "title": "Stranger Things",
+      "rating": "liked early seasons",
+      "genre": ["Sci-Fi", "Fantasy", "Horror"],
+      "notes": "First seasons good, later seasons declined"
+    },
+    {
+      "title": "Breaking Bad",
+      "rating": "liked",
+      "genre": ["Drama", "Crime", "Thriller"]
+    },
+    {
+      "title": "Ted Lasso",
+      "rating": "liked",
+      "genre": ["Comedy", "Feel-good"]
+    },
+    {
+      "title": "The Boys",
+      "rating": "liked",
+      "genre": ["Action", "Superhero", "Dark comedy"]
+    },
+    {
+      "title": "Jack Ryan",
+      "rating": "liked",
+      "genre": ["Action", "Thriller", "Spy"]
+    },
+    {
+      "title": "Reacher",
+      "rating": "liked",
+      "genre": ["Action", "Crime", "Thriller"]
+    },
+    {
+      "title": "Game of Thrones",
+      "rating": "no",
+      "genre": ["Fantasy", "Drama"],
+      "notes": "Didn't enjoy - but likes other fantasy (Wheel of Time)"
+    },
+    {
+      "title": "The Office (US)",
+      "rating": "meh",
+      "genre": ["Comedy", "Mockumentary"]
+    }
+  ],
+  "notSeen": ["The Expanse", "Succession", "The Bear", "True Detective", "Mindhunter", "Ozark", "What We Do in the Shadows"],
+  "preferences": {
+    "likedGenres": ["Sci-Fi", "Action", "Thriller", "Spy", "Crime/Mystery", "Sitcom", "Feel-good comedy", "Fantasy/Adventure"],
+    "patterns": [
+      "Big sci-fi fan (Stargate, Doctor Who, Fallout)",
+      "Enjoys fantasy adventure (Wheel of Time) - GoT was exception not rule",
+      "Loves action thrillers (Reacher, Jack Ryan, The Boys)",
+      "Appreciates light-hearted comedy-action blends (Chuck, Psych)",
+      "Classic sitcom fan (Big Bang Theory, King of Queens)",
+      "Likes clever mysteries (Sherlock)",
+      "Quality drama with tight storytelling (Breaking Bad)",
+      "Prefers shows that stay focused - lost interest when Stranger Things dragged"
+    ],
+    "notLiked": ["Game of Thrones specifically", "Mockumentary style (Office was meh)"],
+    "notes": ["Broad taste but leans toward fun/adventurous over grimdark"]
+  }
+}
--- a/memory/wind-down-log.json
+++ b/memory/wind-down-log.json
@ -1,223 +1,8 @@
 {
-  "goal": {
-    "lastTaskDone": "22:00",
-    "windDownPeriod": "22:00-00:00",
-    "bedtime": "00:00",
-    "windDownNeeded": "~2 hours"
-  },
-  "timezone": "Europe/Vienna",
-  "learningPhase": true,
+  "date": "2026-02-07",
  "entries": [
-    {
-      "date": "2026-01-30",
-      "time": "23:14",
-      "activity": "nose shower (evening routine)",
-      "note": "does this every evening"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:20",
-      "activity": "tinkering with HA automation",
-      "note": "setting up morning briefing hook"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:28",
-      "activity": "setting up CalDAV access for me",
-      "note": "still working on integrations"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:37",
-      "activity": "still working on HA morning hook",
-      "note": "acknowledged wants to finish, 23 min to midnight"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:44",
-      "activity": "testing morning briefing hook",
-      "note": "first test - CalDAV didn't work"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:46",
-      "activity": "testing morning briefing hook",
-      "note": "second test - CalDAV worked!"
-    },
-    {
-      "date": "2026-01-30",
-      "time": "23:49",
-      "activity": "starting wind-down",
-      "note": "watching something, hot beverage"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "01:10",
-      "activity": "bedtime routine",
-      "note": "brushing teeth, putting cats out"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "02:15",
-      "activity": "going to sleep",
-      "note": "watched in bed, surfed internet"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "19:04",
-      "activity": "Gitea → Forgejo migration",
-      "note": "been working for hours already, now on workflow migration to new actions monorepo",
-      "taskType": "infrastructure/devops tinkering"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "19:13",
-      "activity": "decided to defer script to tomorrow",
-      "note": "nudge worked! chose pragmatic stopping point: point to GitHub actions, test workflow, script tomorrow",
-      "nudgeResult": "accepted"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "19:56",
-      "activity": "finished testing workflows",
-      "note": "tested 2 workflows, both worked. Final switch to Forgejo planned for tomorrow morning. Task wrapped up successfully before 20:00!"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "20:13",
-      "activity": "starting semi wind-down (watching something)",
-      "note": "usually finishes before 22:00 - on track for goal!"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "22:05",
-      "activity": "bedtime routine starting (brushing teeth)",
-      "note": "right on schedule! brain-dumped MCP/automation ideas first, then wind-down"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "22:32",
-      "activity": "in bed winding down",
-      "note": "cat, tea, audiobook - proper relaxation mode"
-    },
-    {
-      "date": "2026-01-31",
-      "time": "23:27",
-      "activity": "going to sleep",
-      "note": "33 minutes BEFORE midnight goal!"
-    },
-    {
-      "date": "2026-02-01",
-      "time": "19:16",
-      "activity": "relaxing, watching a series",
-      "note": "already winding down by 19:16 on Sunday - great start!"
-    },
-    {
-      "date": "2026-02-01",
-      "time": "20:36",
-      "activity": "nose shower done",
-      "note": "completed after episode ended, reminder at 20:10 worked"
-    },
-    {
-      "date": "2026-02-01",
-      "time": "23:45",
-      "activity": "going to sleep",
-      "note": "15 minutes before midnight goal - on track!"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "01:01",
-      "activity": "couldn't fall asleep, ate a bit, trying again",
-      "note": "trouble sleeping despite good wind-down timing"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "19:22",
-      "activity": "setting up ci-templates and testing with gbv-aktuell",
-      "note": "working on workflow optimization from earlier discussion"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "22:10",
-      "activity": "stopped working",
-      "note": "fixed Deployer issue, called it a day. I failed to suggest stopping points earlier (19:00-22:00)"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "22:33",
-      "activity": "nose shower reminder",
-      "note": "requested 23 min delay after first reminder"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "22:39",
-      "activity": "nose shower done",
-      "note": "completed"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "22:46",
-      "activity": "identity setup chat",
-      "note": "chose Hoid as my identity - relaxed conversation while winding down"
-    },
-    {
-      "date": "2026-02-02",
-      "time": "23:51",
-      "activity": "in bed winding down",
-      "note": "watching/relaxing more"
-    },
-    {
-      "date": "2026-02-03",
-      "time": "00:05",
-      "activity": "trying to sleep",
-      "note": "5 min past midnight goal"
-    },
-    {
-      "date": "2026-02-03",
-      "time": "19:15",
-      "activity": "watching Prodigal Son (TV)",
-      "note": "visited Marie earlier (16:00), now relaxing with a series - good wind-down activity"
-    },
-    {
-      "date": "2026-02-03",
-      "time": "21:40",
-      "activity": "arrived home",
-      "note": "HomeArrival hook fired, sent nose shower reminder"
-    }
-  ],
-  "patterns": {
-    "2026-01-30": {
-      "stoppedWork": "23:49",
-      "startedBedRoutine": "01:10",
-      "actualBedtime": "02:15",
-      "windDownDuration": "~2.5 hours",
-      "goalMissedBy": "~2h 15min",
-      "notes": "was tinkering with tools until late, then watched TV + internet in bed"
-    },
-    "2026-01-31": {
-      "stoppedWork": "19:56",
-      "startedBedRoutine": "22:05",
-      "actualBedtime": "23:27",
-      "windDownDuration": "~1.5 hours",
-      "goalMetBy": "33 minutes early!",
-      "notes": "nudge at 19:13 helped defer script task. Proper wind-down: TV, then bed with cat/tea/audiobook. Huge improvement from previous night.",
-      "sleepOutcome": {
-        "fellAsleepIn": "~30 min",
-        "briefWake": "5-6am",
-        "wokeUp": "08:00",
-        "totalSleep": "~8 hours",
-        "quality": "good",
-        "morningFeeling": "better than last 2 days"
-      }
-    },
-    "2026-02-02": {
-      "stoppedWork": "22:10",
-      "noseShower": "22:39",
-      "inBed": "23:51",
-      "tryingSleep": "00:05",
-      "windDownDuration": "~2 hours",
-      "goalMissedBy": "~5 min (work ended late at 22:10 instead of 22:00)",
-      "notes": "I failed at wind-down guidance - helped with work from 19:00 to 22:10 without suggesting stopping points. User called me out, updated HEARTBEAT.md. Bedtime itself was close to goal."
-    }
-  }
+    {"time": "19:14", "activity": "Tinkering — troubleshooting Bazzite suspend/resume + VRAM on GPD Win 4", "note": "Saturday evening, tech tinkering not deep work. Active conversation."},
+    {"time": "19:44", "activity": "Still GPD tinkering — fixed VRAM (8GB via UMAF), narrowed suspend crash to CPU hotplug (PowerTools core parking + SMT)", "note": "Making progress, having fun with it. Not work."},
+    {"time": "20:10", "activity": "Done tinkering — nosmt kernel param fixed suspend. Switching to YouTube then TV/audiobook", "note": "Good wind-down transition. Reminded about nose shower."}
+  ]
 }
Author	SHA1	Message	Date
Agent	66423cf66b	Remove redundant work-detection from HEARTBEAT.md (now in AGENTS.md)	2026-02-07 19:20:12 +00:00
Agent	ec18f30f6d	Add wind-down rules to AGENTS.md + first 19:00 nudge to HEARTBEAT.md	2026-02-07 19:18:34 +00:00
Agent	f1db2f3b2d	Add Jellyfin CLI + update wind-down instructions with concrete suggestions	2026-02-07 16:09:01 +00:00
Agent	4be80803d4	Pre-compaction memory flush: full day summary	2026-02-07 15:54:48 +00:00
Agent	cc5baa712b	Add audiobooks CLI for Audiobookshelf API access - current/recent/finished/stats/libraries commands - Tab-separated output for minimal tokens - For wind-down suggestions: know what user is listening to	2026-02-07 15:37:00 +00:00
Agent	10a4c1227b	Rename brain-dump to tasks	2026-02-07 15:15:32 +00:00
Agent	e26f7e141b	brain-dump: fix edit usage message, clean test data	2026-02-07 14:11:22 +00:00
Agent	516d6a1b34	brain-dump: merge --note into --context, single field for extra info	2026-02-07 14:09:40 +00:00
Agent	0352f8d04b	brain-dump: add recurring management, remove --when filter, show full context	2026-02-07 14:07:57 +00:00
Agent	3509d29882	Migrate brain-dump tasks to hex IDs	2026-02-07 14:01:52 +00:00