Part 1: the weekend pick-list — "I want to do something useful but I don't want to work." Sorted by Impact × Leverage × Feel-like, every quest with a time estimate that assumes Claude Code does the heavy lifting. Feel-like scores are my guesses from your behavior — comment a number on any card to correct it (your corrections get written into the Pinecone records). Part 2: the ideas vault headed for your second brain.
⚡ 30-second read
How to use: on a free weekend, open this doc, take the top unfinished quest that fits your available hours, paste its prompt, go. Top 5 quests carry full prompt chains; the rest have kickoff one-liners. Then skim the Ideas Vault below and comment DUMP / KEEP / MODIFY: … on each — approved ideas get upserted to the selfi-journal index in a new ideas namespace so the coach surfaces them in context.
this doc exists so both buttons can be the right button
Feed the "Ignore" dumps (4,100 lines) into coach-intake
3
4
3
36
2–3 h
10
Personal-intro REC ritual (10× ENG + 10× HUN)
3
4
3
36
2 h setup
11
Email volume calculator — host the ready-made app
2
3
3
18
~1 h
12
PARA / notes reorg with AI
2
3
2
12
~4 h
*Feel-like guessed from what you build unprompted (the coach, Studio, animations, starred post-its). XP = Impact × Leverage × Feel-like. Quest 6 is time-boxed by the visa clock — do it first even though it's listed as doc 6.
Quest 1 · ⭐ The booking-niche campaign your starred idea
Why it's leverage: your own niche hypothesis — apartments/campings with websites but no booking system — has a built-in, provable money leak ("you're paying Booking.com 15% for guests your own site could take"). It's the rare cold niche where the pain is visible from the outside, which is exactly what your English-campaign post-it needs (pick niche → Apify → model copy → .com site → test launch). And a booking-system offer is a natural GHL Grand Slam Offer — this quest test-drives the pivot.
# Chain 1/3 — niche list build
Build a scraping plan + run it: find apartments, guesthouses and campings
(start: HU + HR coast + AT lake regions; later EN market) that have their
own website but NO booking engine (no availability calendar, no
book-now flow — just phone/email/contact form). Use Apify or a
Playwright crawler. For each: name, site URL, evidence screenshot-note of
the missing booking flow, contact email, guessed size. Target: 300 rows,
scored. Output: CSV + a summary of the 3 most common site patterns.
# Chain 2/3 — offer + lead magnet
From the CSV patterns, draft the offer: "own-site booking system +
deposit payments" one-pager, and a lead magnet per my post-it: a
personalized "how much Booking.com commission you paid last year"
estimator or a 2-min site-audit video script. Model the email copy on my
best-performing Instantly emails (harvest them first if Quest 8 not done).
# Chain 3/3 — test launch
Set up the .com sender domain + redirect landing page, load 100 leads,
launch a 2-variant test through my existing Instantly infra. Report
open/reply rates after 5 send-days.
Quest 2 · Studio comment-triage automation attacks THE bottleneck
Why it's leverage: your actual delivery bottleneck (the Mirror doc, Q5) is processing client comments, not selling — 70 of your last 100 inbox mails were Studio comment notifications. Automating triage multiplies every future website simultaneously, and it's a feature of a product you own (Studio) — so it compounds into the case-study story too.
# Claude Code — Studio comment-triage v1
In the Clientsflow Studio repo: build a comment-triage layer. Pull all
open client comments across projects (EGYSE, Acélműhely, Ablakdoki, ZZ-*,
"L"). For each: classify = {copy-edit | asset-swap | design-change |
question | bug | out-of-scope}, extract the concrete instruction, group
by page/section, and generate (a) a per-project TODO markdown ranked by
effort, (b) draft replies for the "question" class, (c) a mechanical-edit
batch I can review as one diff (copy edits, phone numbers, image swaps).
Include EGYSE's known review-tool bugs (comment edit/delete, resize) as
tickets. Do not auto-apply anything — everything lands as a reviewable
batch. Also add the digest as a daily email to me instead of 70 single
notifications.
Quest 3 · Sales-call REC → AI feedback coach
Why it's leverage: you already record calls and already have transcript infra (the coach). Every sales call becomes a training rep with a critique — quoted lines, "say this instead" — compounding your close rate, which multiplies every future lead from every campaign. Cheapest skill-gain per hour available to you right now.
# Claude Code — sales-call feedback coach v1
Build a small pipeline: I drop a sales-call recording or transcript into
a folder → it produces a structured critique: (1) call timeline, (2) my
3 best moments (quoted), (3) my 3 costliest moments (quoted, with the
exact alternative line to say next time), (4) discovery-question quality
vs. the SPIN/Hormozi frames, (5) one drill for the next call. Store
critiques in a running log so month-over-month patterns are visible.
Use my existing transcription setup from the telegram-coach project if
reusable. Test it on the most recent recorded sales call you can find.
Why it's leverage: your profit table names reputation ("pro website w/ case studies") as a lever, and every cold email you send links to it — a stunning site raises reply→call conversion on all campaigns at once. Also: you'll actually do it, which beats three better-on-paper quests you won't.
# Chain 1/2 — inspiration + spec
Browse 21.dev, godly.website and awwwards for agency-site patterns heavy
on scroll animations. Collect 8 concrete interaction patterns (hero,
case-study reveal, marquee logos, sticky process section...) as short
specs with reference links. Then propose the CF site structure: hero,
results/case-study strip (use Studio client engagement + campaign
numbers as placeholders), services as the two-package offer, footer CTA
= booked call. Output a build spec I approve before code.
# Chain 2/2 — build
Build it (stack: whatever the current CF site uses, or Astro if
greenfield), with the approved animations, mobile-first, 90+ Lighthouse.
Screenshot-QA at 1280/768/390 with Playwright and iterate until clean.
Deploy to staging; I review before DNS switch.
Quest 5 · Voice-coach upgrade pack your post-it, verbatim
Why it's leverage: the coach only pays compound interest when used daily — and the Mirror doc shows the 3-week silence cost. Your post-it already specced the fix: make it more alive (names), safer to change (sims), and binding (pact + visible tracker).
# Claude Code — voice-coach upgrade pack (telegram-coach repo)
Three upgrades, agent-files first (no redeploy where possible):
1. NAMES: give the coach's subsystems personas/names (cornerman, the
morning caller, the planner...) — update AGENT.md + skills so calls
reference them naturally. Propose names for my approval first.
2. SIMS: run a coach-sim-e2e campaign (dry-run first) covering: missed-
week re-entry, substance-rule tracking (per the new Goals-doc rule),
and the full evening protocol incl. persuasion-when-I-resist — the
exact failure I flagged on Jul 12. Report scores + worst transcripts.
3. PACT + TRACKER: implement the usage pact — a visible streak tracker
(calls answered / week) the coach reads out at weekly planning, plus
the escalation ladder from Mirror Q4 once I've answered it.
Quests 6–12 · kickoff one-liners
6 · Destination sprint: open the Destination doc, run the 7 Perplexity prompts, comment your pick. (Do this one THIS week — visa clock.)
7 · Hormozi planner → skill:"Take the 'kampány tervező' mega-prompt from my Notion To Schedule page and turn it into a Claude Code skill: input = prospect URL + niche, output = the per-campaign table (pain, TAM from KSH, CLV, profit/lead, CPL) + Lead Magnet Guide. Test on Bauart."
8 · Instantly copy harvest:"Log into Instantly (creds in credentials.md), export campaign stats + email variants across all accounts, rank by reply rate, and write the winners into my Notion email-writing template before old campaign data degrades."
9 · Feed the dumps:"Run the coach-intake enrichment spec over the '# Ignore — later feed all to an AI' section of my Notion Personal projects page + the Apádék long-form block from To Schedule; upsert as pending records."
10 · Intro REC ritual:"Draft my 60-second personal intro (agency owner + AI builder), 3 variants ENG + HUN, plus a 30-year-old-self version; set up a coach directive to drill it on evening calls for 20 days."
11 · Email calculator:"Take the email-volume-calculator HTML from my To Schedule page, host it on a subdomain, brand it CF, and add it to my sales-call toolkit."
12 · PARA reorg:"Inventory my Notion + Apple Notes top level, propose a PARA structure with a migration script plan; migrate one area as a pilot."
Part 2 · 💎 The Ideas Vault → Pinecone
📌 Your job on this section (5 minutes)
Comment on each idea below: DUMP (not worth the brain), KEEP (store as-is), or MODIFY: your correction. After your pass, I upsert the keepers — with your edits and feel-like corrections applied — into the selfi-journal index under a new ideas namespace, so the coach and your agents surface them exactly when a conversation touches them. Nothing gets stored before your pass.
Highlighted — the high-leverage eight
I1 · Two-package sales offer
Basic = qualified-lead handoff; Premium = calendar-booked meetings at a much higher price. Arrive at every sales call with the 2-option proposal pre-built. Leverage: price anchoring on every deal, zero extra delivery cost — the fastest revenue-per-client raise available.
I2 · Auto-proposal 10 minutes after the call + 72h window
Proposal generated and sent while the prospect is still warm, with a 10% incentive expiring in 72h. Leverage: attacks the #1 deal-killer (your own follow-up latency — see Gill&Murry's "1 day" promise) with automation instead of discipline.
I3 · The Hormozi campaign-planner mega-prompt
Per-service value-prop reformulation + per-campaign table (pain, TAM from KSH data, CLV, profit/lead, suggested CPL) + Lead Magnet Guide. Leverage: arguably the most valuable single artifact in your Notion — a reusable consulting asset that makes every prospect meeting look over-prepared. (= Quest 7.)
I4 · Value-based pricing reframe
Your pasted note: "$800 for $17,000 annual savings = 4.7%… stop selling the 15-minute setup, start selling the $17,000 savings." Leverage: a sentence that changes every future proposal's framing; pairs with I1.
I5 · List-reactivation service
Upsell after ~10 delivered leads: personalized outreach to the client's own dormant Gmail/CRM list; LinkedIn add-on. Leverage: new revenue line with zero cold-start — the audience already exists and trusts the client.
I6 · n8n copywriter flow
Sales-call + onboarding transcripts → persona, pains, dream outcome → $100M-offer-based copy. Leverage: turns every client conversation you already have into copy inputs automatically.
I7 · AI-assisted coaching business (the life-OS)
"Psychology-based custom software life-OS" per client: free evaluation call → AI-generated profile + flowchart as lead magnet → ~$2,000 setup + $1,000/mo. Leverage: the productization path for everything the telegram-coach already proves on you — the 5-year rung's seed.
I8 · "Sell internally as consultants" culture
Everyone (starting: Dani) pitches suggestions as if an external contractor selling the idea. Leverage: builds Dani's sales muscle for the 80%-delivery future while improving decisions; costs nothing to try for two weeks.
The rest of the vault — one line each
I9 · English cold-email campaign: niche → Apify → modeled copy → .com + redirect → test (post-it; Quest 1 is its first instance).
I10 · Own "Raw Leads" page/pipeline + buy & warm new email accounts (post-it).