REPORT: website pre-Facebook refresh (audit table, deploy verification)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N1W4wBum4JSFrbaEoDkMBy
This commit is contained in:
2026-07-18 21:31:57 +02:00
parent d2fd935a71
commit 28811c207b
+142 -178
View File
@@ -2,203 +2,167 @@
> **Overwrite** this file with a summary of the most recent task only (uniform with the other repos; not cumulative). The cumulative hub history lives in [hub/CHANGELOG.md](hub/CHANGELOG.md); the scripts history lives in [scripts/CHANGELOG.md](scripts/CHANGELOG.md).
## Pre-travel train — R-39 heal · scripts v1.21.0 + ISOs · nav polish · (golden deferred) — 2026-07-18
## Website refresh before the Facebook tester-recruitment post — 2026-07-18
**Commits (this repo):** `bcdb042` scripts v1.21.0 · ROADMAP touch (R-39 diagnosis + R-33 collapse).
**Sibling commits:** `felhom-agent` `9596d5a` (v0.90.1) + `f22f70c` (report) · `felhom-controller`
`24d23b8` (v0.146.0) + `fd93020` (accordion tests).
**Phases 13 shipped. Phase 4 skipped cleanly (its own "time-permitting"). Phase 5 (golden 0.146.0
+ publish) NOT started — see the closing section.**
**Scope:** `website/` only · **Commits:** `be2fc50`, `d2fd935` (pushed to `main`, live via git-sync)
· **Source of truth for every claim:** `documentation/architecture/00-capability-map.md` (verified
2026-07-16, plus the 2026-07-18 N100 rehearsal rows).
---
## Phase 1 — R-39: diagnosis, and the brief's hypothesis refuted
### 1. Audit table — every factual claim, verdict, map row
**The conditional hub fix was NOT shipped, because its condition proved false.** The brief said to
ship a generation-bump fix "only if step 12 pin the mechanism to *re-mint fails to bump the
generation*". It does not:
Map references are section letters from the capability map. "—" = no map row exists, which is
itself the finding.
- `store.SetHostDesired` bumps `desired_generation` **unconditionally** — it went **2 → 3** on the
re-issue.
- `web/configs.go`'s `applyPBSDR` is **exonerated**: its "idempotent … no re-key, no second secret,
no spurious generation bump" comment at ~L618 is accurate and guarded by the
`cur != nil && cur.Namespace != ""` early return. The hub log shows mint #2 came from the
**re-issue** path, not from an Edit-tab Save. The comment-vs-behaviour contradiction the brief
expected does not exist.
#### index.html
**The real mechanism is a signal mismatch between two tiers.** The hub's re-consume signal is *a
generation bump + a poke*. The agent's re-apply trigger is *a change in the descriptor content hash*
(`felhom-agent internal/pbsdr/manager.go` ~L235):
| # | Claim (before) | Verdict | Map row / evidence |
|---|---|---|---|
| 1 | „45+ alkalmazás" | **FIX → 53** | §B *Deploy an app from the catalog* PROVEN-LIVE; real count = 53 `*.felhom.yml` in `app-catalog-felhom.eu/templates/` |
| 2 | „Felhő Felügyelő" (tech tile) | **FIX → „Vezérlőpult"** | Naming ruling; controller login page + R-11 draft both say vezérlőpult |
| 3 | „Kubernetes / Üzleti szintű rendelkezésre állás" tile | **CUT** | — (no customer-facing k8s row; also a bare availability promise) |
| 4 | og:image → `/assets/og-image.png` | **FIX (was a 404)** | File never existed; only `.svg` did. Generated 1200×630 |
| 5 | og:title / og:description (generic pitch) | **FIX** | Rewritten to one honest map-traceable sentence |
| 6 | Hero CTA „Érdekel →" | **FIX** | Retargeted to `#zart-teszt` (the post's landing intent) |
| 7 | Fényképek / dokumentumok / média a saját gépeden | **ADD** | §E Phone photo backup PROVEN-LIVE (6D); §E Documents/OCR PROVEN-LIVE (CAMPAIGN-6C 4P-1) |
| 8 | Hálózati megosztás — Windows Intéző + Mac Finder | **ADD** | §E *Files from Windows Explorer / Mac Finder (SMB server)* **PROVEN-LIVE**; Explorer leg passed 2026-07-18 |
| 9 | Kétszintű automatikus mentés (helyi + titkosított távoli) | **ADD** | §C Tier-2 PROVEN-LIVE (CAMPAIGN-6E P-TIER2); §C Offsite restic PROVEN-LIVE (6D); shares ride both tiers (R-7b, ctrl 0.145.0) |
| 10 | Helyreállítási kód, amit csak az ügyfél ismer | **ADD** | §A *Escrow ceremony … operator zero-knowledge* **PROVEN-LIVE**; wizard first live firing 2026-07-18 |
| 11 | Távoli felügyelet + automatikus frissítés | **ADD** | §G Controller self-update PROVEN-LIVE (demo); §G Publish train floor lift proven on a fresh install (5 s, unattended); §F metrics/monitoring IMPLEMENTED |
| 12 | Internetkimaradáskor a LAN-on minden működik | **ADD (plain wording)** | §E *LAN access when internet is down* **IMPLEMENTED** — allowed as a feature per the governing rule. **Caveat in §5.** |
| 13 | Visszaállítás, ha baj van | **ADD** | §C Offsite restore PROVEN-LIVE (6D); §C Guest-loss DR PBS restore PROVEN-LIVE (CAMPAIGN-2 T-P9) |
| 14 | „Zárt teszt" section + limitations + CTA | **ADD** | See §2 |
```go
if mk := m.loadMarker(); mk != nil && mk.Hash == h && (cf == nil || cf.Hash != h) {
return // idempotent: this exact descriptor already converged
}
```
#### technologiak.html
An ep0 credential re-issue re-keys the **secret of an existing token**, so `token_id` and
`fingerprint` never change and the descriptor stays **byte-identical** — only the side-table
`host_pbs_secrets` row rotates. Same hash → converged agent short-circuits → the fresh secret is
never consumed → the box keeps presenting a revoked credential → **401 forever**. Proof in one line:
`consumed-failed.json` carries hash `a4e5424…`, **identical** to the `marker.json` written two
minutes before the re-issue. The comment at `hub/internal/web/pbsdr.go:320` asserts the reissue
refreshes the descriptor "with the NEW token_id/fingerprint" — false for this op.
| # | Claim (before) | Verdict | Map row |
|---|---|---|---|
| 15 | „Felhő Felügyelő" ×14 (title, meta desc, keywords, og+twitter desc, JSON-LD, TOC, h2, h3, body, stack layer, hedge ×2) | **FIX → vezérlőpult** | Naming ruling |
| 16 | „45+ app telepíthető" | **FIX → 53** | §B |
| 17 | **Whole Kubernetes section** (h2, „Mi az a Kubernetes?", „Mikor érdemes", k3s vs full-k8s cards) + TOC entry + multi-node hardware tier | **CUT** | — No customer-facing k8s/k3s row exists. k3s is *hub* (operator) infra, not the customer box. Also carried „a szolgáltatás nem állhat le" = an availability promise the task forbids |
| 18 | „Tailscale VPN biztonságos távoli elérés" | **FIX → WireGuard** | §E *WireGuard base infra always-on* IMPLEMENTED; §E remote access = Cloudflare Tunnel + Traefik PROVEN-LIVE. Tailscale is not what ships |
| 19 | „RAID adatvédelem" card | **FIX → kétszintű mentés card** | — No RAID row. Shipped model is §C tier-2 second drive + encrypted offsite. Kept the „több lemez ≠ mentés" education |
| 20 | Spec row „4-8 TB (RAID)" | **FIX → „4-8 TB"** | — |
| 21 | „Melyiket válaszd?" referencing the removed multi-node fürt | **FIX** | Consistency with #17 |
| 22 | Hardware price ranges (80k600k Ft) | **KEEP** | Hardware cost guidance, not a service price or product claim; no map conflict. Flagged in §5 |
| 23 | Docker / Proxmox / reverse proxy / Traefik / UFW | **KEEP** | §E Remote access PROVEN-LIVE; Proxmox is the platform per `proxmox-platform.md` |
**Timeline (hub log is CEST; the hub DB is UTC — a split *within one service*):**
#### gyik.html
| CEST | Event |
| # | Claim (before) | Verdict | Map row |
|---|---|---|---|
| 24 | „Minden családtag saját fiókot kap, saját bejelentkezéssel" | **FIX (both JSON-LD + visible copies)** | §E *Multiple household users / per-person accounts* = **MISSING** (R-15). Now: apps do have per-person accounts; the **vezérlőpult has one shared household password** |
| 25 | „Nagyobb rendszereknél RAID konfiguráció **is alkalmazható**" | **KEEP** | Already conditional/on-request; not presented as a shipped product capability |
| 26 | „Jellyfin/Plex — saját Netflix" | **KEEP** | §B catalog deploy PROVEN-LIVE; browser/mobile playback is the app's own function. TV casting handled honestly in the tester block (#14) |
| 27 | Remote access, 3-2-1 backups, updates, power-on-after-outage | **KEEP** | §E, §C, §G rows all IMPLEMENTED or PROVEN-LIVE |
#### biztonsagimentes.html
| # | Claim (before) | Verdict | Map row |
|---|---|---|---|
| 28 | „…**azonnal** értesítést kapsz, és mi is" | **FIX** | §F *Health-degradation email* = **IMPLEMENTED**, and the map states this trigger "has never fired an email live in any doc". Softened, plus a line that the closed test is partly to harden this |
| 29 | Helyreállítási kód / zero-knowledge | **ADD** | §A escrow PROVEN-LIVE. States both sides: nobody else can read it, and without the code neither can we |
| 30 | „az egész rendszered … egy új gépre is visszaállítható" | **KEEP** | §C Guest-loss DR PROVEN-LIVE (CAMPAIGN-2 T-P9-DESTROY-RESTORE) |
| 31 | 3-2-1 rule, incremental, scheduled | **KEEP** | §C nightly dumps + tier-2 + offsite, all PROVEN-LIVE |
#### alkalmazasok.html / kapcsolat.html
| # | Claim (before) | Verdict | Map row |
|---|---|---|---|
| 32 | „45+" ×4 (title, meta desc, og:title, twitter:title) | **FIX → 53** | §B. Page renders 49 cards; catalog holds 53 templates |
| 33 | Contact form → `POST /api/contact` with a stale "TODO: never wired" comment | **KEEP — verified LIVE** | Probed this session: `200 {"message":"Üzeneted sikeresen elküldtük!…"}`. Backend is `manifests/contact-mailer.yaml`. No new backend built |
| 34 | Subject options | **ADD** | New „Zárt teszt — jelentkezem tesztelőnek" → mail subject `Jelentkezés — zárt teszt` |
---
### 2. The „Zárt teszt" section (the post's landing target)
Three cards + CTA:
- **Valódi használat** — the box stays with them, real data, we help with install and migration.
- **Visszajelzés** — what we want back; "I couldn't find this button" is the most valuable report.
- **Ismert korlátok** — stated plainly: **one shared household password** for the vezérlőpult
(§E multi-user MISSING, R-15); **TV-re streamelés hamarosan** (§E DLNA **MISSING**, R-6/R-8);
and "this is a test, expect to find bugs — that's the point".
**CTA wiring:** the audit found the existing contact-mailer live, so per the task the CTA reuses it
`/kapcsolat?tema=zart-teszt`, which preselects the new subject option (unknown `?tema=` values are
ignored with a console warning). **No mailto fallback needed, no new backend.**
---
### 3. Files changed
| File | Change |
|---|---|
| 18:30:51 | `pbsdr provisioned … gen 2; secret stored consume-once` — mint #1, via the WG-registration hook, **with** a generation bump |
| 18:45:51 | agent consumes mint #1`converged state=applied` |
| 18:47:52 | `pbsdr credentials **re-issued** … fresh consume-once secret stored` — mint #2, `consumed_at` stayed NULL |
| `website/index.html` | meta desc, og/twitter block (+width/height/alt), hero CTA, **new „Mit tud a doboz ma?" (8 cards)**, **new „Zárt teszt" section**, 45+→53, 2 tiles retitled |
| `website/technologiak.html` | 14× naming fix, 45+→53, **Kubernetes section + TOC + multi-node tier CUT**, Tailscale→WireGuard, RAID card→two-tier backup, spec row |
| `website/gyik.html` | multi-user answer rewritten in **both** the JSON-LD and the visible accordion |
| `website/biztonsagimentes.html` | new recovery-code block; notification overclaim softened |
| `website/alkalmazasok.html` | 45+→53 ×4 (head only) |
| `website/kapcsolat.html` | new subject option + label mapping + `?tema=` preselect |
| `website/assets/og-image.png` | **NEW** — branded 1200×630 card, generated from the existing logo + v2 tokens via headless Chrome (real brand font, no external assets) |
| `website/CHANGELOG.md` | entry on top |
**A second, independent defect, found while healing.** `configs/felhom-pbs-apply`'s `reconcile`
passed `--server` to `pvesm set`; PVE treats `server` as **create-only** and rejects the whole call
even when the value is byte-identical. So *every* re-apply exited 255 — and because the agent
consumes the one-time secret **before** invoking the wrapper, each re-issue **burned a credential**.
Proven live before writing code: with `--server` → rejected; without → **rc 0**. Fixed in **agent
v0.90.1** (one argv line + red-proof `TestReconcileNeverPassesServerToPvesmSet`, verified red then
green; it handles two vacuous-pass traps — CRLF line endings, and the WHY comment quoting the very
flag under test).
**Cost I incurred:** proving the mechanism consumed the pending secret against the still-unfixed
wrapper, so it burned. The box was already 401 before and after — no functional regression — but the
recoverable state was gone until an operator re-issue. The agent parked correctly in
`consumed-failed.json` with `NOT retrying silently`: **no burn loop**, the fail-safe worked.
**HEALED — Viktor's re-issue click closed the chain in 9 s:** hub re-issued 20:28:44 → agent
consumed 20:28:51 → `converged state=applied` 20:28:53, with the patched wrapper.
| Check | Before | After |
|---|---|---|
| `pvesm status` | `401 Unauthorized` / `inactive` | **`active`** |
| Direct token probe `/api2/json/version` | `401` | **`200`** |
| Real backup | none possible | **`felhom-pbs:backup/ct/9201/2026-07-18T18:31:06Z`, 9 744 319 312 B, 13m36s** |
Encrypted under fingerprint `7e:a6:af:f7:ea:6d:3e:d9` — the **escrowed** key, the one customer zero
holds the recovery code for. The DR tier's **first real backup on the reborn box**. Nothing was
destroyed: `.pw`, `.enc` (K) and the `storage.cfg` entry verified intact (PVE rejects atomically, so
the set-only law held).
**Left for the fleet spec, deliberately not improvised:** (a) make a fresh unconsumed secret actually
un-converge the agent; (b) fix the verify loop's read path — it reads `/etc/pve/priv/storage/<id>.pw`
directly as non-root, a file it can only ever *write* through the root wrapper (`/etc/pve/priv` is
`0700 root:www-data`; sudoers exposes `create|reconcile|grant`, **no read verb**); (c) an auth probe
so `applied` can never mean `401`.
**No CSS, nav, footer, or design changes.** All new markup reuses existing classes
(`.why-grid`, `.why-card`, `.apps-cta`, `.cta-button`) and the existing icon sprite.
---
## Phase 2 — scripts v1.21.0 + fresh ISOs
### 4. Deploy verification (fetched from the live site)
`run_pairing()` now loops **inside** the script (30s sleep — hub-side rate unchanged) instead of
exiting non-zero per poll, so the unit sits in `activating` and systemd prints nothing on the
customer's console. Registration split into `register_appliance()` whose transient failures the loop
retries. Journal quiet but not dark: logged once on entry, then a 10-minute heartbeat; `410` still
exits non-zero on purpose. Console banner every 5 min, single accented spelling, plus the missing
reassurance („Ez a képernyő magától frissül").
- Commits `be2fc50` + `d2fd935` pushed to `main`; git-sync deployed in ~20 s and ~45 s.
- **og tags live on `/`** — `og:title`, `og:description` (honest one-liner), `og:image`,
`og:image:width/height/alt`, `og:locale hu_HU`, `og:site_name`, `twitter:card
summary_large_image`, `twitter:title/description/image`. All verified by `curl`.
- **`og:image` no longer 404s** — `HTTP=200 content-type=image/png bytes=167598`, PNG signature
valid, **dimensions 1200×630** confirmed from the live bytes.
- **Site-wide sweep across all 6 public pages:** `Felhő Felügyelő = 0`, `k3s|kubernetes = 0`,
`45+ = 0`, og-image `200`.
- New sections live: `id="mit-tud"`, `id="zart-teszt"`, `/kapcsolat?tema=zart-teszt`,
`<strong>53</strong>`, „Hálózati megosztás", „TV-re streamelés hamarosan".
- `/kapcsolat` serves `value="zart-teszt">Zárt teszt — jelentkezem tesztelőnek<`.
- **Gates:** `python scripts/site_gates.py` green (BOM, emoji=0, nav/footer identical, analytics,
no CDN, no legacy tokens, no `<style>`, cache-busted assets).
- **Mobile pass — measured, not eyeballed.** At a 380 px viewport:
`documentElement.scrollWidth == clientWidth == 365`**no horizontal scroll**. Cards 317 px,
CTA **275×58** (well above the 44 px tap target), hero h1 40 px, body 16 px, single column.
*Note:* headless screenshots at `--window-size=380` looked clipped, but that is a headless
rendering artifact — the DOM measurement (and the identical artifact on the pre-edit baseline)
disproves it. The only elements extending past 380 px are the off-canvas `.nav-links` menu, by
design.
**The load-bearing half is `TimeoutStartSec=infinity`**a `Type=oneshot` ExecStart is killed at 90s,
so without it systemd would kill the new wait and `Restart=on-failure` would silently reinstate the
exact spam this removes, *after appearing to work for the first three polls*.
**Verified behaviourally**, in a container against a stub hub answering `204` five times then
delivering: **one log line plus one heartbeat, zero exits between polls**, then a clean fall-through
to the direct install and `exit 0`. The old design produced 5 unit invocations and 5 `Failed to
start` console lines for that same sequence.
**ISOs rebuilt (both `--pairing`, `--loader mkimage`, same PVE input `proxmox-ve_9.2-1.iso`
sha `4e88fe41…`), `secret-bearing: no`:**
| ISO | sha256 |
|---|---|
| `felhom-pve-9.2-1-v1.21.0-n100-generic-mkimage.iso` (safety) | `b1b25fd412b779bcacbfaa4c59002ee80a8f35e4c0bd1248cc35d997f967cb90` |
| `felhom-pve-9.2-1-v1.21.0-n100-demo-generic-mkimage.iso` (real) | `90a0fb7da3f3f11d315bf1cdf55a7da84ab9e6d75e06b2942041470f7864031d` |
Both at `180:/mnt/5_hdd/felhom.eu/felhom-iso/out/`. **Verified the fix actually shipped inside the
artifact**, not just in git: extracted the embedded first-boot payload with xorriso and decoded it —
the shipped `felhom-bootstrap.sh` is **byte-identical to the committed source**, carries the new
cadence constants and the `while true` loop, the old `will poll again in 30s` exit line is **gone**,
and the embedded unit carries `TimeoutStartSec=infinity`. Viktor flashes the stick.
> **Zero „Felhő Felügyelő" occurrences remain** — confirmed both in the working tree (all 7 HTML
> files, including the non-public draft) and on the live site (all 6 public pages).
---
## Phase 3 — controller v0.146.0 nav polish
### 5. Things that need a ruling (not fixed)
Built, pushed and **deployed to guest 9201** (`0.146.0 Up (healthy)`).
- **Scrollbars:** thin + hairline-coloured; `scrollbar-width`/`scrollbar-color` for Firefox **and**
`::-webkit-scrollbar` (8px, thumb `--line`, hover `--text-3`, `--radius`) for WebKit/Blink, since
neither alone covers the browsers customers use. `.sidebar``--bg-2` track, `html``--bg-0`.
Tokens only.
- **Collapsible groups:** Tárhely / Biztonsági mentés / Megosztás as accordions, chevron, exactly one
open. Header is a **real `<button>`** with `aria-expanded` + `aria-controls` + `:focus-visible`, so
keyboard/AT reachability is real rather than simulated. **Nothing became unreachable** — checked
first: every group's landing page is also its first sub-item. **Progressive enhancement** — the
active group is opened **server-side**, so it is correct before any JS runs. **No layout jump**
`grid-template-rows: 0fr → 1fr` rather than `max-height`, animating to the content's real height
with no magic number to drift; the toggle reserves its active border as transparent; both
transitions off under `prefers-reduced-motion`.
All design-v2 gates PASS (`template_id_gate`, `emoji_gate`, `native_confirm_gate`,
`offbox_rename_gate`, `mojibake_gate`, `app_row_dedup_gate`); build/vet/tests green.
`docker_run_volume_path_gate` still fails on `estimate.go:179` — that is **R-29**, pre-existing,
verified to fail identically on the untouched tree, and deliberately not bundled.
**Screenshot leg NOT done, and here is the honest reason.** The demo controller's password is
**customer-owned** since the claim flow — Viktor set it during the rehearsal — so the credentials on
the build server are stale and a curl-login returns the `Bejelentkezés` page. Instead of asserting
nothing, four **render tests** (`internal/web/nav_accordion_test.go`) pin the server-side half through
the real shared layout: every sub-page opens its own group with `aria-expanded=true` and an `.active`
toggle and **exactly one** group open (the count is asserted, not just the expected group); a page
outside any group opens nothing; every group's landing page still exists as a sub-link; the toggle is
a real button whose `aria-controls` targets a real element. **Red-proofed** — removing the `is-open`
marker fails two assertions on both storage pages. The visual leg still wants Viktor's browser.
---
## Phase 4 — skipped cleanly
Its own instruction was "time-permitting; skip cleanly if not". Nothing was started, so nothing is
half-done. Note that its (a) auto-mint self-bind link, (b) post-RESET health card and (c)
unprovisioned-offsite flash correspond to **R-36 / R-37 / R-36** and remain open as written. The
conditional Phase-1 hub fix is **not** part of any v0.67.0 train, because its condition was refuted.
## Phase 5 — golden 0.146.0 baked + published (STOP → Viktor)
Run on the nested drill VM on 180, following the recorded Phase-C procedure
(`pilot/RUNBOOK-publish-0.85-0.120-2026-07-12`).
**Artifact:** `felhom-golden/0.146.0/golden.tar.zst`
| | |
|---|---|
| **GOLDEN_SHA256** | **`4834c703162c5437467a329144b1a523019bf5693ab9d439558be7323587e955`** |
| Size | 612 696 588 B (archive 584 MB) |
| Controller baked in | **0.146.0** (confirmed in the log, not assumed) |
**Every pass marker green:** unit `Result=success` / `ExecMainStatus=0` · **0** FATAL or exclusions ·
`docker OK (overlay2; data-root /var/lib/docker)` · **all three mounts included** — rootfs `/`,
mp0 `/var/lib/docker`, mp1 `/mnt/sys_drive` · pre-delete **HTTP 404** (the pre-gate — the version did
not previously exist) · upload **HTTP 201**.
**Integrity verified independently of the build host** — anonymous `GET | sha256sum` matches
`4834c703…e955` **byte-for-byte**, ranged GET returns **206**, and `content-length` matches the bytes
the bake reported. The golden now appears in the hub's dropdown alongside 0.136.0 and 0.143.0.
**Teardown per GL-1 discipline:** bake log copied out first as evidence
(`180:/mnt/5_hdd/felhom.eu/drill/bake-0.146.0.log`), then build guest 9100 `pct destroy --purge`
(0 guests remain), token + script + log `shred -u` in-VM, VM powered off, qemu confirmed gone (via
`ps -eo comm`, not the self-matching `pgrep -f`), drill disk reverted to the `virgin` snapshot
**exactly as found**. **Token-leak grep = 0** against the literal token value, on the bake log and on
both ISO build logs from this session.
> ### STOP — the two remaining saves are yours (password-gated)
>
> 1. **Day-0 manifest** → Golden **0.146.0** (sha auto-reads as `4834c703…e955`).
> **Agent stays 0.90.0, MinAgent stays 0.90.0** — the v0.146.0 CHANGELOG declares no new agent
> coupling, so nothing justifies moving either. Save.
> 2. **Then** the floor → **v0.146.0**, saved **LAST**.
>
> The traveling box converges over the tunnel — that would be the **second** live floor lift, this
> time from the road. Worth noting in CONTEXT when it lands.
1. **Naming ruling was applied as the default, per the task** — the surface is „vezérlőpult",
the brand is Felhom. One line vetoes it.
2. **„Internetkimaradáskor is működik" is the softest claim on the new page.** §E rates it
**IMPLEMENTED**, and the map explicitly notes it was *"never drilled as a customer experience"*
(R-19). It is allowed as a feature by the governing rule, but it is the one card on the landing
page with no live customer-facing evidence behind it. If R-19 stays undrilled before the post
goes out, consider cutting that card — it is one `<div>`.
3. **Hardware price ranges (80k600k Ft) on `/technologiak` were left in place.** The task said not
to *add* pricing; these predate it and are hardware cost guidance, not a service price. But a
recruit arriving from "volunteer tester" messaging then hitting price tables is a tonal
mismatch — worth a ruling.
4. **Foreign WIP in the shared worktree — I did not touch it.** `website/assets/logo.svg`,
`og-image.svg`, `og-image_wide.svg` were already modified and `og-image_wide.png` untracked when
I started. Staged only my own paths (per the never-`git add -A` rule); those four remain
uncommitted. **Someone appears to be working on OG imagery in parallel — reconcile before it
collides with the `og-image.png` this task shipped.**
5. **The R-11 one-pager does not exist yet** (ROADMAP status: *idea*). The task referenced its
„Egy mondat a bizalomról" wording; since there was nothing to reuse, the trust copy was sourced
from the two places that *are* ruled: the shipped escrow wizard template (`backups_escrow.html`)
and `RUNBOOK-onboarding-draft-v3.md` §C2. If the one-pager later lands with different wording,
align the site to it.
6. **I sent one real test message through the live contact form** (`name=Teszt`, `message=probe`)
to confirm the mailer was actually wired before building the CTA on top of it. It will be in the
info@felhom.eu inbox — safe to delete.
7. **`gyik.html` stores every answer twice** (JSON-LD `FAQPage` + visible accordion) and the two had
already drifted (the payback-cost answer differs between copies). The multi-user fix was kept in
sync across both; the pre-existing drift elsewhere is unfixed and is a standing trap.