Files

T

admin 9e079c7883 RECON: a Felhom-issued subdomain works in the product — the blocker is Cloudflare edge-cert depth

Question A: YES, no code change. customer.domain is a trimmed string with no
UNIQUE, no CHECK, no format rule (store.go:114, configs.go:673), copied verbatim
into controller.yaml (configgen.go:48), and every one of its 30 consumers on the
box interpolates it without parsing. Zero hits for registrable/eTLD/publicsuffix
across both repos. Nothing creates DNS records (zero hits for dns_records) — the
two Cloudflare clients are WAF-only. And the zone-ownership assumption is a
SWITCH, not a requirement: traefik.yml.tmpl selects DNS-01 when cf_api_token is
set and HTTP-01 when it is empty.

The real blocker is Cloudflare, proven live: the edge certificate covers exactly
one wildcard level (SAN = demo-felhom.eu, *.demo-felhom.eu), so a two-label
hostname — which a per-tester subdomain forces — gets "tls alert handshake
failure" and no peer certificate at all. That makes Advanced Certificate Manager
a prerequisite of the separate-domain plan, not an optional extra. Whether ACM is
available on the account could not be established read-only: the only Cloudflare
tokens in reach are the Zone:DNS:Edit tokens on the demo boxes, which the fence
forbids using.

Question C, measured rather than reasoned: r.Cookie returns the FIRST match and
never tries the others (BOGUS+real = 302, real+BOGUS = 200), so a tossed cookie
wins outright — DoS and confusion, not takeover, since it fails closed on
mutations. CSRF is a single choke point (server.go:256) and the token carries the
whole load against a same-registrable-domain attacker. But it is SKIPPED entirely
when no session cookie is present, which with browser-cached Basic auth is
cross-origin CSRF on every mutating route (R-135).

Agreeing with the separate-domain recommendation, with the caveat the brief asked
for: it is necessary but not sufficient. It does not solve Question D, because
that is a shared-zone problem and the new domain is a shared zone.

Filed R-133..R-138: duplicate domains accepted; hub/controller zone-resolvers
disagree on depth; CSRF skipped on the no-cookie path; __Host- rename (one line,
preconditions verified met); geo-WAF rules zone-scoped and non-namespaced (four
cross-tenant faults, blocks shared-zone onboarding); shared-zone cf_api_token is
a zone-wide DNS-write capability on a customer's box.

Nothing created: no customer, DNS record, tunnel, route or code change.

2026-07-31 09:00:11 +02:00

DIAGNOSIS-f9-storage-registration-gap-2026-06-14.md

…

FIX-M18-NOTES.md

…

FIX-M19-NOTES.md

…

FOLLOWUP-golden-default-controller-tag.md

…

FOLLOWUP-nas-automount-guest-reboot-reassert.md

…

OPEN-ITEMS.md

RECON: a Felhom-issued subdomain works in the product — the blocker is Cloudflare edge-cert depth

2026-07-31 09:00:11 +02:00

README.md

docs: file R-29 to the register; attach the gate-orphan instance to its class

2026-07-29 11:13:23 +02:00

ROADMAP.md

hub v0.84.0 — break-glass console credential on the host page

2026-07-31 08:19:36 +02:00

SPEC-r85-phase4-5-2026-07-26.md

R-85: demo-felhom restore-test PASSED — rotation proven on BOTH boxes; correct a bad estimate

2026-07-27 08:17:00 +02:00

README.md

documentation/backlog/

OPEN-ITEMS.md is the register of open work and the file to read first — it holds only what is open, one row per item, every row with a state and an owner. ROADMAP.md is the full history and reasoning behind the R-n IDs, including shipped and killed items; an ID is minted there, and a new instance of an existing item attaches to that ID rather than getting its own.

The rest of this folder: verified-LIVE findings with implementable fix plans that are not yet implemented. Preserved here (instead of on git branches) per the trunk-based, no-branches rule — the fix itself is implemented later directly on main, during a normal/supervised session.

FIX-M18-NOTES.md — dump re-validation runs every 5 min (perf). FIXED in controller v0.62.0 @ f8afe5c (2026-06-14). (was on the deleted felhom-controller branch fix/m18-dump-validation-cache.)
FIX-M19-NOTES.md — deriveStackName misattribution edge (low-incidence correctness). FIXED in controller v0.62.0 @ 6bab68b (2026-06-14). (was on the deleted branch fix/m19-stackname-crossref.)
FOLLOWUP-golden-default-controller-tag.md — the golden bakes a stale controller (:0.43.0 when queued; had rotted again to :0.85.1 by resolution). FIXED in felhom-agent @ ceca355 (2026-07-03): build-golden.sh v2.0.0 makes the controller tag a MANDATORY argument (a required arg cannot rot) and golden 0.98.3 was baked + clean-room-validated (bake → first-boot-current → self-manage → app deploy, on the drill VM — no supervised touch of live guests needed) + published + vouched. Evidence: ../audits/DRILL-golden-098-2026-07-03.md.

Related: the live-drive fixspec (../audits/live-drive-fixspec-2026-06-14.md) carries the deferred supervised items F9 (HDD provisioning/guest-attach), F20-BUG2 (durable_id scheme), F20-BUG3 (async mkfs) — to be implemented in the agent/golden supervised session.