Commit Graph

518 Commits

Author SHA1 Message Date
admin 8e69e5edda docs: scripts CHANGELOG - manifest_bearer_gate.py entry (incl. the .gitignore *secret* un-tracking gotcha) 2026-07-13 15:11:46 +02:00
admin 3616dcbb7c manifest: hub 0.53.0 (closing bundle GitOps bump; rides with the report-api secretKeyRef - Secret created out-of-band 2026-07-13 before this sync) 2026-07-13 15:05:06 +02:00
admin 3efb1b1af1 docs: hub v0.53.0 closing bundle - CHANGELOG + CONTEXT (both operator rulings verbatim) + take-two F-14 disposition (SHIPPED, gated on escrow-acked deletion provenance) 2026-07-13 15:02:44 +02:00
admin 3682303131 scripts: manifest bearer gate lands as manifest_bearer_gate.py (the .gitignore *secret* pattern silently un-tracks any 'secret'-named gate file - the previous commit's gate never made it into git); references updated 2026-07-13 14:59:31 +02:00
admin 80aca3a3a1 hub + manifests: operator bearer out of git (secretKeyRef report-api)
The global bearer (api.report_api_key) leaves manifests/hub.yaml: the
ConfigMap field is an empty placeholder; the Deployment injects
REPORT_API_KEY from out-of-band Secret/report-api (deliberately NOT
optional - a missing Secret fails Ready rather than booting a hub with an
empty bearer). main.go gains the REPORT_API_KEY env override (the
RESEND_API_KEY twin). scripts/manifest_secret_gate.py blocks bearer-shaped
(64-hex) literals across manifests/ (felhom.secret.yaml's pre-existing
umami APP_SECRET is a visible non-fatal KNOWN-BACKLOG line - secrets.md
backlog, out of scope here).

Ordering (load-bearing): create Secret/report-api with the CURRENT value
BEFORE syncing; this manifest change rides the same sync as the 0.53.0
image (older images ignore the env override). The exposed git-history copy
dies only with the supervised ROTATION - procedure + consumer list in
documentation/runbooks/secrets.md ("Operator/global bearer key"); both
publish runbooks' ROTATION notes now point at that disposition. The
controller repo's example-config copy of the literal is scrubbed
(felhom-controller 51c871a).

Red-proof E: reintroducing a 64-hex literal into manifests/hub.yaml ->
gate exit 1 (recorded); restored clean.
2026-07-13 14:57:09 +02:00
admin 36c72138f1 hub: dead-host roll-up honesty - customer status folds worst expected host
Customer status (dashboard row, /configs list, detail header + strip) is
now worst(controllerDerived, hostStatusOf(each expected host)) via the ONE
staleness definition (Server.hostStatus, hosts.go - shared with the
HostStalenessChecker; no second threshold). Any host down/stale caps the
customer at WARN with a cause chip naming the host ("host down: <id>");
pending (never-reported) hosts worsen only once the customer has reported
(onboarding exclusion). The three previously-inlined controller-status
chains collapse into controllerStatus() (rollup.go). Display + derivation
only - checker alerting untouched.

Live shape pinned (drill-1 / Peti cluster): host down 23h + controller
report minutes old rendered a GREEN row - TestRollup_DeadHostMasking now
fails that exact outcome. Red-proof: short-circuiting foldHostStatus to
controller-only flips C + two D subtests red ("dashboard row is GREEN
over a 23h-dead host").
2026-07-13 14:52:04 +02:00
admin 04861a7ed3 hub: F-14 part 2 — gated auto-Reissue on the PBS-DR enable dead-end
pbsdrProvisionAtom, on tenantsync token_exists: consult the customer's
MOST RECENT host-deletion record. escrow_acked=true -> invoke the
EXISTING tenantsync Reissue op, store the audit event (pbsdr_auto_reissue,
hub-source, "Previous key destroyed (acknowledged deletion) - credentials
re-issued automatically."), proceed to secret+descriptor as a normal
provision. No record / un-acked -> the pre-existing refusal, byte-unchanged
(never-silently-re-key law; operator ruling 2026-07-13).
Scenario A test drives the REAL handler path over a REAL escrow-ack delete;
scenario B pins the exact non-effect (zero Reissue calls, no state).
Red-proof: bypassing the gate flips both B subtests to 303 (silent re-key).
2026-07-13 14:44:42 +02:00
admin 2321077800 hub: F-14 part 1 — host-deletion provenance (host_deletions, in-tx)
Every DeleteHost writes a provenance row INSIDE the cascade tx: host_id,
customer_id, deleted_at, escrow_acked. escrow_acked = ack given over a
PRESENT escrow row (acknowledged destruction, not a vacuous checkbox).
LatestHostDeletion(customer) serves the F-14 gate — newest record only,
so an old acked deletion never whitelists a newer un-acked one. No
backfill by design: pre-record deletions stay on the manual path.
Red-proof: dropping the in-tx INSERT fails TestDeleteHost_ProvenanceRecord
("no deletion record written") + the part-2 scenario-A test.
2026-07-13 14:44:42 +02:00
admin 74fa61c7bc audits: disposition — backups-IA finding 2 (legacy-boot SystemDisks dead wizard) RESOLVED by felhom-agent v0.87.0 (slaves-walk, operator ruling 2026-07-13); live evidence drill replay + felhom-pve byte-identical regression; remaining: supervised wizard leg + Day-0 manifest vouch 2026-07-13 13:21:14 +02:00
admin a92e18b21f docs: hub v0.52.0 polish batch — CHANGELOG + REPORT + CONTEXT + REUSE (inline_confirm partial); take-two §4 dispositions: F-15/F-16/zero-toggle SHIPPED
Claude-Session: https://claude.ai/code/session_01GzammAMzsJTgpQHqxwM2bC
2026-07-13 08:48:17 +02:00
admin 0e05c6966d manifest: hub 0.52.0 (polish batch GitOps bump)
Claude-Session: https://claude.ai/code/session_01GzammAMzsJTgpQHqxwM2bC
2026-07-13 08:30:34 +02:00
admin 5d66c62bd6 hub: F-16 — inline two-step confirm partial replaces every native confirm() (offsite/PBS re-issue, freeze, telemetry reset, dismiss-all, regen-password, claim-resend, block/delete, geo-disable); hub_confirm_gate.py enforces zero
Claude-Session: https://claude.ai/code/session_01GzammAMzsJTgpQHqxwM2bC
2026-07-13 08:19:54 +02:00
admin bad9203daa hub: F-15 instant reset codes — reset-request response carries the rotated {code_hash, generation, issued_at} (same shape/guarantee as the report ACK); claim-resent flash now states the operator-resend ACK lag
Claude-Session: https://claude.ai/code/session_01GzammAMzsJTgpQHqxwM2bC
2026-07-13 08:04:53 +02:00
admin 11d0f55303 take-two re-drill record: zero-stop Day-0 on v1.15.0/0.86.0/0.51.0 — 1h22m, hands-free DR cascade, floor-update first proof, offsite round-trip; new findings F-14/F-15/F-16
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 23:13:29 +02:00
admin ccf89c0a67 manifest: hub 0.51.0 (DR-tier-by-default GitOps bump)
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 20:48:20 +02:00
admin fda783bdaf docs: DR-tier-by-default batch — decisions 1-5 in CONTEXT, day0 ACL advice reversed, ceremony decision-4 policy block, WG disclosure (tester agreement + GO-LIVE D3), drill F-disposition table, REPORT
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 20:43:25 +02:00
admin 448a68237a hub v0.51.0: DR-tier-by-default — per-customer dr_tier flag (default ON, legacy backfill from reality), cascade stages, WG-registration auto-provision hook, offsite-requires-DR guard (F-6 policy), host-page capability chips (inactive=neutral)
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 20:37:00 +02:00
admin 007946faf4 installer v1.15.0: DR-tier-by-default sweep (F-7 pbs-apply, F-10 age, F-9 wg default-on, F-1 single version + hub const, F-2 dry-run auth, F-8 rotation pointer, ACL felhom-pbs kept) + hostinstall_gates.py
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 19:52:56 +02:00
admin e205a2d6a1 test: claim engine — reset requests never alter claimed state (Scenario C non-DoS; leftover from the claim-arc session)
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 19:38:42 +02:00
admin 6040c7b93d drill report: customer-claim arc resolution + post-claim-arc snapshot + follow-up observations
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 19:20:39 +02:00
admin 8ded485a58 docs: customer-claim arc — REPORT/CONTEXT, G9 supersede, tester-agreement claim step, day0 D.3/D.4 gate, drill F-4/F-5 RESOLVED
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 18:56:46 +02:00
admin 9a14f04819 manifest: hub 0.50.0 (customer-claim arc GitOps bump)
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 18:47:03 +02:00
admin 3ded6cf7c1 hub v0.50.0: customer-claim arc docs (CHANGELOG, default template claim_code_* fields, REUSE rows)
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 18:45:43 +02:00
admin 6b40eb8619 hub: customer-claim password arc parts 1+2 — code engine, emails, ACK, configgen bake, UI (v0.50.0)
Closes DRILL-day0-vm F-4 hub-side: per-customer claim state (customer_claims,
bcrypt-only custody), the claim engine (issue at real config retrieve = Day-0
bake; first-report issue for live boxes; resend rotates generation; reset
rate-limited 3/day), three Hungarian emails via the dispatcher, report-ACK
claim object {code_hash, generation, issued_at} + set-only claimed ingest,
web.claim_code_* baked into generated controller.yaml, Setup-tab status chip
+ resend button, POST /api/v1/claim/reset-request (self-scoped), claim_lockout
event allowlisted. 13 new tests; full repo green.

Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 18:12:48 +02:00
admin b904477ed9 REPORT: hub v0.49.0 live validation complete — all 7 browser probes green
Dirty-form 103s no-reload, live-tab auto-reload, settings→edit alias,
(paused) hint, no-op save flash→Edit tab, style.css?v=0.49.0 network check.
Notes the never-reset dirty-flag semantic vs the spec's 'clear it resumes' wording.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:54:51 +02:00
admin e96e2b2361 REPORT: hub v0.49.0 edit-a train — results, red-proofs, deploy verification, operator follow-ups
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:43:46 +02:00
admin b1ab0ea0e5 docs+manifest: hub v0.49.0 — Edit tab merge train (CHANGELOG, REUSE rows, CONTEXT, GitOps bump)
Version note: the task spec targeted 0.48.0 but that tag had already shipped
(app_start_failed); published tags are never re-pointed → this train is 0.49.0.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:39:09 +02:00
admin 1d94b1a9e9 hub: scoped auto-refresh + style.css cache-bust + staging rule (v0.48.0 parts 3-4)
- Auto-refresh: the 60s reload fires only while a live tab is active
  (nav data-live-tabs="overview,applications,events,host") AND no form is
  dirty (delegated input/change listener; never reset — a reload clears it).
  Skipped ticks reschedule; a muted (paused) hint shows next to the toggle on
  non-live tabs / dirty forms. Toggle, localStorage key, 60s cadence, and
  default-on behavior unchanged. The refresh script resolves the legacy
  settings→edit hash alias like the tabs script.
- Rider 4a: every template's stylesheet link is /style.css?v={{hubVersion}}
  (the v0.47.0 gotcha: max-age=3600 served stale styling for up to an hour
  after a deploy). Red-proof run: a reverted bare link fails the test.
- Rider 4b: CLAUDE.md standing rule — never git add -A in this repo
  (the 146d165 sweep incident); explicit paths + pull-rebase + one writing
  session per clone.
- Tests: Group C structural pins (attribute read, dirty listeners, alias x2,
  hint element, cadence/key survivors) + Group D cache-bust over six pages.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:37:04 +02:00
admin 2e03de1e0c hub: merge the customer edit page into the Edit tab (v0.48.0 edit-a, part 2)
- Settings tab renamed Edit; embeds config_form_body (.ConfigForm via the
  builder) + Controller Update + Geo + a new Danger zone card holding the
  relocated Block/Unblock/Delete forms (endpoints + confirm() unchanged).
  All cards are SIBLINGS after </form> — never nested in the config form.
- Customer Info header loses the Edit link and Block/Delete forms; only the
  config-less Create Config action stays.
- GET /configs/{id}/edit is a 302 to /customers/{id}#tab=edit; tabs JS gains
  the settings→edit legacy-hash alias.
- Post-action redirects land back on their tab: update/block/unblock/
  offsite-reissue/offsite-freeze/pbsdr-reissue → #tab=edit, regen-password
  → #tab=setup; delete unchanged (/configs).
- handleConfigUpdate gains the server-side twin of the form's required
  fields; the error path re-renders the STANDALONE page with the SUBMITTED
  overrides (B3 red-proof: nil overrides → typed values reset → test FAILS;
  header red-proof: restored header buttons → count=2 → test FAILS; both run).
- Tests: Group A (panel surface, sibling forms, header cleaned by COUNT),
  Group B (B1 302, B2 create unchanged, B3 typed-values, B4/B5 anchor table).
  Amended pins: customer_tabs_test settings→edit; pbsdr_test postUpdate now
  supplies the required fields + FormRendersState asserts the embedded render.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:33:07 +02:00
admin e74014730a hub: extract config_form_body sub-template + configFormData builder (v0.48.0 part 1)
Behavior-neutral extraction (the host_detail_body pattern): the config form's
<form> + in-flight script move to a {{define}} sub-template; config_form.html
keeps the chrome. renderConfigForm/handleConfigNewForm/handleConfigEditForm
now go through the one configFormData builder (nil overrides = parse stored
ConfigJSON). Prepares the customer page Edit tab embed.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZc5w5jDhFLv6qDC32KN5v
2026-07-12 17:24:59 +02:00
admin 3e949bc513 DRILL-day0-vm 2026-07-12: full report + runbook corrections (F-4/F-6 headline findings)
Audit doc for the Day-0 VM drill: appliance install, floor-at-first-report,
escrow ceremony + auto-confirm FIRST LIVE FIRING (~7.5 min, zero clicks),
offsite backup + restore round-trip, PBS-DR/WG fork (Peti-sequence rehearsal).
Corrects day0-install.md A.2 (git creds optional since v1.11.2, anonymous
fetch is the normal shape) and RUNBOOK-escrow-ceremony.md (identity-only mode
does NOT exist — F-6; age prereq — F-10). REPORT.md overwritten; CONTEXT.md
one-liner added.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NptTCFtu7dz2Ru89qHRagN
2026-07-12 17:16:25 +02:00
admin 0ed87f5dc4 docs(pilot): finalize 0.85/0.120 train — DB floor saved, Peti deferred no-ETA
Operator saved DB global floor 0.120.0 (hub-logged 11:19:43Z); effective floor
now 0.120.0 aligned with env fallback, no skew. Peti D/E/G deferred
indefinitely (box offline, no return ETA) — fleet not waiting; gate protects
him on return.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-12 12:00:41 +02:00
admin 168e86b77d docs(pilot): publish-train 0.85/0.120 results (fleet half complete; Peti parked)
Fleet artifacts published + verified (agent 0.85.0, golden 0.120.0), hub
manifest vouches both (MinAgent 0.81.0), env floor moved to 0.120.0 (0eb9dc2,
Synced/Healthy). DB global floor = operator's last save. Peti D/E/G parked —
box offline (dead fan); gate holds him at 0.113 until his agent lands.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-12 11:15:27 +02:00
admin c322bac21a docs(break-glass): document operator-set FIXED root@pam credential (demo-felhom-01)
§5: add the standing note that a host's root@pam may be pinned to a known
operator-chosen password instead of the random day-0 one — set via chpasswd +
re-vault (PUT /hosts/<id>/recovery-credential, host api_key), guarded by the
break_glass state marker so plain installer re-runs skip it. Caveats: never
--rotate-recovery such a host; a full from-scratch reinstall wipes state.json
and re-randomizes (re-run the set-and-vault). Captures why demo-felhom-01's
root@pam "kept changing" until it was pinned 2026-07-12.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 11:15:07 +02:00
admin 0eb9dc2e9b manifests: raise global controller floor 0.113.0 → 0.120.0 (publish train)
Env fallback DEFAULT_MIN_CONTROLLER_VERSION for the CAMPAIGN-3 wave. Day-0
artifact manifest already vouches agent 0.85.0 + golden 0.120.0 (MinAgent
0.81.0). Demo already 0.120.0 (no churn); Peti offline + gated (agent 0.79 <
MinAgent 0.81 → controller held at 0.113 until his agent lands). DB global
floor to be aligned to 0.120.0 by the operator, saved last.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-12 11:12:42 +02:00
admin d02a216d0f hub v0.48.0 CHANGELOG + logging-conventions: TRACE level + ring sizing/spill (controller fix-3/fix-6, CAMPAIGN-3 closed)
🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 10:31:26 +02:00
admin 9095981d0c hub: bump manifest to v0.48.0 (app_start_failed event acceptance — controller fix-3)
🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 10:13:14 +02:00
admin d7b3c82144 hub: accept app_start_failed event (controller fix-3, CAMPAIGN-3) — allowlist + customer message
🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 10:09:54 +02:00
admin f875ac913d docs: network-storage-nas — F8 one-classification-two-surfaces + F4 mapped_uid range (Task C)
🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 09:44:35 +02:00
admin 45bff0184f docs: backup-architecture — F7 atomic volume dumps + F6 no-single-copy + F5 stale sweep + NAS locality (Task B)
🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 09:18:21 +02:00
admin 0d73ca579b CAMPAIGN-3 Task A: agent v0.85.0 boot/recovery plane docs + host-install deployment_mode (--mode gates node self-heal)
network-storage-nas.md gains Boot-ordering (F12), reassert-hardening (F11/F10/F9),
F13 limitation, and Node self-heal (appliance) sections. host-install v1.14.0
templates deployment_mode from --mode + byo assert. Live matrix: host reboot x2 =
0 ordering-cycle lines (F12 dead); appliance self-heal recovered attempt 1.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 08:26:55 +02:00
admin f97f7803c7 CAMPAIGN-3: morning RCA — host loss = F12 CRITICAL automount ordering cycle (hardware exonerated); box recovered to defined state
journalctl proved boot -1 came up at 23:31 WITHOUT networking: the agent automount
template's After/Wants=network-online.target cycles with local-fs.target and systemd
deleted networking.service to break it; the 06:45 power-cycle boot hit the same cycle
and deleted the automount instead. Every boot with an enrolled share is a coin flip
until the template drops that ordering. Recovery executed: automount started, apps
healed, privatebin removed via real flow, verification backup green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 06:55:40 +02:00
admin bed9bd2c5a CAMPAIGN-3 audit: unattended night run — data plane solid; NAS automount re-arm plane broken (F10 CRITICAL, F9/F11/F7 HIGH); demo host DOWN after 4e reboot (HUMAN)
30 PASS / 17 FAIL / 11 FINDING / 1 DISCREPANCY. Fix-6 ring numbers (6.5min horizon
under load, wiped per restart), backup-vs-NAS policy discovery, morning recovery
runbook. Evidence: 180:~/campaign3/.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017CDMFpFx84pfviCTVuGGhf
2026-07-12 04:29:48 +02:00
admin ff7a5d1c56 REPORT: hub v0.47.0 live validation ALL PASS (contrast + tabs + dr-drill deletions + offsite probe)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:59:09 +02:00
admin 658401c36d REPORT: fix self-referencing commit hash (77247f9)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:44:59 +02:00
admin 77247f9710 REPORT: hub v0.47.0 UI reorganization — implementation + deploy complete; live validation pending operator login
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:44:19 +02:00
admin 10a7bf09f7 build runbook 2026-07-11 21:43:09 +02:00
admin 95d71fef2e hub v0.47.0: CHANGELOG + README tab/host-lifecycle/offsite docs + REUSE map + manifest bump to 0.47.0
CONTEXT.md: v0.47.0 arc entry (resolves the 146d165 swept-WIP flag) + the
multi-endpoint-allocation deferral + pre-GitOps stale-template note.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:39:44 +02:00
admin 0daddcd1c4 hub: offsite multi-endpoint management UI (v0.47.0 part 5)
- /offsite lists ALL wg_endpoints rows as cards (id, address, pubkey, subnet,
  PBS addr, peers-in-subnet count) + add/edit/delete forms
- store: ListWGEndpoints (id order) + DeleteWGEndpoint (plain delete; the
  peers-in-subnet guard lives in the handler where the refusal is built);
  SetWGEndpoint upsert reused, single-expected comment updated
- guards: subnet edit refused 409 while peers sit in the current subnet;
  endpoint delete refused 409 while peers sit in its subnet; full form
  validation (CIDR, pbs ip in subnet, port 1-65535, pubkey, id charset) -> 400
- peer table gains an Endpoint column (first id-ordered subnet match; em dash
  when none); pubkey-change edit gets a type-to-confirm noting pull-based
  convergence
- allocation/reconciler/desired-state STAY lowest-endpoint-id (page notes the
  deferral); GetWGEndpoint semantics untouched
- tests: E1-E6 incl. guard red-proofs; store list/delete test

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:35:24 +02:00
admin 068427a729 hub: restore the host-delete escrow-ack gate (red-proof mutation swept by 146d165)
The parallel session's git add -A swept the v0.47.0 Part-4 WIP mid-red-proof:
hosts.go landed with the RED-PROOF 2 mutation (deleteEscrow := true), i.e. the
escrow acknowledgement bypassed. This restores the real gate; the swept store/
template/test files were captured in their correct final state (full go test
green). Red-proofs 1/2/5 all ran and FAILED as required before this point.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Vvz1NCu22p8dGkRCpeX9re
2026-07-11 21:28:04 +02:00