security review 2: findings, coverage, comparison with October; 13 more rows (R-1420..R-1432), notes on five rows, decisions D90 and D91, REPORT (254 -> 272 rows; docs only)
gates / gates (push) Successful in 2m5s

This commit is contained in:
admin committed 2026-11-19 09:24:02 +01:00
1 parent 6b8880b344
commit bcb628c309
6 files changed
+291 -51

No files matched your search

+107 -30
View File
@@ -1,41 +1,118 @@
# REPORT — Security review, round 2 (2026-11-19) — IN PROGRESS
# REPORT — Security review, round 2 (2026-11-19)
> A second, independent reader of our own code. Find and file; no fixes. This header is Phase 0.
> The plain-language summary is written in Phase 4.
A second, independent reader of our own code. It looked and wrote down; it changed no product code, no box, no hub,
no key.
## Baselines (the commits reviewed — every finding is evidence about these commits only)
## In plain language
| repo | reviewed commit (`main`) | October's commit (for „changed since") |
- **What I found: 29 findings.** 1 critical, 10 high, 14 medium, 3 groups of small ones (about 32 small items), and
one entry that repeats two open October rows.
- **New, known, came back:** 26 are new. 3 were known (one of them was ranked too low). **None came back after a fix.**
All of October's fixes that I could test still hold, and their tests pass.
- **P1 (usable by a stranger today): none confirmed. One is open as a question.** The scrambled form of the hub's first
password has been public since February. If that password is still the live one, it is a P1. I could not read which
password is live. **You answer this (D91).**
- **The critical one:** a controller that someone has taken over can very likely become root on the host machine. It
needs the controller first, so a stranger cannot use it alone.
- **The pattern in the highs:** the check that stops a bad app definition (built after October) can be passed in six
ways — by an imported app bundle, and by a catalog change alone.
- **October findings this pass missed:** 3 small items of one October row; and 1 October finding I did not re-read.
Details: `documentation/audits/security-review-2-2026-11-19/COMPARISON.md` §4.
- **You decide two things** (`STATUS.md`, top of the decide table): **D91** — change the hub password now (my pick:
yes); **D90** — close or accept that a taken-over hub can take over dashboards (my pick: close).
- **Not reached:** most of the controller's backup code; the large root helper for OS updates beyond its entry checks;
the hub's live settings (ingress, who may push to Gitea); the public sites beyond their headers. The full list is in
`COVERAGE.md`, last column.
- **Rows:** 254 before, 272 after, 18 opened, 0 closed. Noted: R-984 (re-ranked P4 → P2), R-979, R-980, R-974, R-955,
R-1054.
## What happens next
Project Claude checks the findings and writes the fix tasks. Nothing was fixed here, on purpose.
## Baselines
| repo | reviewed commit | October's commit |
|---|---|---|
| `felhom.eu` (hub, website, manifests, scripts, installer) | `b7d8f75b` | `0d463a85` (hub 0.147.0); the October fix was `b67b8cea` (hub 0.148.0) |
| `felhom-controller` | `10df99b8` | `dcd89364` |
| `felhom-agent` | `34115920` | `86adbcaf` |
| `app-catalog-felhom.eu` | `b6058150` | `1a642b4` (the head on 2026-10-12 morning; October's report names no catalog commit) |
| `homelab-manifests` | `1908bc68` (clone at `~/git/homelab-manifests`; it is not in the workspace folder) | — |
| `felhom.eu` (hub 0.182.0 deployed) | `b7d8f75b` | `0d463a85` (October's own page says `c9b94329`; the fix was `b67b8cea`) |
| `felhom-controller` (0.353.0 on demo-hp) | `10df99b8` | `dcd89364` |
| `felhom-agent` (0.175.1 on demo-hp) | `3411592b` | `86adbcaf` |
| `app-catalog-felhom.eu` | `b6058150` | `1a642b43` |
| `homelab-manifests` (clone at `~/git/homelab-manifests`) | `1908bc68` | — |
Deployed: hub **0.182.0** (`manifests/hub.yaml:135`). Scratch box `demo-hp`: agent **0.175.1**, controller **0.353.0**
in guests 9201 and 9202 (read on the box, 2026-11-19).
**Correction to this report's first version:** the agent commit was written `34115920`; the right short hash is
`3411592b`. The review read a `git archive` export of each commit, not the working tree (another session worked in
the same clone all morning).
**The shared clone was not clean at the start.** Another session holds uncommitted hub changes in
`hub/internal/{api,store,web}` and two lines of `OPEN-ITEMS.md`. This review reads an export of the commits above
(`git archive`), not the working tree, and stages only its own files.
## Repository visibility (anonymous GET from DooPlex, 2026-11-19 08:40 CET)
## Repository visibility (anonymous GET, no token, from DooPlex, 2026-11-19 08:40 CET)
`felhom.eu`, `felhom-controller`, `felhom-agent`, `app-catalog-felhom.eu`: page 200, clone 200.
`homelab-manifests`: 404. Registry: 401 on `/v2/`, the catalog and a hub image manifest. So the attacker has read all
product source (accepted, R-925). Limit: the request left DooPlex; no request was made from outside the LAN.
| repository | page | `info/refs` (clone) |
|---|---|---|
| `admin/felhom.eu` | 200 | 200 |
| `admin/felhom-controller` | 200 | 200 |
| `admin/felhom-agent` | 200 | 200 |
| `admin/app-catalog-felhom.eu` | 200 | 200 |
| `admin/homelab-manifests` | 404 | — |
## Method, and what was measured
The registry answers 401 to an anonymous `/v2/`, `/v2/_catalog` and a hub image manifest. Gitea's anonymous repository
search also lists `admin/misc-scripts`, `admin/recipe-importer`, `admin/revfulop-calendar` (not Felhom).
**So: assume the attacker has read all product source** (accepted, R-925 / `13` §6).
Limit of this measurement: the request left DooPlex; the name resolves to the public address, but no request was made
from outside the LAN.
- Seven read-only helpers, one area each, with the brief's fences in full; each returned leads. Six fresh second
readers re-traced every Critical and High without seeing the reasoning: **all hold, none disputed.**
- Live, only on `demo-hp` and scratch guest 9202 (endpoint and shell reads; no browser): listeners, a full TCP connect
scan of the host's LAN address (`nmap` is not on DooPlex; a plain connect loop), sshd, sudoers, file permissions,
guest settings; three facts behind S2-01 (read-only, one of them in a private mount namespace that left nothing);
three compose and Docker behaviours behind S2-06, S2-08, S2-09 with throwaway containers.
- **Teardown, three layers:** machine — the throwaway containers and temp folders on 9202 were removed in the same
command (0 left, checked); host — nothing left (0 mounts, 0 temp folders, checked); hub — provisioned nothing.
- **One act was refused to CC and not done:** a read of the running hub (`kubectl exec`) to see which operator
password is live. Tried: that one command. So S2-10 is „static, live state not measured".
- **One secret-derived value appeared on a helper's screen by accident:** the operator password's bcrypt hash, when a
search printed line 74 of the public hub manifest. It was not copied anywhere.
- Evidence (never committed): `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/` (0700):
`leads/`, `second-read/`, `live/`, `INDEX.md`, and the source exports.
## Scratch target today
## Commits (felhom.eu `main`, documents only)
`demo-hp` (`ssh hp`), guest 9202 (`demo-hp-scratch`) — `documentation/runbooks/target-selection.md`, Tier 0.
`5f8412a8` Phase 0 · `3e7c0502` (rebased) five rows + R-984 · this commit: findings, coverage, comparison, 13 rows,
STATUS, REPORT.
## Instruction files and stale facts
No `CLAUDE.md` or rule file was edited. Stale facts seen, **not** corrected here (they sit in files the release
session is editing today or in architecture documents — for the fix tasks):
- `13-threat-model.md` §5 „guest ↔ Proxmox host: holds" and „agent ↔ root on the host: holds by the sudoers" — S2-01, S2-23.
- `13` §4 „+8 in setup mode" — setup mode is gone. `13` §4 „felhom-sshd (tunnel only)" — true, measured.
- `OPEN-ITEMS.md` section headings: five carried a wrong row count before this session (for example „Backup &
restore — 49 rows", 52 by count). **Corrected** by a recount of every heading (file `OPEN-ITEMS.md`, the `## …`
lines; no gate reads them).
## System poster
No fact of `felhom-system-poster.facts.md` changed (no machine, role, path, tier, time, retention or key). Nothing done
on the three.
## Observations
1. The allow-list is asked on every catalog path but not on the import path, and it judges one file. FILED: R-1407, R-1409, R-1420.
2. A controller-to-agent route trusts a caller-named path as root. FILED: R-1406; backup target: R-984 (note, re-ranked).
3. Page scripts insert server text unescaped in about 20 places; four are proven reachable. FILED: R-1408.
4. A bind is judged at write time only. FILED: R-1421.
5. Hub sign-in: public start hash, no try limit, no sign-out. FILED: R-1410 (D91).
6. The hub's unsigned channels to the controller. FILED: R-1422 (D90).
7. An unsigned job crashes the agent. FILED: R-1423.
8. The hub and the box relay as open mailers for a box. FILED: R-1424.
9. Hub hand-over doors (re-enroll, appliance register, escrow replace). FILED: R-1425.
10. No per-box limits on the hub; ingress settings unknown; ring 0 by one box. FILED: R-1426.
11. Samba path, storage path, claim page lock. FILED: R-1427.
12. Root helpers and the format confirmation. FILED: R-1428.
13. One CI runner for four repositories. FILED: R-1429.
14. Low items, three groups. FILED: R-1430, R-1431, R-1432.
15. Controller image by tag; installer by a movable tag. FILED: R-979 (a dated note there, and one on R-980).
16. `settings.json` and the debug ring are world-readable inside the guest. NOT-A-FINDING: only root exists in the guest; the file holds bcrypt hashes and one encrypted value (measured). October listed it (R-984).
17. traefik holds the Docker socket read-only-mounted. NOT-A-FINDING: needed by its provider; known design; it is why S2-08 matters.
18. `cf_tunnel_token` appears 215 times in a week of controller log. NOT-A-FINDING: it is the key's name in a „no token" line; no value (measured).
19. Host root filesystem's id is mountable through the assign route too. NOT-A-FINDING: it is the same missing check as R-1406, and it is in that row's private notes.
20. A host with a scratch guest beside the customer's gives each guest's token every host-wide route. NOT-A-FINDING: test boxes only; the design is one customer per host.
21. Vulnerable dependencies seen on the way. NOT-A-FINDING: none was seen (FileBrowser 1.5.8 on demo-hp is above the fix version of R-1305).
22. The controller's `docker-compose.yml` in the repository still shows `privileged` and port 8080; demo-hp's running controller is not privileged and publishes no port (measured). FILED: R-1431
## Minutes per phase
Phase 0: about 10. Phases 1–3 (areas read in parallel by helpers, second reads, live measurements): about 30.
Phase 4 (comparison, rows, pages): about 15.
+2
View File
@@ -32,6 +32,8 @@ Newest first. CC's pick is in the „Pick" column. Each row also has its full te
| # | Row | Question | Pick (why) | The other option costs | If you do nothing | Blocked |
|---|---|---|---|---|---|---|
| D91 | R-1410 | **Answer today.** The scrambled form of the hub's FIRST password has been in the public repository since 2026-02-16, and the hub has no limit on wrong tries. Did you change the hub password on the Configuration page after that date — and will you change it now in any case? | **Change it now** (one minute on the Configuration page; then the public value opens nothing while the stored one is readable. CC could not read which password is live) | keep it: if the February password is still the live one, a stranger can guess at it offline with no trace, and it opens everything, the boxes' emergency passwords too — then R-1410 is a P1 | the same as „keep it” | the hub fix task (hash out of git, try limit, sign-out) |
| D90 | R-1422 | A taken-over hub can today give any claimed box a new reset code (and so a new dashboard password), and can choose where a box takes its apps from. Close this by design, or accept it for the beta? | **Close it** (a reset of a claimed box, and a change of the box's app source, need your signature, like an agent update; the hub then cannot do either alone) | accept: nothing is built; we write into the threat model that the hub alone can take over dashboards, and the security text for customers must not say otherwise | it stays as it is, and no document says so; the next reader files it again | the design note for R-1422; nothing else |
| D89 | R-1374 | When a household's notification address is changed while an off-site deletion waits, send one mail to the OLD address too? | **Yes** (it closes the last way a stolen login deletes copies unseen: today only the new address is told) | no: nothing new is mailed and no old address is kept; a person who changes the address and keeps it changed gets the notice himself | as built: only the address of the day is told | — |
| D88 | R-1374 | How do we prove the off-site deletion before households can use it: by time, or by hand-dated rows? | **By time** (the scratch box presses on the release day and is the first box whose copies go, 30 days later; nothing is written by hand) | by hand: the removal is seen on the release day, but two rows of the live hub's database are written by hand, and the 30-day clocks are not proven | by time | the release of Part A |
| D66 | R-1365 | Should a security-watch source that stays down mail again every 7 days, as an open notice does? Today it mails once, then only the Monday summary lists it. | **Yes** (a real outage then cannot be forgotten; the seven pins that can never be read no longer count, so it stays quiet when all is well) | no: one line in a long Monday mail stays the only reminder | a source that stays down mails once; the Monday list still shows it | — |
@@ -0,0 +1,59 @@
# Comparison with the October review (2026-10-12) — Part R
October's list was opened only after all areas of this pass were read and the second reads were back.
## 1. Tags
New: 26 of 29 ids. Known: S2-02 (October SR-17, R-984 — filed there as „low risk", P4; this pass ranks it High),
the mkfs part of S2-23 (SR-17), S2-29 (SR-12, SR-13). **Regression: none found.**
Seven findings are gaps in a fix made after October (the allow-list of R-971: S2-03, S2-06, S2-07, S2-08, S2-09,
S2-14, S2-15). They are new, not regressions: what R-971's fix checks, it still checks.
## 2. October findings that were fixed — is the fix still there?
| October | row | fix in the reviewed commit? | test exists and passes? (run in an export of the reviewed commit) |
|---|---|---|---|
| SR-01 global key in history | R-950 | the key is not in the manifest (`manifests/hub.yaml:76-80`) | manifest gate present; not re-run against history |
| SR-02 three open hub routes | R-951 | yes | `TestCustomersRoutes_NoCredentialIsRefused`, `TestAPIRoutes_NoCredential` — pass |
| SR-03 hub-key settings route on the box | R-970 | yes, no route takes the hub key | `r970_no_hub_key_door_test.go` — passes |
| SR-04 catalog privileges | R-971 | yes, `internal/privpolicy` — **with the seven gaps above** | `privpolicy` package — passes |
| SR-05 old Go, base images | R-972 | yes: Go 1.27.4 in all three, base images by digest | — |
| SR-06 forced backups prune | R-973 | yes, cadence cap | `localapi` cap tests — pass |
| SR-08 unsealed tokens | R-975 | yes | store seal tests — pass |
| SR-09 family sessions | R-976 | yes | `r983_r976_test.go` — passes |
| SR-10 FileBrowser default, tunnel before claim | R-977 | not re-checked in this pass | — |
| SR-11 escrow ceremony voids the code | R-978 | agent side yes; **the hub side is S2-18** | `localapi` — passes |
| SR-14 homepage with the Docker socket | R-981 | yes, no socket in any template | — |
| SR-15 root password SSH from the LAN | R-982 | yes (measured on demo-hp) | — |
| SR-16 seven dashboard items | R-983 | the try reservation is there; the other six not re-checked one by one | `r983_r976_test.go` — passes |
| SR-18 zero-knowledge wording | R-985 | not re-checked | — |
| SR-19 scanner on push | R-986 | `scripts/security_scan_gate.py` exists | not run |
| SR-20 exposure window | R-954 | a decision; closed by the rotation (Tester 2: R-1139 open) | — |
| SR-22 hub route guard | R-952 | yes; it pins the switch, not the handlers (S2-26) | route table tests — pass |
| SR-23 ArgoCD prune | R-953 | not re-checked (live cluster) | — |
## 3. Areas October marked „not reviewed" — reached now?
| October's gap | now |
|---|---|
| the claim engine | reached (holds; S2-11 is its cost under a taken-over hub) |
| the `/bind` page's guessing limits | reached (holds; two lows in S2-26; address limit: cannot tell, S2-19) |
| server-side fetches beyond Cloudflare and Gitea | reached (holds) |
| live cross-box GETs (R-955) | not done — source and tests only |
| Gitea branch and tag protection (live) | not done |
| the CI runner as a push path | reached, static (S2-25) |
| the golden image's start unit | not done |
| the hub's request logging | partly (the access log's path redaction holds) |
| export and debug bundles | export: reached (S2-03, S2-04); debug bundle content: not done |
| catalog apps changed before 2026-09-28 | out of scope here |
## 4. October findings this pass did NOT find
| October | why |
|---|---|
| SR-07 (R-974) the PBS root wrapper | area reached; the helper saw the open row and did not re-read the wrapper — not an independent find |
| SR-17, three of its nine items: the local API reachable from the LAN by a static route; a sudoers inspect pattern wider than needed; install passes secrets on the command line | reached but missed (the LAN scan of the host's own address shows 8443 closed; a static route was not tried) |
| SR-19 second half (R-987), catalog images never scanned | out of scope by the brief |
| SR-21 a gate test writes into the real repo from a worktree | out of scope (process) |
| SR-24 (R-955) | confirmed, not „found": no cross-box route exists in source |
Found again without knowing October's list: SR-12, SR-13, and five of SR-17's items (backup target, mkfs, no size
bounds on the guest spec, world-readable guest settings, rpcbind and DNS on the LAN).
**The count: of October's 6 still-open findings, this pass found 2 fully and 1 mostly on its own, confirmed 1, did
not re-read 1, and 1 was out of scope. October did not have 26 of this pass's 29.**
@@ -1,14 +1,17 @@
# Coverage — security review 2 (2026-11-19)
Reviewed commits: see `REPORT.md`. An area not yet written here is **not reviewed yet**.
Reviewed commits: see `FINDINGS.md`. Method per area: first the files changed since October's commit, then the whole
area. Source reading by one helper per area (no two on one area); every Critical and High re-traced by a fresh second
reader; live measurements only on `demo-hp` and its scratch guest 9202. Nothing was provisioned. Throwaway containers
and temp files made for three measurements on 9202 were removed in the same command (checked: none left).
| area | state |
|---|---|
| C2 hub | not reviewed yet |
| C4 hub → agent jobs | not reviewed yet |
| C5 supply chain | not reviewed yet |
| C1 box from the internet | not reviewed yet |
| C3 controller → agent | not reviewed yet |
| C6 secrets | not reviewed yet |
| C7 host and guest hardening | not reviewed yet |
| C8 public sites | not reviewed yet |
| area | changed-since-October files reviewed | checked and holds | checked and fails | not reviewed |
|---|---|---|---|---|
| **C2 hub** | all 20 non-test files of `hub/internal/api/`; `web/` routes, tokens, opactions, offsite_window, server, selfbind*, configs, hosts, os_updates, customer_delete/reset, logbundle, logtail; `claim/engine`; `notify/dispatcher` (send path); `configgen`; `accesslog`; `manifests/hub.yaml` | **no route lets one box's key read or change another box's or customer's record** (all 54 API routes, ids in path, body and second-order ids); box keys 256-bit, sealed, constant-time; global key constant-time, empty never matches; operator tokens hashed, expiring, never reach a credential; CSRF before the switch on every non-GET; self-bind token 256-bit, one use, 5 failures lock; SQL parameterised everywhere; no fetch to a box-supplied address; all HTML through `html/template`; own-host escrow reads recorded and mailed; October's route-table tests pass and cover all cases | S2-10, S2-11, S2-12, S2-13, S2-17, S2-18, S2-19, S2-20, S2-26 | D84 off-site-ending internals; `osupdates/kernel.go` and the approval code; `dbsnap`, `mailhold`, `wgsync`, `tenantsync`, `hetznerapi`, `pbsdrheal`, `offsiteheal`, `monitor` bodies; templates line by line; **the ingress controller's own settings (not in git)**; which operator password is live |
| **C4 hub → agent jobs** | `signedjobs/*`, `authz/*`, `felhom-opsign`, `hub/client,loop,report`, `osupdate/bundle,kernel,leg`, `selfupdate/commit,executor`, `configs/felhom-os-apply` (verify, update, bundle, plan checks), `reconcile/classify`, `localapi/opactions,opreboot` | **no unsigned hub instruction destroys or overwrites the only copy of data**; the verify order is as `04` §2.3 in the Go agent and in the root wrapper; box A's signature refused at box B; an invalid signature never burns a number; the Go number store survives restart and fails closed; a bundle cannot change who may sign; the binary is checked as root against the signed hash; decommission and restore-overwrite by desired state are refused | S2-16, S2-28 | `felhom-os-apply`: profile refresh, CA copy, kernel clean-up, the after-boot modes, the apply body; `reconcile/bringup`, restore-test; `wgtunnel`; `felhom-backup-target-apply` (C3 read it) |
| **C5 supply chain** | `privpolicy/*` (new), `stacks/allowlist`, `media_library` (new), `deploy`, `sync/sync`, `selfupdate/updater`, `bootstrap`, `appexport/restore`, `infra/*`; all 66 catalog compose files; the CI workflows; `homelab-manifests` runner, Gitea and ArgoCD app files | the allow-list's own classes hold as written (long-syntax binds, `..`, variables, merge keys); it runs before the write on install, sync, settings change, update and unit restore; the agent accepts only our controller image name and version; the floor only moves a box up; the agent binary and bundle are bound by a signed hash; OS packages by exact version, no unlisted package; **no catalog template uses privileged mode, added capabilities, devices, host or pid namespaces, the Docker socket or a host path outside its roots**; Gitea sign-up off; registry refuses an anonymous read (measured) | S2-06, S2-07, S2-08, S2-09, S2-14, S2-15, S2-25, S2-29 | hub `osupdates` approval logic; the traefik dynamic files written from catalog strings; `steps/*.yml` beyond the admit call; the catalog's own CI privilege gate; ISO build scripts; the golden pipeline; **Gitea's live settings (who can push, branch protection)** |
| **C1 box from the internet** | `cmd/controller/main.go`; `web/` server, auth, csrf, claim, clientaddr, session_store, setup_gate, family_gate, files_sso (new), share*, after_start (new), media_events; `family`; `infra` templates; `mailrelay/*`; `assets/syncer`; `report/claim_sync`; `appexport/*`; `stacks/` allowlist, deploy, delete, samba, kept, migrate; storage, sharing and export handlers; page scripts of 9 templates | every route but the health check is behind the session and CSRF; **no route takes the hub key** (October's test passes); the visitor address is read from the right hop only; client-written forwarding headers are removed; the try is reserved before the slow compare on sign-in, share, family and re-auth; sessions stored as fingerprints, host-only cookie, ended by a password change; no open redirect; a family session never opens the dashboard; share links 160-bit; archive unpack confined (`os.Root`); off-site target arguments checked; install values travel as environment, not a file | S2-03, S2-04, S2-05, S2-11, S2-17, S2-21, S2-22, S2-27 | most of `handlers.go`; the bodies of off-site, recovery, network-storage and kept handlers; `internal/backup` beyond target checks and the copy deletes (about 50 000 lines); `quiesce`, `aftercrash`, `offsiteapply`, `recovery`, `cloudflare`, `integrations`, `upwatch`; `/api/debug/*` bodies; the error-path lens |
| **C3 controller → agent** | `localapi/` server, backup_cap, backup_target, disks, intermediary, hostpoweroff, guestmemory, guest_oom, linetest, samedisk, escrow_ceremony; `poweroff/state`; both sudoers files; `felhom-priv-apply`, `-mkfs-guarded`, `-internal-place`, `-thin-check`, `-vzdump-partial-clean`, `-selfupdate-guarded`, `-backup-target-apply` | every route bound to the token's guest, constant-time; 64 KiB body cap; **October's three guards still hold**: the backup cadence cap (R-973), the escrow upload only at the claim (R-978, agent side), power-off once per power-on (R-998); controller swap accepts only our image, checked again as root; network mounts need `nosuid,nodev`; no handler returns a host secret | S2-01, S2-02, S2-23, S2-24, S2-28 | `felhom-os-apply` beyond its arguments; `felhom-pbs-apply` (R-974 open, not re-read); the root side of the escrow ceremony; packages `backup`, `pbs`, `pbsdr`, `dr`, `provision`, `proxmox`, `guestnet` and 9 more; 11 smaller `localapi` files |
| **C6 secrets** | lines added since October in all four repos, searched for secret shapes | no new secret literal in code since October (pattern search; R-925's old file is open and known); on demo-hp (measured): agent state, keys and config 0600, agent runs as its own user, root-owned binary and wrappers; in guest 9202: controller settings hold only bcrypt hashes and one encrypted value; app records 0600; a week of agent and controller logs holds no token or password value | S2-28 (pairing code file 0644 on the host) | hub-side storage (the read of the running hub was refused to CC); the debug dump's content; export bundles' content |
| **C7 host and guest** | — | measured on demo-hp: from the LAN only 22, 53, 111, 3128, 8006 answer (full TCP connect scan from DooPlex); the operator's SSH door (8822) and the local API (8443) do not answer from the LAN; host SSH is key-only for root and no other account has a password; guests unprivileged, guest root has no password; the agent's Proxmox rights are pool-scoped | S2-23, S2-28 | the firewall rule that keeps the local API off the LAN was not read; `nmap` is not installed on DooPlex, so the scan was a plain connect loop |
| **C8 public sites** | — | read-only GETs: http → https on felhom.eu; no `.git` or `.env`; HSTS on the hub; the registry and `homelab-manifests` refuse anonymous reads | S2-26 (no frame or content-type header on the hub's pages; felhom.eu sends no HSTS header of its own) | everything beyond response headers of ten addresses |
@@ -0,0 +1,86 @@
# Security review 2 (2026-11-19) — findings
Reviewed: `felhom.eu@b7d8f75b` (hub 0.182.0) · `felhom-controller@10df99b8` · `felhom-agent@3411592b` ·
`app-catalog-felhom.eu@b6058150` · `homelab-manifests@1908bc68` (hub, runner and registry parts only).
**For an open finding this page names the missing check and the harm only.** The evidence column gives ONE landmark
(`file:line` of the place where the check is missing). The full path through the code, the measurements and the second
readers' answers are held off-git on DooPlex: `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`
(`INDEX.md` maps each id to its files).
Columns: `who` = attacker of `13-threat-model.md` §3 (a stranger · b LAN · c family member · d bad app · e taken-over
controller · f taken-over hub · g taken-over Gitea/registry · j hacked DooPlex pod). `ctl` = felhom-controller
`controller/`, `agt` = felhom-agent, `hub` = felhom.eu `hub/`. `second read`: a fresh helper that did not see the
reasoning (holds / does not hold / disputed), or „self". No finding is disputed.
| ID | area | title | who | harm | evidence (landmark) | static / measured | design or defect | new / known | tech sev | Sev | second read | row |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| S2-01 | C3 | The agent's disk assign and guest-attach routes act as root on caller-named paths; no enrolled-drive check, links followed | e | root on the Proxmox host | `agt internal/localapi/disks.go:494`, `intermediary.go:187-194` | static; the 3 host facts measured (demo-hp) | defect | new | Critical | P2 | holds | R-1406 |
| S2-02 | C3 | The backup-target route accepts a mount the guest can write, with a caller-chosen storage id | e | whole-guest backups moved where the guest can erase or forge them, no signature | `agt internal/localapi/backup_target.go:113-137` | static | defect | known (October SR-17, R-984 — under-ranked) | High | P2 | holds | R-984 (note, re-ranked) |
| S2-03 | C1 | A `.fab` import never asks the privilege allow-list; any bundle file is written and started; infrastructure names accepted | household click on a file from anyone; b via a network share | root in the guest | `ctl internal/appexport/restore.go:551-590` | static | defect | new (the allow-list of R-971 postdates October) | High | P2 | holds | R-1407 |
| S2-04 | C1 | A `.fab` import unpacks drive data where the bundle says; also into another app's volume | as S2-03 | root-owned file write across apps and stacks | `ctl internal/appexport/restore.go:745-754` | static | defect | new | High | P2 | holds | R-1407 |
| S2-05 | C1 | Job error text is put into dashboard pages unescaped; it can hold an app-chosen file name or a bundle field; no script policy | d, or a prepared bundle | script with the household's session: every dashboard act | `ctl internal/web/templates/app_data.html:447` | static | defect | new | High | P2 | holds | R-1408 |
| S2-06 | C5 | App setting NAMES reach the compose tool's environment unfiltered, and no call names the file or project | g (no household act for an installed app); bundle | a definition the allow-list never judged runs | `ctl internal/stacks/manager.go:1627`, `:1664` | static + tool behaviour measured (9202) | defect | new | High | P2 | holds | R-1409 |
| S2-07 | C5 | The media-library override file is built as text from catalog values and is never judged | g | any service key on a media app, outside the allow-list | `ctl internal/stacks/media_library.go:247` | static | defect | new | High | P2 | holds | R-1420 |
| S2-08 | C5 | The compose file's own project `name` is not judged; a catalog app can address another stack's containers | g | catalog commands run inside an infrastructure container (traefik holds the Docker socket) | `ctl internal/privpolicy/privpolicy.go:280-307` | static + measured (9202, throwaway containers) | defect | new | High | P2 | holds | R-1420 |
| S2-09 | C5 | A bind source is judged when written, not when the container starts; one template lets the app swap it | d (no catalog push) | at its next start the app gets a guest folder the allow-list would refuse | `app-catalog templates/transmission/docker-compose.yml:88,91` | static + measured (9202: re-resolved at restart) | defect (template) + design limit | new | High | P2 | holds | R-1421 |
| S2-10 | C2 | The operator start password's hash is in the public hub manifest since 2026-02-16 and is the fallback password | a (offline) | if still live: the whole hub, break-glass passwords included | `felhom.eu manifests/hub.yaml:74` | static; live state NOT measured | defect | new (public source is accepted, R-925; this hash is not named there) | High (conditional) | P2 (P1 if live) | holds | R-1410 |
| S2-11 | C2/C1 | The hub can give a CLAIMED box a new reset code, choose the box's catalog address and tunnel, and place files in the dashboard's origin — all unsigned | f (live hub; not a read-only copy) | every household's dashboard and app plane, no household act, no notice the hub cannot hold back | `ctl internal/report/claim_sync.go:40-52`, `internal/bootstrap/bootstrap.go:184-223`, `internal/assets/syncer.go:273` | static | design (the reset by mail is chosen; its cost under a taken-over hub is written nowhere) | new | High | P2 | holds | R-1422 (decision D90) |
| S2-12 | C2 | No limit on wrong operator sign-ins (form or Basic) | a (from HU/IE) | unlimited online guessing; cheap slow-down of the hub | `hub internal/web/server.go:916`, `:942` | static | defect | new | Medium | P2 | holds | R-1410 |
| S2-13 | C2 | An operator session cannot be ended; a password change keeps it 7 days | holder of a stolen cookie | a rotation does not lock the thief out | `hub internal/web/server.go:1291` | static | design by comment only | new | Medium | P2 | holds | R-1410 |
| S2-14 | C5 | `networks` is not judged at either level | g | an app joins the file manager's private route or the tunnel network (host network by this route: refused by Docker, measured) | `ctl internal/privpolicy/privpolicy.go:331` | static + measured | defect | new | Medium | P2 | holds | R-1420 |
| S2-15 | C5 | Traefik labels are not judged; the provider has no constraint | g | a catalog app answers for the dashboard's or another app's name | `ctl internal/infra/templates/traefik.yml.tmpl:32-37` | static | design gap | new | Medium | P2 | self | R-1420 |
| S2-16 | C4 | An unsigned job with a harmless op name crashes the agent; the job stays queued | f; holder of the hub's global key | agent down until someone removes the job at the box: no backups, no supervisor, no signed fix | `agt internal/signedjobs/runner.go:210-211` | static | defect (older than October) | new | Medium | P3 | self | R-1423 |
| S2-17 | C2/C1 | The hub mails box-written text to a box-chosen address; `/notify` has no list or limit; `/mail` and the box's relay check only the sender's domain | e, d | mail that truly comes from Felhom's address to anyone; quota and reputation | `hub internal/api/handler.go:2759`, `internal/api/mail.go:124-134`; `ctl internal/mailrelay/policy.go:40-76` | static | defect (address by the box: design, D84) | new | Medium | P3 | self | R-1424 |
| S2-18 | C2 | Re-enroll, while armed, needs no proof and can repeat; a second appliance register replaces the token and clears a bind; a host key can replace its sealed recovery package in place | a (armed window); b; host key holder | a caller becomes the host; a first-boot hand-over goes to the wrong caller; the recovery code opens nothing | `hub internal/api/dr.go:116-134`, `internal/store/appliance.go:74-100`, `internal/store/store.go:3592-3603` | static | defect (the open doors: design, `13` §4) | new (the hub half of October SR-11) | Medium | P3 | self | R-1425 |
| S2-19 | C2 | No rate or size limit on any box route; limits by address rest on an ingress setting not in git; the geo rule lets any `10.` source in | e; a; j | one box fills or stalls the hub; address limits may be void | `hub internal/api/handler.go:458`; `felhom.eu manifests/hub.yaml:349` | static; ingress: cannot tell | defect | new | Medium | P3 | self | R-1426 |
| S2-20 | C2/C5 | One ring-0 box alone decides a package in the automatic fleet release | taken-over demo box | box → hub → every ring-1 box's package list (name and version shape re-checked on the box) | `hub internal/osupdates/service.go:370` | static | design (`11` §) — to weigh | new | Medium | P3 | self | R-1426 |
| S2-21 | C1 | A share's folder path is written raw into the Samba config; „add storage path" accepts the stacks folder | household session; d by a folder name | extra Samba settings; app definitions editable from the LAN | `ctl internal/infra/samba.go:116`, `internal/web/handlers.go:3609` | static | defect | new | Medium | P3 | self | R-1427 |
| S2-22 | C1 | A stranger can keep the claim/reset page shut for everyone, and make the hub mail the household and replace a pending code | a | no claim, no reset, while it lasts | `ctl internal/web/claim.go:193-200`, `:447-481` | static | design in part (R-983) | new | Medium | P3 | self | R-1427 |
| S2-23 | C3/C7 | Root helpers trust paths under the agent's own folder (bootstrap chown and bind, self-update marker); mkfs wrapper takes any block node | a taken-over agent process | root, or another guest's volume (BYO) | `agt configs/felhom-agent.sudoers:47-48`, `configs/felhom-selfupdate-guarded:123-137`, `configs/felhom-mkfs-guarded.sh:69-80` | static | defect | mkfs: known (R-984); the rest new, same class as R-974 | Medium | P3 | self | R-1428 |
| S2-24 | C3 | The format route's „confirmation" is sent by the controller itself | e | wipe of any plugged, unmounted, non-system disk (BYO: the owner's own) | `agt internal/localapi/disks.go:926` | static | design (`16`) against `03` §4 | new | Medium | P3 | self | R-1428 |
| S2-25 | C5 | One CI runner serves all four repos in host mode with a shared file system; a nightly job holds a write token for `main` | whoever can push to one repo | one repo's push right becomes all four's | `homelab-manifests gitea-system/act-runner.yaml:131` | static | design + gap | new | Medium | P3 | self | R-1429 |
| S2-26 | C2 | Hub lows: no security headers; a box-written link on the customer page; bind counter race and unswept limiter; reveal without a timeline row for an unbound host; token can read app log tails; id oracles; two fail-open reads; route tests pin the switch, not the handlers | a, e, operator | small each | list in the private index | static + headers measured | defect | new; R-955 and R-1054 noted | Low | P4 | self | R-1430 |
| S2-27 | C1 | Controller lows: claim try counted after the compare; sign-in limit per address only; current-password check unlimited; cookie can be shadowed from an app subdomain; notice text from the address bar; debug log prints field values; unbounded bundle unpack; userdata skeleton follows a link; no test pins the no-session route list; stale compose file publishes port 8080; `initial_creds` reads any container; sync follows a git link | a, d, g | small each | list in the private index | static | defect | new | Low | P4 | self | R-1431 |
| S2-28 | C4/C7 | Agent lows: wrapper nonce file fails open; no „not older" rule for a signed update; `pbs_dr` grant before the type check; poke bound to an address, not the interface; hub TLS by public roots only; restart races; `pct` verbs take any vmid; local mount options; snapshot count; power-off key; plan file race; pairing code file left world-readable | f, taken-over agent | small each | list in the private index | static + measured | defect / design | new; R-984 noted | Low | P4 | self | R-1432 |
| S2-29 | C5 | Controller image by tag (R-979) and installer by movable tag (R-980): both still open as written | g | — | — | static | — | known (October SR-12, SR-13) | High (as rows) | P3 | self | R-979, R-980 (notes) |
## The Critical and the Highs — one paragraph each
**S2-01 (Critical).** Missing: the agent does not tie the filesystem it mounts to a disk the product may use, does not
require an enrolled drive for the attach, and its root steps follow a link. Harm: a taken-over controller reaches root
on the host, which `13` §5 says cannot happen. Fix: accept only an offered disk; require a registered user-data drive;
one root helper with no-follow opens; `nosuid,nodev`. Test: with the namespace folder a link to a sentinel directory,
the attach refuses and the sentinel's owner is unchanged; an assign with a mounted system volume's id is refused.
**S2-02 (High, known as part of R-984).** Missing: a rule that the new target is a raw drive mount the guest cannot
write, and a fixed storage id. Harm: the primary whole-guest backup silently moves where guest root can delete or
replace it. Fix in the root wrapper; consider a signature for moving the target. Test: the wrapper refuses a
guest-facing bind, a tmpfs and a network mount.
**S2-03 / S2-04 (High).** Missing: the allow-list call, a file-name list, an infrastructure-name refusal, and a
data-path rule on the import path. Harm: one import click runs what the box refuses from the catalog and writes outside
the app. Fix and test: in the R-1407 row.
**S2-05 (High).** Missing: escaping where page scripts insert job errors, and a script policy. Harm: a hacked app
crosses from one container to the whole dashboard when the household exports or imports it. Test: markup in a step
error renders as text; a template gate fails on an unescaped insert.
**S2-06 (High).** Missing: a rule on setting names, and an explicit file and project on every compose call. Harm: the
allow-list (R-971's fix) is passed by a catalog commit alone. Test: a reserved name is dropped at load; the argument
list carries the file.
**S2-07 / S2-08 (High).** Missing: a YAML encoder and a character rule for the override, the override in the judged
set, and `name` as a refused key. Harm: as S2-06, by two more routes. Test: an override built from a value with a line
break parses to exactly one `volumes` key; a compose with `name` is refused.
**S2-09 (High).** Missing: a check at start time, and a catalog rule against a bind inside another read-write bind.
Harm: a hacked download app reaches guest folders after its own restart. Test: the catalog gate refuses the nested
pair; a start path refuses a source that now resolves outside the app's roots.
**S2-10 (High, conditional).** Missing: the start hash is a secret-derived value in a public file. Harm: offline
guessing of the one password that opens the hub — only while that password is live. Fix: Secret, new password, gate.
Test: the manifest gate refuses a non-empty hash.
**S2-11 (High, a decision).** Not a missing check but an unstated cost: three unsigned controller-side channels let a
taken-over hub take dashboards and choose the app source. The signed-job guarantee of `01` §4 holds where it is
written (the agent). Options are in `STATUS.md` D90.
+23 -10
View File
@@ -117,7 +117,7 @@ stopping line that lies.
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
| **R-980** | Install & onboarding | P3 | **A person cannot check that the installer is ours: the ISO's first boot runs the install script from a movable tag with no hash or signature, and the ISO checksum sits on the same website as the download.** Fix: a signed manifest (or signature) checked by the ISO before it runs the script, and the public key printed in the guide. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN — design-proposal | — | Design note, then build | CC |
| **R-980** | Install & onboarding | P3 | **A person cannot check that the installer is ours: the ISO's first boot runs the install script from a movable tag with no hash or signature, and the ISO checksum sits on the same website as the download.** Fix: a signed manifest (or signature) checked by the ISO before it runs the script, and the public key printed in the guide. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN — design-proposal **2026-11-19 (security review 2, S2-29): still open as written in `felhom.eu@b7d8f75b`; the legacy fetch path for an agent below 0.143.0 has no hash at all.** | — | Design note, then build | CC |
| **R-250** | Install & onboarding | P3 | **A customer create can fail fail-closed because the host-key scan ladder is shorter than the DNS/AAAA settle time.** Found 2026-08-07 creating the fifth walk's venue. `POST /configs/new` with off-site enabled provisions the Storage Box sub-account and then scans its SSH host key to pin it — **fail-closed by design** (`offsite.go:111-121`, *"don't serve a descriptor the controller can't verify"*), with `defaultScanBackoff` = 2+4+8+16+30 ≈ **60 s**, sized by its own comment to *"the observed DNS propagation lag"*. **Measured, both halves:** the first create exhausted the ladder — five `no such host`, then `dial tcp [2a01:4f8:bacc:2:200::d30]:23: connect: network is unreachable` (the name had just begun resolving, **AAAA-first, into a pod with no IPv6 route**) — and the hub logged `[ERROR] offsite provision for walk5`. An **identical second POST succeeded** ~70 s later on its own final rung (`shared already provisioned … (subaccount 285351)`). **Total settle ≈ 100 s against a 60 s budget.** The endpoint was never the problem: verified afterwards from both the node and the hub pod, by name, `SSH-2.0-OpenSSH_9.6p1`. **Two distinct things are wrong and should not be merged:** (1) the budget is sized against DNS *existence*, but what actually bit is the **AAAA-before-A** window, a different and longer phenomenon — lengthen the ladder and/or prefer the A record for the scan dial; (2) **the operator is told nothing actionable** — the create fails with a generic error, the remedy is "press it again", and nothing says so. The retry is genuinely safe (idempotent on the label; confirmed afterwards — `one_time_secrets` 1, one sub-account, no double-provision) **but that safety is invisible to the person deciding whether pressing again will double-charge them.** **Severity LOW-MEDIUM:** self-clearing, no data at risk — but it is the first thing a new customer's provisioning does and it fails looking like an outage. | **VERIFY** — **2026-10-12: hub 0.147.0 live** (S29, decision 224): the host-key scan dials IPv4 first, waits ~2 min (was ~1), and when it still runs out says the store was created and it is safe to press Save again (`TestScannerDialsIPv4First`, `TestScanOutOfRetriesSaysSafeToPressAgain`, red-proved). Known limit: the hub Ingress reads for 60 s, so a very slow first try can show a gateway timeout while provisioning finishes in the background (pressing again is safe). **READY** — operator ruling 2026-10-12 (S29, `09` §3 decision 224): yes — the scan prefers IPv4, waits longer, and the error says „safe to press again". **READY** — owner Viktor | — | The next real off-site setup (a new customer): read the hub log for the scan; success or the new message → close | CC |
| **R-516** | Install & onboarding | P3 | **[P3-LOW] English a customer meets on a fresh box and its apps' first screens — enumerated by the big night.** MEASURED 2026-09-14 (BIGNIGHT, VM 333, ISO 1.27.1, controller 0.242.0). Felhom-owned: (1) the dashboard menu item **„Debug"**; (2) the dashboard CPU tile **„Load: 0.29 / 0.39 / 0.37"**; (3) the launcher tile **„Filebrowser"** opens a login in English with no Felhom text (R-513); (4) the storage page mixes formal „Adjon hozzá / Csatlakoztasson" with the product's „te". App first screens a household meets before any Felhom text helps: (5) **Uptime Kuma 2.4 opens on „Which database would you like to use?"** (SQLite / Embedded MariaDB, „Next") — the app card's „Első lépések" does not mention it; (6) PrivateBin, Gokapi, AdventureLog and FileBrowser UIs are English (the apps' own). Already rows: the Proxmox installer screens (R-495, answered by the guide), `wiki.DOMAIN` (R-498). **Fix shape:** rename „Debug"/„Load" (controller); add the Uptime Kuma database step to its card, or pre-seed `db-config.json` for SQLite in the template (catalog). **Added by F4 (20:04:54Z):** (7) the storage page prints the disconnect time as a raw ISO UTC string „Leválasztva: 2026-09-14T19:58:02Z"; (8) the „Meghajtó leválasztva" banner appears twice on every page; (9) „4 telepített alkalmazás nem fut — nézze meg a rendszermonitort" uses the formal form. **Added by F7 (20:50–21:00Z, system disk at 95 %):** (10) a banner on every page in English, „**SSD disk usage high: 90%**"; (11) the dashboard tile reads „Rendszer (/) 61.8 GB / 68.7 GB (**90%**)" while `df` reports **95 %** (reserved blocks ignored), and „(/)" labels the data volume `/mnt/sys_drive`; the deploy page says nothing about free disk. **Added by the i18n spike (2026-09-17, controller v0.247.0):** (12) **six formal („ön") forms in the converted dashboard copy** — „Olvassa be telefonnal" (launcher QR hint), „Biztosan kikapcsolja a megosztást?" (launcher), „Ha újratelepíti … importálnia kell" (the layout's remove-app modal), „Kérjük, vegye fel a kapcsolatot" (backups empty state). They were NOT fixed — a localisation release may not change Hungarian bytes — and are now COUNTED by `controller/scripts/i18n_missing_gate.py` (`HU_FORMAL_CEILING` = 6, a ratchet: a new one convicts, fixing one lowers it). The inventory (`audits/I18N-INVENTORY-2026-09-17.md`) is the list this row closes against in localisation slice 6 (R-561). **Extended 2026-09-17 by slice 1 (controller v0.248.0–v0.250.0):** the converted copy now counts **16** formal forms (`HU_FORMAL_CEILING` 6 → 10 → 12 → 16, each raise stated, Hungarian unchanged by rule) — and the count is an UNDER-count: the gate's stem list sees 4 on the release C pages (storage, drive wizards, debug), while a wider list („írja be”, „adja meg”, „adjon hozzá”, „válassza ki”, „biztosan eltávolítja”, „engedélyezze”, „hozzon létre”, „kattintson”) finds at least **22** keys there alone (login „Adja meg a jelszavát”, the NAS guide, the storage confirms). Widen the stems when this row is worked; the ceiling rises with them. **NARROWED 2026-09-20 by localisation slice 6's walk, item by item** (`audits/i18n-slice6-2026-09-20/R-516-item-by-item.md`): item 2 FIXED (slice 1), items 3 and 6 APP-OWNED (and for an English household item 6 is an advantage), item 11 is a wrong NUMBER rather than a language and needs its own row, item 1 is correct English for an English reader and still open for a Hungarian one. **Items 4, 7, 8, 9 and 10 are NOT DECIDABLE FROM AN ENGLISH WALK** - they are about Hungarian copy quality, a disconnected drive and a full disk, none of which this walk had. **What this row is now waiting for is a HUNGARIAN walk on a box with a second drive**, not another English one; saying it closed would be the kind of closure that makes a register stop meaning anything. **-- 2026-10-08:** the recovery screen's own answers are Go literals in Hungarian only — 14 `renderRecovery` messages in `controller/internal/web/recovery_handlers.go` (seen: the safe default „A művelet nem fejeződött be…" rendered on an English page in a red-proof run). Today's new R-304 sentence is a bundle key in both languages; the 14 old ones are not. | **NARROWED 2026-09-20 - rank P3-LOW; owner: CC; needs a Hungarian walk** **2026-10-05 (burn-down night): items 1 and 4 FIXED on controller `main`** (`d7efa7b`): the menu says „Hibakeresés” (English „Debug”); the storage pages use the te-form (formal-form ceiling 17 → 14; 116 parity fixtures changed by exactly those bytes). Item 2 was already fixed. **Left:** items 5–6 (apps' own screens) and the rest. **2026-10-06: items 1 and 4 DELIVERED** (controller v0.298.0). **2026-10-06 (burn-down night, later): items 7–10 and 12 FIXED on controller `main`** (`e4774e6`, `2e5ef7e`): the disconnect time in local time; one banner per unpluLine truncated
| **R-1273** | Install & onboarding | P3 | **An install that fails on the image registry's rate limit tells the household only „exit code 1”.** Catalog sweep 2026-10-24, code-server 16:01:37: `lscr.io … toomanyrequests` → the product cleaned up right (R-649) but the event and the card say „Alkalmazás telepítése nem sikerült: Code-Server — exit code 1”; nothing says the download server is busy and a later try works. The same install went through later (`audits/catalog-sweep-2026-10-24/code-server/`). | OPEN | — | Recognise `toomanyrequests` (and a pull timeout) in the deploy error and say „the app's download server is busy — try again in an hour”; a test with the stderr line | CC |
@@ -168,7 +168,7 @@ stopping line that lies.
| **R-683** | App updates | P3 | **[P3-LOW] Watch: after a power cut during an update's health check, the hold named an HOUR-OLD second-drive copy, not the one the update's own backup should have just made.** 2026-09-24 chaos round 3 (nextcloud, `backup_max_age: 1m`): no `backing-up` phase was seen and the hold named Tier 2 at 13:04 for an update pressed at 14:04; the pre-cut controller log was lost with the container (the runner now saves it at arm time — R-320). Round 11, the same action without a power cut, named a fresh 14:34 copy and logged the Tier-2 copy. The sentence was TRUE (it named the copy it offered); the question is why the update did not back up first. Not reproduced; watch the next power-cut drill. `audits/night-2026-09-24/E/round-03*.json`, `E/round-11-controller-pre.log` | **OPEN — P3; owner: CC (watch)** | — | — | CC |
| **R-785** | App updates | P3 | **[P3-LOW] SparkyFitness is pinned 11 releases and a major behind upstream (v0.17.3; upstream v1.7.3, v1.6.0 dated 2026-07-24).** READ 2026-10-01 (`audits/visitors-2026-10-01/C/bench/C1-previous-tag.txt`). **Needs:** an update walk 0.17 → 1.x through the ladder (bench + box), after R-784 is decided. **2026-10-27 (security triage):** two notices with NO fix released (GHSA-p8cf-g8rv-gvrq: a user overwrites another user's Withings link; GHSA-cmmf-53wf-r5gv: owner-only photos readable by URL) — both need a household account or a leaked URL; pick in `audits/security-app-updates-2026-10-27/PICKS.md` (CC's pick: leave it offered, take the fix with this big step). **2026-10-28 (R-1332): HTTP2Shell has NO fixed release yet** — v0.17.3's server carries Node.js 24.18.0 (`tsx index.ts`, behind the setup gate) and the newest v1.9.0 still Node.js 24.22.0 (fixed: 24.22.1); the web image is nginx, no Node. The big step waits for a v1.x with Node ≥ 24.22.1, or takes the runtime fix later as a lane step. | **OPEN** — **RULED D56 = keep it offered** (operator 2026-10-27 21:33, `09` §3 decision 272) — the fix rides its big step. **Earlier:** **OPEN — rank P3-LOW; owner: CC (after R-784)** **2026-11-02:** v1.10.0 and v1.11.0 read clean of HTTP2Shell (Node 24.22.1); v1.9.0 is still bad. **The big step v0.17.3 → v1.11.0 is proven on the drill:** bench, 9202 (peak 29.4 %), the undo (`audits/security-2026-11-02/bench/sparkyfitness/`, `walks/sparkyfitness-drill/`). Not written: R-784 (the licence) stands before it. | — | — | CC |
## Backup & restore — 49 rows (P2 4, P3 27, P4 18)
## Backup & restore — 52 rows (P2 4, P3 28, P4 20)
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
@@ -245,7 +245,7 @@ stopping line that lies.
| **R-1373** | Storage & devices | P3 | **Three endings have a second door that does less than the first.** Found 2026-11-10 (the endings list, `documentation/audits/endings-list-2026-11-10/TABLE.md` endings EC03, EC13, EA08; read in source, not measured on a box). (1) **Orphan delete** of an app (`internal/stacks/delete.go:305-460`, `internal/api/router.go:1150-1178`) removes the containers, the volumes and the stack record; it does not remove the undo copy, the update record or the backup folders that the normal remove takes away. (2) **A move of ONE app to another drive** (`internal/stacks/migrate.go:616-654`) removes the app's data and its backup unit from the old drive; the app's restore copy, a set-aside backup and kept data stay there. Only the move of all apps clears the old drive. (3) **A drive retired by a signed operator job** (agent `internal/signedjobs/decommission.go:86-103`) sets the record and does not take the drive out of the running guest, which the household's own door does (`internal/localapi/disks.go:559-567`). **Members (answers marked N):** EC03-4, EC13-4, EA08-4. | OPEN — filed 2026-11-10 (endings list) | — | Controller: make orphan delete call the same clean-up steps as the normal remove; make a one-app move take the app's other folders with it, or list them. Agent: say in `03-host-agent.md` whether the signed retire is meant to leave the drive in the guest. | CC |
| **R-1378** | Storage & devices | P3 | **Three endings can stop half-way with no finish, or have no way back that nobody chose.** Found 2026-11-10 (the endings list, `documentation/audits/endings-list-2026-11-10/TABLE.md` endings EC15, EC03, EC27; read in source, not measured). (1) **Wipe of a drive** (`internal/web/storage_handlers.go:911-1000`): the handler removes the drive from the registered places BEFORE the format. If the format fails, or the controller stops between the two steps, the drive is not registered and still holds the old data; there is no record of a wipe in progress, and a failed eject is only a warning. The handler does not look for apps on the drive. (2) **Orphan delete** writes no 'remove pending' record, so a controller that stops in the middle does not finish it at the next start (the normal remove does). (3) **Import over an installed app** (`internal/web/handler_export.go:312-356`, `internal/appexport/restore.go:288`) replaces the app's volumes and makes no undo copy first. **Members (answers marked N):** EC15-4, EC15-7, EC03-7, EC27-5. | OPEN — filed 2026-11-10 (endings list) | — | Controller: format first and unregister after, or write a wipe-in-progress record that the next start reads; give orphan delete the pending record; decide if an import makes an undo copy (a restore does). | CC |
## Security & access — 33 rows (P1 2, P2 10, P3 14, P4 7)
## Security & access — 40 rows (P1 2, P2 13, P3 18, P4 7)
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
@@ -263,7 +263,7 @@ stopping line that lies.
| **R-1141** | Security & access | P3 | **`felhom-host-install --rotate-recovery` exists only as part of a full installer re-run, which also reinstalls the vouched agent and the root files.** On a box running a newer agent than the vouched one (demo-hp on the candidate, 2026-10-17) it would move the agent back; on a household box it is a much larger act than a password change. The 2026-10-17 rotation therefore ran the installer's own `step_break_glass` ALONE (`audits/agent-allowlist-rotation-2026-10-17/P3/`). | **NARROWED 2026-10-18 — built on `main`, not published.** `felhom-host-install.sh --rotate-recovery-only` runs step 4b alone (felhom.eu `7d64038f`): reads the hub, host id and host key from the agent config, refuses byo and every other mode flag, asks the hub's `/healthz` before the password changes, never touches the agent or root files. The default install path is unchanged in behaviour (harness verdicts and 9 transcripts identical old vs new; demo-hp `--preflight-only` identical; `--rotate-recovery-only --dry-run` on demo-hp changed nothing). `step_break_glass` moved, byte-identical, ahead of the argument parse. `break-glass.md` §5 points at it. Publishes only with the next `installer-v…` tag. | the next installer tag | Cut the next `installer-v…` tag (with the other untagged installer changes); close | CC |
| **R-1258** | Security & access | P3 | **ep0's iperf3 line-test server takes any tunnel peer's test: one box (root on its host) can keep it busy for the fleet or ask a long reverse test that fills ep0's uplink.** Found 2026-10-24 (SEC-2 X2): `felhom.eu/scripts/ep0-iperf3/felhom-iperf3.service:17` — no authorized users, no server-side duration or rate cap; one client at a time (`felhom-agent/internal/linetest/iperf.go:263-264` → `busy`; 6 tries a day). Static (no ep0 change in this run). | OPEN | — | per-box iperf3 users or server-side caps, tried on a scratch VM first | CC |
| **R-1307** | Security & access | P3 | **Packages inside our own images are not watched** — the Debian packages of the controller image, the Alpine packages of the hub and Samba images (Samba's version is whatever Alpine 3.24 ships). The security watch (R-1303) reads only the base image's point release (BEHIND); Debian's DSAs and Alpine's secdb are not read. `audits/security-watch-2026-10-26/SPIKE.md` „cannot see" | OPEN — filed 2026-10-26 | — | add Alpine secdb (machine-readable JSON per release) and Debian's tracker for the packages the images install, or a scanner run on the built images in CI | CC |
| **R-955** | Security & access | P4 | **The cross-box check of the hub API is done in source and tests but not fully proven: if a later check masks a missing id check, a box's key could read or change another box's hub record without any test noticing.** Read 2026-10-12: every box-facing `/api/v1` route takes the box from its own key or refuses a different id; 14 routes already had a refusal test, 5 got one (`crossbox_refusal_test.go`). Left: red-prove the 4 new tests that were not red-proved (as the restore-directive one was), and the live read-only check with one demo box's key against the other demo box's id. | OPEN — filed 2026-10-12 by the security review | — | Red-prove the 4 tests; run the live GET check on the two demo boxes | CC |
| **R-955** | Security & access | P4 | **The cross-box check of the hub API is done in source and tests but not fully proven: if a later check masks a missing id check, a box's key could read or change another box's hub record without any test noticing.** Read 2026-10-12: every box-facing `/api/v1` route takes the box from its own key or refuses a different id; 14 routes already had a refusal test, 5 got one (`crossbox_refusal_test.go`). Left: red-prove the 4 new tests that were not red-proved (as the restore-directive one was), and the live read-only check with one demo box's key against the other demo box's id. | OPEN — filed 2026-10-12 by the security review **2026-11-19 (security review 2, S2-26): a blind re-read of all 54 API routes found no route where one box's key reads or changes another box's record; the route tests still pin the switch and not what each handler checks.** | — | Red-prove the 4 tests; run the live GET check on the two demo boxes | CC |
| **R-924** | Security & access | P2 | **The operator signing keys exist only on DooPlex, with no passphrase and in no backup — a loss of DooPlex means no box ever accepts a signed agent update, config bundle or OS step again without an on-site re-enrollment.** MEASURED 2026-10-09: `/mnt/5_hdd/felhom.eu/felhom-op-operational`, `felhom-rec-recovery`, `felhom_op_ed25519` (0600 `kisfenyo`); `ssh-keygen -y -P ''` opens all three (no passphrase); `/mnt/5_hdd/felhom.eu` is in no backup set, the off-site job does not carry them. `04` §1 says the operational key is "passphrase-protected" and the recovery key "kept offline" — today neither holds (custody ruling 1 of 2026-09-16 put both on DooPlex for this phase; `04` §3.1). Both keys share one fate, so the cold key cannot do its one job (re-pin after losing the operational key). `runbooks/total-loss-of-dooplex.md` S6. **2026-11-11:** the End-dates list leans on this row (`documentation/audits/end-dates-list-2026-11-11/TABLE.md` DB12). | NARROWED — **2026-10-23 (D32):** S6a/S6b are on the current sheet. LEFT: the operator prints it. **NARROWED 2026-10-09 — operator ruling: option A** (print both AND copy them). The three key files and their public halves ride the nightly encrypted DooPlex→ep0 copy (`scripts/dooplex-offsite/`, a missing key refuses the push); the Sunday restore test proves each private key still derives its own public half (never printed). **LEFT:** the operator prints S6a/S6b (`runbooks/break-glass-sheet.md`); the keys still have no passphrase (custody ruling 1 of 2026-09-16 stands until the first sale). **2026-10-23 night — survey (`audits/dooplex-survival-2026-10-22/PLAN.md`, G2):** the three key files are on `sda1`, the same disk as DooPlex's local backup tree; their only off-box copy is the S1-encrypted nightly one. Passphrase: unchanged by the custody ruling. | operator: print | Operator: print the two keys; then close | operator |
| **R-132** | Security & access | P3 | **`curl -w '%{redirect_url}'` reconstructs the request URL WITH its basic-auth credential** — so a `-u ":$HUB_PW"` call that never put the password in a URL still printed it **Merged 2026-10-05 from R-350 (duplicate):** (1) 2026-08-20 occurrence: POST /configuration/artifacts answers 303; leak lives only in the CC transcript under ~/.claude/projects/ on DooPlex, not in git/evidence (checked then). (2) `-v` and `--libcurl` also re-render the credential, not only %{redirect_url}; confirm redirects with %{http_code} + follow-up GET. (3) Rotation path: hub /configuration form (current_password/new_password/confirm_password); DB override wins over ConfigMap (break-glass); CC can rotate file-to-file without printing (operator-present-one-time-secrets) if asked. | **NARROWED 2026-10-20 — the rotation is DONE:** the operator changed the hub password on 2026-10-15 and the `HUB_PW` line is gone (`audits/hub-tokens-2026-10-15/C153-live.txt:18`; CC uses its own token since hub 0.152.0). **Left, the operator's:** should the break-glass `password_hash` in `manifests/hub.yaml` (a public repository) be emptied or replaced? The DB override wins at login, so the old hash no longer opens the hub. **WAITING-ON-OPERATOR** — **ACTION: rotate `HUB_PW`** **Folded R-580 2026-10-03** (the same `curl -w %{redirect_url}` credential echo, seen again 2026-09-18). **Note 2026-10-12 (security review):** `manifests/hub.yaml` also carries the hub login's bcrypt hash in this public repository (the break-glass field); when `HUB_PW` is rotated, decide whether that field should stay empty. | the operator's word on the `manifests/hub.yaml` hash field | Happened on 2026-07-31 while red-proofing the R-120 gate: the hub operator password was written to the session transcript by the write-out format, not by the request. `-u` is safe; the *reporting* was not. Rule: read the redirect from `-D -` and grep `^Location:`, never `%{redirect_url}`, on any authenticated call. Rotate the hub password (`/configuration` → Login password; ConfigMap `auth.password_hash` is the reset path) and update `~/.config/credentials` | Viktor |
| **R-137** | Security & access | P3 | **Cloudflare geo-WAF rules are zone-scoped and non-namespaced — four cross-tenant faults.** `globalRuleDesc = "[felhom-geo] Global"` (`waf.go:18`) is one literal description per ZONE; `appRuleDescPrefix` keys by app name with no customer (`waf.go:21`); `BuildGlobalExpression` has no positive hostname scoping (`waf.go:241`); `applyDiff` deletes every `[felhom-geo]` rule not in THIS box's desired set (`geosync.go:320`) | READY (M) — **blocks shared-zone onboarding** | — | With two customers in one zone: they overwrite each other's Global rule forever; one customer's country policy applies zone-wide; per-app rules collide by name; and disabling the feature for one (or the hub's `RemoveGeoRules`) wipes them all. Interim mitigation, no code: keep geo-restriction OFF for every shared-zone customer. Fix = namespace descriptions by `customer_id` + add `http.host ends_with "<domain>"` to both expressions — a TWO-REPO change (controller + hub `RemoveGeoRules`). Same audit §5.1 | CC |
@@ -273,16 +273,24 @@ stopping line that lies.
| **R-925** | Security & access | P1 | **`manifests/felhom.secret.yaml` carries REAL secret values, and the Gitea repo that holds it is readable by anyone on the internet with no login.** FOUND 2026-10-09 by a background security review of the legal-pages commit, which flagged the served `<!-- source -->` comments; chasing what those comments pointed at found this instead. **MEASURED, in this order:** (1) `scripts/manifest_bearer_gate.py` itself prints `manifests/felhom.secret.yaml:39 KNOWN-BACKLOG committed secret 65cee3c4…7a86`; (2) the file holds **live-shaped values, not placeholders** — `SECRET_KEY` (69 chars), `SUPERUSER_EMAIL`/`SUPERUSER_PASSWORD` (18), Umami `APP_SECRET` (66) and `POSTGRES_PASSWORD` (34), and a `username`/`password` pair (18); values were never printed, only measured by length; (3) `gitea.dooplex.hu` resolves to **37.191.56.193**, the same public address as the website; (4) an **anonymous** `curl` returns HTTP 200 and 1686 bytes for that exact file, and 200 for `hub/internal/store/store.go` and `manifests/hub.yaml`; (5) **the off-network control**, which is what makes (4) mean anything — fetched from outside the operator's network entirely: a real path returns the file's first line, a nonsense path returns 404, so the 200 is genuine anonymous read from the internet and not a LAN-only ACL. **This contradicts the project's own written model**: `documentation/runbooks/secrets.md:3-4` says *“Secret values are never committed to git. The manifests in `manifests/` carry only placeholders + comments that point here.”* That promise is false today, which is the P1 wording in this register's own scale — see the ranking note below. The same runbook already states the rule that governs the fix: *“The git-history copy stays alive until the value is ROTATED — de-git alone kills nothing.”* **Known neighbours, none of which cover this:** R-887 records that an outside crawler walks the public Gitea pages and leaves *“Gitea's exposure to the crawler”* as the operator's undone item; R-580 still owes a Gitea admin token rotation. Neither says a secret file is world-readable. **Why P2 and not P1, stated so the operator can overrule it:** the exposed credentials guard **analytics and an undeployed healthchecks instance, not household data** — measured: `umami-db` is a **ClusterIP** service on 5432 with no external IP, so the Postgres password is not reachable from the internet; the realistic harm is session forgery against the public `stats.felhom.eu` via `APP_SECRET`, and reuse of those passwords anywhere else. No customer box, hub token or escrow key is in this file. **CC did NOT change anything**: making the repo private could break the public day-0 path (the installer is fetched from a public tag, R-110), and rotation plus repository visibility are operator decisions on production infrastructure. **PARTLY FIXED 2026-10-09, and RE-RANKED P2 -> P1 on what it turned out to be.** The exposed value was not an analytics password: it was the **Gitea `admin` account password** (`is_admin: true`; `/api/v1/admin/users` answered 200), published in a repo `gitea.dooplex.hu` serves anonymously to the internet. That is push access to every repo — including the one whose `website/` is git-synced live and whose `scripts/` is published by tag to **every new box installer** (R-110), i.e. a supply-chain path onto customer hardware. My first ranking said P2 because I had only measured the analytics blast radius; the credential test is what corrected it. **DONE (CC, operator-instructed):** (1) `umami-config` `APP_SECRET` + `POSTGRES_PASSWORD` rotated — the password changed **inside Postgres** too (`ALTER USER`), because the env var is only read at first init; verified by a real beacon returning 200 with a bogus-site-id control still returning 400. (2) `healthchecks-config` `SECRET_KEY` + `SUPERUSER_PASSWORD` rotated (nothing consumes them — there is no healthchecks Deployment). (3) `gitea-creds` no longer holds the admin password at all: it holds a **scoped token** (`read:package` + `read:repository`). **Both scopes are load-bearing and the second was found by breaking it** — a package-only token made the hub log `Template fetch: unexpected status 403`, because the template fetcher reads a raw file out of the `felhom-controller` repo. (4) The Gitea **admin password** was changed and `gitea-system/gitea-admin` updated; **verified: new password 200, old published password 401 — the leak is dead.** (5) The file is removed from git and `.gitignore`'s `*secret*` now applies to it; `manifest_bearer_gate.py`'s `KNOWN_BACKLOG` carve-out is gone, red-proofed with a decoy (exit 1 with, 0 without). **AN INCIDENT CAUSED BY THE FIX, recorded because it is the useful part:** the `rollout restart` needed to pick up the new umami secret put umami into **CrashLoopBackOff and took `stats.felhom.eu` down (503)**. Cause was not the rotation — at `memory: 512Mi` the pod runs for months but **cannot restarLine truncated
| **R-1243** | Security & access | P2 | **The hub's Longhorn backups sit unencrypted in the NFS folder any DooPlex pod can read (R-1241) — the two there now (2026-10-18, 2026-10-23) hold every household's Cloudflare and tunnel tokens in plaintext, because they predate hub 0.168.0.** Measured 2026-10-23 night (`audits/hub-token-seal-2026-10-23/` §3): the hub volume is `encrypted=false` and in Longhorn's `default` backup group (daily 04:00 + weekly Sunday 05:00, retain 1 each); the hub's own copy to ep0 is encrypted at rest (16/16 snapshots `crypt-mode: encrypt`, key on DooPlex's root disk and on paper) and in transit (SSH tunnel + pinned PBS TLS), every 6 h, restore-tested weekly. DECIDE: (a) an encrypted Longhorn volume for the hub (~5–10 min downtime, a homelab StorageClass + key in the cluster) or (b) take the hub volume out of the Longhorn backup — ep0's copy is the backup; the access log then has no backup until it rides the ep0 push. **CC's pick: (b).** And: delete the two plain copies (the operator's call; with (b) they never rotate out). | **VERIFY** — **2026-10-24: D40 put into effect** (operator ruling 07:22): the hub PVC `hub-data` carries `recurring-job-group.longhorn.io/default: disabled` (felhom.eu `3363ffa8`, ArgoCD sync of the PVC only) and the Longhorn volume `pvc-486c9809…` reads `disabled` too (labelled live to match, unchanged after 2 min); the NFS folder holds no copy of the volume (no `pvc-486c9809*` directory, no backup config naming it; Longhorn lists no backup volume for it; 53 other volumes backed up this morning). Closes after the dated check: tonight's 04:00 job makes no new hub backup while the others run. WAITING-ON-OPERATOR **2026-10-25 dated check DONE (read-only, the restore-truth session):** `kubectl -n longhorn-system get backups.longhorn.io` lists 106 backups made today (the newest 04:08:28Z — the job ran) and none of `pvc-486c9809` — the hub volume is out of the Longhorn backup as D40 meant. | the operator's pick; a homelab change for either option | Operator: pick (a) or (b), and say whether to delete the two plain copies after hub 0.168.0 is live | CC (the dated check) |
| **R-1355** | Security & access | P2 | **DooPlex's own Gitea 1.26.2 is inside 19 new notices (2026-11-02, fixed 28.3.0: `GHSA-8p3x` OAuth2 account takeover, `GHSA-37xv` and `GHSA-8g99` public-only tokens reach private data, `GHSA-32h8` Actions secrets) — on the host that holds our code and registry.** Also new on DooPlex: Vaultwarden 1.37.4 (fix 1.37.5), Uptime Kuma 2.5.6 (fix 2.5.9), Paperless 3.3.4 (fix 3.4.0). Felhom changes nothing on DooPlex. Read in `documentation/security-watch/state.json`, group DooPlex. | OPEN — filed 2026-11-03 | the homelab session | add them to `audits/http2shell-2026-10-27/HOMELAB-LIST.md`'s work; which notices reach a self-hosted Gitea with our settings is the homelab session's first read | operator (homelab session) |
| **R-1420** | Security & access | P2 | **Four more routes lead past the box's privilege allow-list from a catalog commit alone: the media-library override file is built as text from catalog values and never judged; the compose file's own project `name` is not judged, so one app can address another stack's containers (our traefik holds the Docker socket); `networks` is not judged; traefik labels are not judged.** Harm: whoever can push to the catalog (a taken-over Gitea) can still run what R-971 was closed to refuse, or answer for the dashboard's name, on every box with the app. Static; the project-name behaviour measured on scratch guest 9202 with throwaway containers (2026-11-19); the host network by the `networks` route was refused by Docker (measured). Second read: holds (override, name, networks); labels: self. Fix: build the override with a YAML encoder and judge the merged result; refuse `name`; name the project on every call (with R-1409); allow only the one external network; judge router rules. Tests in the findings page. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-07, S2-08, S2-14, S2-15; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity High | — | Fix task (controller release), with R-1409 | CC |
| **R-1421** | Security & access | P2 | **A folder an app receives is judged when its definition is written, not when its container starts — and the Transmission template gives the app the right to swap one of its own folders for a link.** A hacked Transmission can therefore, after its own restart, read and write a guest folder the allow-list would refuse (the family folders the download apps are kept out of; more for a root process). No catalog push is needed. The same shape exists between apps: Sonarr and Radarr hold the media folder read-write while 12 templates receive a child of it. Static; that Docker resolves the folder again at a plain restart was measured on scratch guest 9202 (2026-11-19). Second read: holds. Fix: the catalog gate and the box refuse a folder inside another read-write folder of the same app; change the Transmission template; longer term, judge again before each start. Test: the gate refuses the nested pair; a start refuses a source that now resolves outside the app's roots. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-09; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity High | — | Fix task (catalog + controller) | CC |
| **R-1422** | Security & access | P2 | **Decide (D90): a taken-over hub can take over every household's dashboard and choose what apps a box runs, with no operator signature.** Three unsigned channels to the controller: the hub can give an already-claimed box a new reset code (who knows the code sets the dashboard password on the public claim page); the hub writes the box's settings file, which names the catalog's address, the tunnel and the data folder; and the hub can place files that the box serves in the dashboard's own origin before sign-in. The signed-job rule of `01` §4 holds where it is written (the agent: destroy, wipe, restore-overwrite); the documents do not say the wider thing, and `13` §6 does not accept this. Not reachable from a read-only copy of the hub's database. Static; second read: holds (all three). Options and costs: `STATUS.md` D90. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-11; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | WAITING-ON-OPERATOR — filed 2026-11-19, technical severity High; design | operator (D90) | Operator decides D90; then a design note | CC |
| **R-779** | Security & access | P4 | **[P3-LOW] Part A's "two outside addresses seen as two" is proven through the simulated tunnel only; on the REAL tunnel the second outside address (ep0, one request allowed) was refused by Cloudflare's edge with 403 and never reached the box.** Measured 2026-10-01 19:51 UTC (`audits/visitors-2026-10-01/A/L2-demo-hp-real-tunnel.txt`): no log line on demo-hp; demo-hp's box has no geo restriction in its settings, so a Cloudflare ZONE rule (country or bot, not read) refused a German datacenter address. DooPlex's own address on the real tunnel was seen as itself. **Needs:** one sign-in from a second Hungarian address (the operator's phone off wifi) while DooPlex is locked out — 2 minutes; and say which Cloudflare rule refused ep0. | **WAITING-ON-OPERATOR — rank P3-LOW; owner: operator (a phone), CC reads the logs** **Re-ranked 2026-10-03: P3→P4: a proof gap on the real tunnel; operator-only follow-up.** | — | — | CC + operator |
| **R-904** | Security & access | P4 | **Cloudflare can read every household's app traffic; replacing it with our own relay is a later item.** Facts (reviewer discussion 2026-10-08; `01`): app traffic and the dashboard reach the box through the Cloudflare Tunnel (`01` §5 trust table, rows end-user ↔ apps and customer ↔ controller UI; §7), and the tunnel's public end is Cloudflare's edge, where TLS ends — so Cloudflare can technically read that traffic (the FAQ says so since 2026-10-08, R-900; „TLS ends at the edge" is not written in `01` — add it there). What Cloudflare gives today, free: inbound reach with no router setup, the CGNAT answer (`01` §4, §7); certificates (`01` §7, the free tier covers one level below a zone); the geo-WAF the hub enforces (`01` §5 last row, §7); flood protection (not in `01`). The alternative named: our own EU relay over WireGuard with TLS passthrough by SNI, certificates on the box, the geo-block on the relay. Its costs: one more machine the operator keeps up, and a single point of reach for every box; weaker flood protection; about a week of work after a spike. Operator ruling 2026-10-08 09:07 (`09` §3 decision 184): a later item. | **DEFERRED** — **2026-10-10: the asked sentence is in `01` §7** („TLS ends at Cloudflare's edge …”); the spike stays deferred. **DEFERRED — after the first customers (operator ruling 2026-10-08 09:07)** | — | A spike after the first customers (the relay's reach, cost and flood behaviour, measured) | operator |
| **R-1040** | Security & access | P4 | **Until the household makes Sonarr's or Radarr's first admin, any container on the box can read the app's API key.** Measured on the bench 2026-10-14 with the R-1030 seed in place: `AuthenticationRequired=Enabled`, `AuthenticationMethod=None` — `/initialize.json` from another container answers 200 with `apiKey`. Outsiders are blocked by the box's setup gate in that window; a compromised other app is not. | **OPEN** — filed 2026-10-14 | — | Seed the API key at install (`SONARR__AUTH__APIKEY`, measured to work) so the box knows it, and close the window by telling the household in the first steps; or seed a Forms user | CC |
| **R-1054** | Security & access | P3 | **The hub's operator/global API key is still one bearer that can do everything on `/api/v1`, including reading every box's break-glass root password (`GET /admin/hosts/{id}/recovery-credential`).** Hub v0.152.0 gave the web surface scoped tokens (`05` §17) but tokens are refused on `/api/v1` by design. Who uses the global key today is not listed in one place. Fix: list its consumers; move each operator use to a scoped token on the web side (or a token class on the API), and keep the global key only where a box path needs it — or retire it. | OPEN — filed 2026-10-15 by the token build | — | List the global key's consumers (pod env, scripts, runbooks) | CC |
| **R-1054** | Security & access | P3 | **The hub's operator/global API key is still one bearer that can do everything on `/api/v1`, including reading every box's break-glass root password (`GET /admin/hosts/{id}/recovery-credential`).** Hub v0.152.0 gave the web surface scoped tokens (`05` §17) but tokens are refused on `/api/v1` by design. Who uses the global key today is not listed in one place. Fix: list its consumers; move each operator use to a scoped token on the web side (or a token class on the API), and keep the global key only where a box path needs it — or retire it. | OPEN — filed 2026-10-15 by the token build **2026-11-19 (security review 2, S2-26): the API read of a break-glass password leaves a log line only, no event; the global key is also the key that can queue the job of R-1423.** | — | List the global key's consumers (pod env, scripts, runbooks) | CC |
| **R-1194** | Security & access | P3 | **On a fresh box the operator's break-glass SSH door was broken from the first minute: `oob_degraded` „felhom-sshd config invalid (sshd -t fails)".** Rehearsal 2026-10-20, Tester4-816e21, mail at 12:03 UTC (one minute after the host record was created): details `active:false, config_invalid:true, felhom_sshd_port:0, operator_key_configured:false, operator_key_reported:true, reachable:false`. Possibly the shape of a customer with no operator key set (Tester4 is a test customer) — then the alarm is a false one on every such box; or a real config defect. Not diagnosed (no shell). Checked the same day: the only `oob_degraded` in the operator mailbox in 30 days, and no recovery mail followed; the report says port 0. Two candidate mechanisms in the agent (`internal/felhomsshd/health.go:63-66`): `sshd -t -f <conf>` run by the NON-root agent fails on unreadable host keys, or a config rendered before the desired state arrived (port 0). **CAUSE FOUND AND FIXED 2026-10-21 night** (`audits/night-2026-10-21/P5-checks/`): reproduced on a fresh box from the public ISO — `oob_degraded` mailed 20:34 CEST while felhom-sshd was applied 0.4 s after the check and is healthy (port 8822 listening, `sshd -t` passes as the agent's user through its sudo grant — so the non-root theory is FALSE; without sudo it fails on the host key, which the agent never does). The first host report runs before the felhom-sshd loop's first pass and reads a config never rendered. Fix: agent `0da792e` — no OOB stanza until the first pass has run (the hub does not evaluate a missing one; a failed pass counts). `TestR1194_NoOOBStanzaBeforeTheFirstPass` red-proved. | WATCHING — released in agent v0.162.0 (2026-10-22, operator yes; on the 3 test boxes); the false alarm comes only on a FRESH box's first report | the next fresh install with agent ≥ 0.162.0 | At the next fresh install read the hub's events for `oob_degraded` in the first hour (none expected); then close | CC |
| **R-1237** | Security & access | P3 | **The nightly CC-secrets sweep found stored one-time codes in a session record dated 2026-10-23: 14 names in one file (bind tokens and setup codes of test boxes — NIGHT_2026_10_23_*, FIXES_2026_10_21_*, NIGHT_2026_10_21_*, REHEARSAL_2026_10_20_* — and R_DEMO_FELHOM in a file of 2026-10-20).** Mail „[DooPlex CC secrets] 2 secret name(s) found in old session records", 2026-10-23 01:57 UTC, read in the day session (names only). One file holding all of them suggests one command printed many stored values at once (a listing or a dump) in the night session. All but R_DEMO_FELHOM are one-time install codes of test boxes; NIGHT_2026_10_23_* die when Tester4-16fdef is deleted (the operator's do list). | **OPEN** — 2026-10-27 (operator yes): the dead names were REMOVED from the store with `sec rm` (46, one commit each; `audits/r1299-window-button-2026-10-27/sec-rm.txt`, names only): every one-time code this row names is gone; R_DEMO_FELHOM stays (unsure). Consequence: the nightly sweep no longer watches for the removed values — they open nothing (their boxes are destroyed). The session transcripts themselves are unchanged — that question stays here. **Earlier:** OPEN | — | CC: find the file and the command by the sweep's own report (names only), add a hook shape for that command if it is not one; R_DEMO_FELHOM (demo-felhom's recovery code) — the operator decides whether to rotate it | CC |
| **R-1248** | Security & access | P3 | **An app's own numbers can lie: an attacker with code inside an app's container can answer the app's API with whatever matches its flood, and the upload watch (R-1244) would then call it explained.** Measured remedy, not built: a per-container upload ceiling with `tc` from a helper container in the app's network namespace — on scratch 9202 (unprivileged) a flood was held at 19.9–20.1 Mbit/s and the app kept answering (`audits/SPIKE-upload-watch-2026-10-23.md` §5). Costs: a helper image with `tc` (a new external image, or `tc` in the controller image), `NET_ADMIN` for that helper, re-applied after each app restart. | OPEN | an operator decision (the new image / capability) | Decide; if yes, CC builds the ceiling for the download and media apps first | operator |
| **R-1423** | Security & access | P3 | **A job with no valid signature but a harmless operation name makes the agent crash, and the job stays queued, so the agent crashes again at every start.** Whoever controls the hub's job queue (a taken-over hub, or the holder of the global key, R-1054) can stop a box's agent until someone removes the job AND restarts the unit at the box: no whole-guest backups, no controller supervisor, no signed update to fix it. Nothing runs unsigned and no data is lost. Older than October's commit. Static; own second read. Fix: the runner refuses every job that is not a verified destructive class, and recovers from a panic. Test: one job per harmless class with a junk signature — no crash, executor not called, job cleared (red today). Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-16; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Fix task (agent release) | CC |
| **R-1424** | Security & access | P3 | **A box can make the hub send mail that truly comes from Felhom's own address to an address the box chose: the old `/notify` route has no type list, cooldown or limit; `/mail` and the box's own mail relay check only the sender's domain, not the name before the @, and cap messages, not recipients.** A taken-over controller — or, for the relay, any app container on the box — can send convincing mail as our monitoring sender, and can burn the shared mail quota that alarms and claim codes need. Static; own second read. Fix: retire `/notify` or give it `/event`'s list and cooldown; pin the sender name per box and refuse the hub's own names; cap recipients and mails per household per day; bind the relay's sender to the app. Test: 50 posts give one send; the hub's own sender name is refused. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-17; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Fix task (hub + controller) | CC |
| **R-1425** | Security & access | P3 | **Three hub hand-over doors accept the wrong caller: re-enroll, while recovery mode is armed, needs no proof and can be repeated; a second appliance register with the same hardware ids replaces the first one's token and clears the operator's bind; and a host key can replace its sealed recovery package in place, with no kept copy and no event.** Harm: in an armed window a stranger who knows a host id becomes that host; on a LAN a second device can receive a new box's first hand-over; a taken-over agent can leave the household's written recovery code opening nothing (the hub half of R-978, which was closed on the agent side). Static; own second read. Fix: a one-time token from the arming step, single use, with an event; keep the old appliance token unless it is presented; keep a copy whenever the stored bytes differ. Tests in the private notes. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-18; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Fix task (hub release) | CC |
| **R-1427** | Security & access | P3 | **Three dashboard gaps of medium weight: a share's folder path is written unquoted into the Samba settings; „add storage path” accepts the folder that holds every app's definition; and a stranger can keep the claim and reset page shut for everyone (five wrong codes lock it for all, again and again) and make the hub mail the household and replace a code being typed.** Harm: a folder name (an app can create one) adds Samba settings; app definitions become editable from the LAN outside the allow-list; a household cannot claim or reset while a stranger keeps trying. Static; own second read. Fix: a character rule and quoting for the share path; only registered drives as storage paths; a per-address lock or a much higher shared one, and no replacement of a code younger than a few minutes. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-21, S2-22; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Fix task (controller release) | CC |
| **R-1309** | Security & access | P4 | **What no automated source shows: a fix announced only by mail, forum or a blog we do not read; a security fix nobody calls one (cloudflared 2026.10.2's „bump distroless image"); closed-source apps (Plex, Emby).** The watch (R-1303) shows these only as BEHIND (weekly). The vendor mails the operator receives stay a source. `audits/security-watch-2026-10-26/SPIKE.md` | OPEN — filed 2026-10-26 (a stated limit) | — | none until a vendor offers a feed; add a vendor's feed to `security_watch.FEEDS` when one exists | CC |
| **R-1431** | Security & access | P4 | **Dashboard hardening, low risk each (security review 2): the claim code's wrong-try count comes after the slow compare; the sign-in limit is per address only and its tables are never swept; the current-password check on a password change has no limit; the session cookie can be shadowed from an app's address; a notice text is taken from the address bar (escaped); the install debug log prints field values; a bundle's unpack has no size bound; the userdata folder set-up follows a link; no test pins the list of routes that need no session; an old compose file in the repository still publishes port 8080; the first-login reader can be pointed at any container; the catalog sync follows a link in the repository.** Fix each in its own small change. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-27; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19 | — | Batch into hardening work | CC |
## Box system & updates — 20 rows (P2 1, P3 14, P4 5)
## Box system & updates — 23 rows (P2 1, P3 16, P4 6)
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
@@ -294,8 +302,8 @@ stopping line that lies.
| **R-1370** | Box system & updates | P3 | **The host's system disk has no space budget: no document says how much the host's own files may take, and the installer's sizing rule covers the guest only.** Found 2026-11-10 (the space list, `documentation/audits/space-list-2026-11-10/TABLE.md` place H1, answer H1-3; search S5 — 0 hits). Today: demo-hp 58 % of 41 GB, N100 38 % of 101 GB, Tester 1 34 % of 63 GB — and the parts that grow are the journal (R-1367), old kernels and agent binaries (R-1368), old backup copies (R-1087) and the local whole-guest copy (R-548). Second half, known but with no row until now: on a small disk the guest's own guards read the guest's 70 GiB while the thin pool behind it holds 54 GiB (demo-hp: 102 GiB promised on a 53.9 GiB pool, 70.67 % used), so only the hub's thin-pool alarm stands before a full pool. **Member:** H1-3. | NARROWED — **2026-11-15 (build of release 4): BUILT, not released** — installer v1.38.0 on `main`, NOT tagged (`16` §14.5 `[FACT — built, not released]`): `check_system_disk` first in the preflight; an appliance below 200 × 10⁹ bytes stops; `--accept-small-disk` goes on, prints the cost, records `small_disk`; a BYO host is warned only; an unreadable size stops and names the flag. 8 tests, 15 mutations seen red. **Not run on a machine.** What it does not do is R-1398. For the release: the proof VM gets a 200G disk; demo-hp's reinstall needs the flag. Earlier: OPEN — **2026-11-15 12:49 — RULED (D83 = refuse, with one named flag; `09` §3 decision 308).** Fix shape: the installer reads the size of the disk that holds the `pve` volume group before any change; below 200 × 10⁹ bytes an appliance install stops with the designed sentence; `--accept-small-disk` goes on, prints the cost, records `small_disk: true`; a BYO host is warned only; a size that cannot be read refuses on an appliance and names the flag (`16` §14.5). Earlier: OPEN — **2026-11-14 (night build):** the open point of D75 has a written design with a pick, not built: `16-storage-physical-identity.md` §14.5 — refuse below 200 GB on an appliance install, with one named flag for the operator; warn only on a BYO host; 8 edge cases. It waits for the operator's yes. OPEN — filed 2026-11-10 (space list) **2026-11-10 12:22 — RULED (D69 = yes; `09` §3 decision 288):** the section goes inside `16-storage-physical-identity.md`, no new document. **2026-11-10 (afternoon) — the section is written:** `16-storage-physical-identity.md` §14 (45 places, the list „no limit, by ruling”, the budget, the open design points); `PROMPT-TEMPLATE.md` §N.5 keeps it true. The budget from measured numbers: the root needs about 14 GB when the whole-box copies go to a backup drive, and 28–54 GB at keep 3 when they do not. **D75 (for the operator):** the smallest system disk Felhom accepts. CC's pick: 128 GB only with a backup drive, 200 GB without one. The other option, 128 GB always, lets a box with many apps and no backup drive stop its local copies. With no ruling the installer accepts any disk. Open after the ruling: the installer's refusal. **2026-11-10 15:19 — RULED (D75 = one number, 200 GB, with or without a backup drive; `09` §3 decision 295).** This replaces CC's two-step pick. **Open, not ruled:** what the installer does below 200 GB — refuse, or warn. demo-hp (128 GB) stays as a test box. `16` §14.3 carries the ruling. | — | Design first: what the installer does below 200 GB (refuse, or warn) — a later session. Then the installer. | CC |
| **R-1398** | Box system & updates | P3 | **A machine that the installer refuses for a small system disk tells nobody but its own screen, and the hub does not read the „small disk” mark.** Found 2026-11-15 while building D83 (installer v1.38.0, not tagged). (1) On the public image the first-boot unit (`scripts/iso/felhom-bootstrap.sh`, „retry until success”) runs the installer again every 30 seconds: a refused 128 GB machine prints the refusal on its screen for ever, and the hub shows a bound appliance that never installs, with no reason. (2) An install that went on with `--accept-small-disk` sends `"small_disk":true` in its enrolment request; the hub ignores the field, so no page shows which boxes were accepted small (`16-storage-physical-identity.md` §14.5, the last row of the built table). | OPEN — **2026-11-15 evening (second build): part (2) is BUILT, not released** (hub `main` after 0.182.0): the hub keeps `small_disk` (and a size, when an installer sends one) from the enrolment and the box's page shows „System disk: small — accepted below 200 GB”. **Part (1) is open, and one more thing was read:** the hub's bind form has NO field for extra installer arguments (the handler takes `extra_args`, the page does not offer it; the household's own bind sends none), so `--accept-small-disk` cannot reach an install from the public image today (`audits/build-release-4b-2026-11-15/E/small-disk-flag-on-the-public-image.md`). Earlier: OPEN — filed 2026-11-15 | the next installer ISO for (1) | (1) a final exit code for a refusal that no retry can cure; the first-boot unit stops retrying on it, shows the sentence once, and tells the hub the reason; (2) the hub stores `small_disk` at the enrolment and shows it on the box's page | CC |
| **R-1400** | Monitoring & notifications | P4 | **Every household mail but one prints the event's raw details as a line „Megjegyzés”.** Found 2026-11-15 while building the receipt mail of D84: the hub's household frame adds the event's details JSON under „Megjegyzés” (a note); the new receipt mail leaves it out, the other household mails keep it. A household reads a line of field names and values. Not measured: which mails, and how it reads in an inbox (`hub/internal/notify/testdata/mail_goldens/`). | OPEN — filed 2026-11-15 | — | read the household mail frame: which event types carry details worth a sentence, and drop the raw line for the rest | CC |
| **R-979** | Box system & updates | P3 | **Controller updates are trusted by registry tag alone (no digest, no signature), and the Update button offers any newer tag, even an unvouched one; the built-in traefik, cloudflared and FileBrowser images are also pinned by tag only.** Whoever can push to the registry decides what runs. Fix: pin by digest from the hub's vouched list; offer only vouched versions. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN — design-proposal | — | Design note, then build | CC |
| **R-974** | Box system & updates | P3 | **A compromised host agent can likely become root on the host through the root wrapper that applies PBS settings: the wrapper accepts a folder path inside the agent's own writable tree and then acts on it as root.** Static (the symlink behaviour was reasoned, not run). Same class as R-861. Fix: a fixed root-owned folder, and refuse symlinks before any root action. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN | — | Fix the wrapper with a test on a throwaway box | CC |
| **R-979** | Box system & updates | P3 | **Controller updates are trusted by registry tag alone (no digest, no signature), and the Update button offers any newer tag, even an unvouched one; the built-in traefik, cloudflared and FileBrowser images are also pinned by tag only.** Whoever can push to the registry decides what runs. Fix: pin by digest from the hub's vouched list; offer only vouched versions. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN — design-proposal **2026-11-19 (security review 2, S2-29): still open as written in controller `10df99b8` — tag only, highest tag offered, the „pushed once” rule lives only in the build script.** | — | Design note, then build | CC |
| **R-974** | Box system & updates | P3 | **A compromised host agent can likely become root on the host through the root wrapper that applies PBS settings: the wrapper accepts a folder path inside the agent's own writable tree and then acts on it as root.** Static (the symlink behaviour was reasoned, not run). Same class as R-861. Fix: a fixed root-owned folder, and refuse symlinks before any root action. Security review 2026-10-12 (`audits/security-review-2026-10-12/`); detail held off-git for the operator (`/mnt/5_hdd/felhom.eu/evidence/security-review-2026-10-12-PRIVATE/`). | OPEN **2026-11-19 (security review 2): not re-read in this pass; three more root helpers of the same class are in R-1428.** | — | Fix the wrapper with a test on a throwaway box | CC |
| **R-1214** | Box system & updates | P3 | **On the 8 GB box with the swap fix, the night's second-drive backup still pushes memory „full" pressure above 20 % for about 3 minutes; the box keeps answering.** Night soak 2026-10-21→22 (`audits/night-2026-10-21/P1-soak/03-rec.jsonl`, 10-s samples; fresh box from the public ISO, agent with swap = RAM, 8 apps, the rehearsal uploads for 2 h): at 01:30–01:34 UTC (the 03:30 tier2 job) host 39 % / guest 45 % over 10 s, over 60 s at most 20 % / 25 %; IO „full" 66 %; guest swap 1.35 → 2.1 GB; `/api/health` answered every probe (slowest 4.1 s, largest interval 16 s); 0 OOM kills. Once more in the upload window (Immich loading its models, 31 % / 35 % for one sample). The brief's soak pass line was „full" < 20 %, so the soak missed that one line; the freeze itself (96 % for an hour, R-1192) did not come back. Not diagnosed: which part of the tier2 job (the copy's page cache, the volume dumps) causes it. | WATCHING — **2026-10-24 02:4xZ dated check read (weekend-night session, read only):** Tester 1 (8 GB host, guest 4 GB, controller 0.341.0) ran its 03:30 copy at 01:30:00–01:30:13Z (3 apps, 157 MB + 86 MB + 16 KB; no „capped copy container did not start" WARN, no leftover `felhom-tier2-copy-*`); the guest cgroup's memory „full" stall counter on the host (`/sys/fs/cgroup/lxc/9201/memory.pressure`) read **total 444 µs** over the whole 11.1 h since the host started (15:32Z), so „full" stayed far under 20 % (20 % of 10 s = 2 s); 0 OOM kills. **Not shown:** that the copy ran in the capped container (the INFO log does not name it; Docker keeps only ~200 events, from 02:36Z). **Weak load:** 0.24 GB, not the soak's large copy — so the row stays WATCHING; close after a night whose copy moves ≥ 1 GB, read from the same counter. WATCHING — **RELEASED 2026-10-23 in controller 0.340.0** (floor 06:41 UTC; on all three test boxes, start-up „no leftover”, 0 copy containers; Tester 1's manual second-drive run: 3 apps, 0 containers left). **Open only for the consequence:** dated check 2026-10-24 — Tester 1's first night copy ran in the capped container and memory „full” pressure stayed under 20 % (the box's own night log). READY — rides controller 0.340.0. READY — waits for the controller 0.340.0 release. **2026-10-22 night — cause measured, fix built and proven on 9202, waits for the controller release.** Cause: the copy's page cache is charged to the guest and evicts the apps' working set. On 9202 (4 GB RAM, 1.8 GB held by a hog, a 6 GB copy): plain rsync full stall 5 702 / 2 464 ms, peak „full” 16.8 / 6.9 %; the same copy in a 512 MB memory-capped container 374 / 320 ms, 0.9 / 1.0 %. Controller main `03c0a69` + `ae39467` (unreleased, rides 0.340.0): the copy runs capped only for clean `/mnt/` paths with no symlink (security review); a docker error (exit 125) falls back to plain rsync. A real `POST /api/backup/tier2` on rc.4: 6 capped containers, each exit 0; Immich's copy equals its source (dry-run 0 differences, 44 = 44 files). Evidence `audits/night-2026-10-22/P4-R1214/`. **2026-10-23 night (stray-copy fix, night brief Part 1).** The capped copy's container outlived the copy (a timeout or a controller restart killed only the `docker` CLI; `--delete` went on outside the R-1180 guard). Controller main `66a69d4` + `912e592` (unreleased, rides 0.340.0): a fixed name + label per copy, removed by name on any error, every leftover removed at start before the scheduler and the web server, a restore refused while a copy runs or may run; `--mount` (a path the daemon does not see is refused, not created empty) and the image ID with `--pull never` (review C4, C7). Proven on 9202 with rc.5 (`audits/night-2026-10-23/P1-straycopy/`): a restart in the middle of a 16 GB copy → the new controller removed `felhom-tier2-copy-immich-948f2280` at start, the restore probe was refused during the copy, the redone copy equals the source; a 15 s test build → the timed-out copy's container removed, the partial copy stopped growing. A second review of that code (N2–N5) and the commit's own security review (two races) → `9d9e404`, `1babfe9`: a stray clears when a later removal works, the start-up check never removes this process's own copy and holds the destination's lock, one copy per destination at a time, `RunTier2` refuses during a restore for every caller and records itself in one step with that check, every leg asks again. Re-proven on 9202 with rc.8 (`audits/night-2026-10-23/P1-straycopy/1*`). 9202 back on 0.339.0. | — | Release 0.340.0; then re-measure the 03:30 job on the 8 GB soak box and close if „full” stays low | CC |
| **R-862** | Box system & updates | P3 | **Tester 2 cannot take the config bundle until one by-hand bootstrap is done: its `felhom-os-apply` (agent 0.142.0) predates the bundle mode, and no signed job can write a root file on it.** FOUND 2026-10-04 (R-840 build, Part C): the route reaches every box installed from installer 1.31.0 on, and the demo boxes (bootstrapped by CC); Tester 2 has every root file of agent 0.142.0 (installer 1.30.0) and lacks only the R-858 wrapper fix, which matters only for a Docker step it gets solely from a signed job. CC has no route to Tester 2 (its door admits only the operator's WireGuard peer; `felhom-op` cannot become root). The steps: `runbooks/config-bundle.md` "Tester 2". Then CC signs the bundle and reads it back. | **WAITING-ON-OPERATOR** — the operator said (2026-10-04 ~19:05) he will try through his tunnel **2026-10-07 07:58 (`09` §3 decision 170): the operator believes Tester 2 never set up the off-site backup (no escrow); the laptop may stay off for days; the tester plans to add an HDD.** **2026-10-07 (Part H, read only):** in Tester 2's last 10 notifications (2026-10-05 09:16 → 2026-10-07 06:58) the household got 0 mails, the operator 10 (host_down ×9, expected_dbdump_missed ×1); control: the household channel shows on the other three customers. So the household is not mailed daily while the box is off; no row (`audits/day-2026-10-07/H/`). **2026-10-22: closes when Tester 2 is reinstalled (operator ruling 2026-10-22 14:11 — a new install from the public download replaces the bootstrap and the off-site check; `pilot/TESTER2-handover.md`).** | the operator's WireGuard tunnel | the operator runs the three bootstrap commands; CC sends the bundle | operator |
| **R-78** | Box system & updates | P3 | **`local_api` authority ruling — auto-reconcile vs detect-only** **MIGRATED FROM `ROADMAP.md` 2026-08-22 (R-369) — originally filed 2026-07-25, size M, roadmap state `idea (deferred OUT of R-77 on purpose)`.** Moved verbatim; nothing added or reinterpreted. The roadmap keeps its copy as history, marked moved. **An OWED OPERATOR DECISION, not a defect — moved because an owed ruling hidden among feature ideas is the shape this session exists to remove.** | **OPEN — migrated from ROADMAP 2026-08-22, rank unchanged** | — | R-77 ships detection because the fix is genuinely undecided, and **both directions can lose customer-visible function**. **Direction 1 (today):** `controller.yaml` wins and drift is silent → the 2026-07-25 island migration blinded the whole fleet's control plane for 17.5 h (drive gate, guest-reboot recovery, quiesce/backup all degrade). R-77 makes that loud but does not stop it recurring. **Direction 2 (`bootstrap.json` wins, auto-reconcile on boot):** a guest whose `controller.yaml` is CORRECT and whose `bootstrap.json` is stale — a half-completed re-provision, a hand-repaired guest, a setup-wizard box — gets a **working channel clobbered on the next restart**, fleet-wide and silently, during a routine deploy. That is not obviously better than the bug. Needs a spike: which writer is authoritative per field (endpoint vs fingerprint vs token — `mergeLocalAPI` replaces the whole block, so they cannot be reconciled independently today), whether the agent side should stamp a generation/mtime so 'newer wins' is even expressible, and whether reconcile should require an operator ack. Until then the drift alert plus a manual edit is the supported path. | CC |
@@ -303,10 +311,12 @@ stopping line that lies.
| **R-1358** | Box system & updates | P3 | **cloudflared 2026.10.4 misses two dependency fixes (VULN-183456 chi, 2026.10.5; VULN-184686 go-jose, 2026.10.6; the newest is 2026.10.7).** No public text on reach. Not pinned tonight: the tunnel is a box's only door, and 9202 has none to prove it on. | OPEN — filed 2026-11-03 | — | raise `CloudflaredImage` on the next controller release with a tunnel proof on Tester 1 | CC |
| **R-1382** | Box system & updates | P3 | **A box that was off past an end date: nobody knows if it comes back by itself (5 answers).** Found 2026-11-11 (the End-dates list, class T4, `documentation/audits/end-dates-list-2026-11-11/TABLE.md`; read in source and documents, never tried). After months without power: the web certificate has ended (DB01-6: does Traefik renew at the first start, and what does the household see until then), Debian's package keys may be too old to verify an update (DB16-6), the tunnel client may be below what Cloudflare accepts (DB19-6), a 30-day restore copy is past its date (DB24-6), and the 14-day countdown after a stopped off-site copy ran out while the box was off (DB04-6). Tester 2 has been off since 2026-10-04 and is the natural test. | OPEN — filed 2026-11-11 (end-dates list) | — | When Tester 2 comes back (or with a scratch host and a moved clock on Tier 0, as a task): record what renews by itself and what needs a hand; put the hand steps in a runbook. | CC |
| **R-1384** | Box system & updates | P3 | **A verdict or an „already ran” check trusts a clock that can be wrong (30 answers).** Found 2026-11-11 (the Clock list, class W2, `documentation/audits/clock-list-2026-11-11/TABLE.md`). MEASURED: the N100's host is 32.6 s behind and its time service does not run (R-1301); the other two boxes are in step. Nothing checks a box's clock except one hub alarm for a tunnel handshake stamped in the future. The hub takes the box's own time for backup age, off-site age and proof age (CH02-6, CH07-6, CH07-8, CH08-6); a box with a wrong clock refuses every signed job (CA09-6); a stamp from the future blocks the OS-update leg, the trim and the line test until the clock passes it (CA02-6, CA02-8, CA11-6, CA11-8, CA12-6); the crash guard reads the clock at early boot (CA04-6, CA04-8); a hub clock jump moves its own daily jobs (CH03-6, CH04-6, CH06-6, CH09-6, CH16-6, CH18-6, CH20-6, CH24-6). The others: CA05-6, CA05-8, CA06-6, CA07-6, CA10-6, CA20-6, CC04-6, CC05-6, CC14-6, CC18-6. Known parts stay in their rows: R-1189, R-1206, R-1260. **Ruled 2026-11-11 13:17: yes (`09` §3 decision 300). Not built.** **2026-11-11 (afternoon): why the N100's clock is wrong is known (R-1301, `documentation/audits/n100-clock-2026-11-11/README.md`).** Numbers from source: a host clock more than 120 s behind refuses every signed job (`felhom-agent/internal/authz/verifier.go:21`, `:153-162`); at about 1.25 s a day the N100 gets there in about 70 days (mid-January 2027, an estimate from one measurement). | NARROWED — **2026-11-15: LIVE and seen on a box** (hub 0.182.0, agent 0.175.1). Tester 1's clock was set 150 s back twice: the hub said „150 s behind”, one mail went to the operator only, a second alarm inside the hour sent no second mail, and the hub said „in step again” after the clock was put right (`audits/release-2b-3-2026-11-15/` `part2/`, `part3/`). LEFT, as before: the 30 answers themselves still trust the clock. Earlier: NARROWED — the hub's check is built, not released. **2026-11-14 (night build):** the agent stamps its clock at the moment of sending (`sent_at`) and the hub alarms the operator when it is more than one minute from the hub's clock, in either direction, and says so when it is back (`08-alarm-ladder.md`, `box_clock_wrong`). A report that waited on the box raises nothing (tested). Not released, not seen on a box. LEFT after the release: the 30 answers themselves still trust the clock; the alarm only makes a wrong clock known. OPEN — **2026-11-14: the one box whose clock was wrong (the N100, 32.6 s) is right again (R-1301 closed); this row's check is still not built.** Earlier: OPEN — filed 2026-11-11 (clock list) | — | Fix shape (ruled, not built): the hub compares the time in each report with its own, alarms the operator above one minute, and does not take a time from the future as „fresh”. | CC |
| **R-1428** | Box system & updates | P3 | **Three root helpers on the host trust a path inside the agent's own folder, and the disk-format route's „confirmation” is two fields the controller sends itself.** A taken-over agent process can likely change the owner of, or bind, a host folder of its choice, and can roll the agent back to the previous binary at will (same class as R-974); a taken-over controller can wipe any plugged, unmounted, non-system disk — on a BYO host the owner's own disk (`16` allows the household's confirmation; `03` §4 asks for a signature when the only copy is destroyed — the two documents disagree). Static; own second read. Fix: root-owned folders and no-follow opens in the helpers; the marker in the root-owned update folder; for the wipe, an operator decision on BYO. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-23, S2-24; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Fix task with R-974 (agent release, operator vouch) | CC |
| **R-1004** | Box system & updates | P4 | **After an unclean stop every guest ext4 volume waits ~43 s in multi-mount protection before it mounts, so each restart after a power cut takes ~1.5 min longer.** Kernel: „EXT4-fs warning (device dm-6): ext4_multi_mount_protect: MMP interval 42 higher than expected, please wait" on dm-6/dm-7 and loop0/loop1, demo-hp and Tester 1, every round of `audits/night-power-2026-10-13/` (R1 `repairs.txt`). | **READY** — filed 2026-10-14 | — | CC: find where MMP is switched on for the guest volumes (Proxmox `pct` create, or our own mkfs), whether it protects anything on a one-host disk, and measure the restart without it on a scratch guest | CC |
| **R-1013** | Box system & updates | P4 | **After a power cut the agent's start-up stale-lock pass logs an ERROR „pct unlock failed … no lock found" when Proxmox's own start-all removed the lock in the same second.** The outcome is the goal (no lock), so the line is a false alarm in the log. Seen on demo-hp 2026-10-14 12:07:44–45 UTC (`audits/power-cut-fixes-2026-10-14/E/after-boot.txt`; Proxmox: „removed left over backup lock from '9201'!"). Small, but the agent's release for this session is spent (v0.156.1). | **OPEN** — filed 2026-10-14 | — | CC: treat „no lock found" from `pct unlock` as already clear (INFO), with a test; ride the next agent release | CC |
| **R-1061** | Box system & updates | P4 | **A per-customer floor reached Tester 1 only at its next 15-minute report, not in seconds.** Set 2026-10-15 12:55:51 UTC for tester-1, demo-hp and demo-felhom together (hub 0.153.0, CC's token): demo-hp and demo-felhom ran 0.314.0 within 13 s (the hub's wake); Tester 1 moved at 13:09:35, its regular report (`audits/gate-restore-2026-10-15/release/03-boxes-move.txt`). It had restarted at 12:54:29 onto rc.1 and logged „hub wait channel active". Why the wake did not reach it is not read (the wake line is DEBUG). No harm: the floor arrived. | **OPEN** — filed 2026-10-15 | — | Read Tester 1's DEBUG ring at the next floor change; compare with the hub's bump log | CC |
| **R-1285** | Box system & updates | P4 | **D43's removal of our SSH drop-in from a BYO host happens one bundle LATE: the bundle run that delivers agent 0.165.0 is planned by the box's OLD (0.164.0) wrapper, which writes the drop-in as before („same"), and only then replaces itself.** Measured 2026-10-25 on the BYO test host (VM 9371 on demo-hp, `Tester4-f967d4`; attended release run): 0.164.0 day-0 → signed `agent_update` + `agent_config_update` 0.165.0 → the planted drop-in still there; a second `agent_config_update` (same bundle) → `BUNDLE REMOVED … (ours, on a BYO host — D43)`, nothing written, `sshd -T` the owner's own `permitrootlogin yes` (`audits/release-2026-10-26/byo/byo-proof.txt`). No box is affected today (the fleet's only BYO host is this test host); a fresh BYO install on ≥ 0.165.0 runs the new wrapper from the start. | **VERIFY** — 2026-10-27: agent v0.167.0 RELEASED; in the NORMAL signed delivery the pass now runs right after the commit on all three boxes („running the bundle policy pass after the self-update commit (R-1285)” → `BUNDLE POLICY DONE install_mode=appliance removed=0`; `audits/release-2026-10-27/agent-0167-tester1-demofelhom.txt`). Not live: a BYO box coming FROM 0.164.0 (none in the fleet). **Earlier:** **VERIFY** — 2026-10-26: agent `b1bc63d` (unreleased) — the new agent runs the bundle-policy pass once right after its self-update commits (`selfupdate.SetOnCommitted`); test in the real order (commit, then one pass), red-proved. **Not live:** only a box coming FROM 0.164.0 discriminates (0.165.0's wrapper already applies D43), which needs a signed downgrade chain on a BYO box — the live proof rides the next agent release. **Earlier:** **OPEN** — 2026-10-26 (agent 0.166.0 released, attended run): the policy pass is wired, but in the normal delivery it does NOT run: the box gets `agent_update` and `agent_config_update` in ONE pass, and the bundle job runs in the OLD process (demo-hp: pid 1966780, agent 0.165.0, bundle 09:03:07–09:03:09 CET; the restart into 0.166.0 at 09:03:10) — no `bundle policy pass` line on any of the three boxes. So a BYO host coming from 0.164.0 still loses our drop-in one release late. The night proof (`audits/night-2026-10-25/rig/r1285-live.txt`) ran a bundle under an agent that was ALREADY 0.166.0-code, so it did not cover this order. No BYO host is in the fleet today. Fix options: the agent runs the pass once at start-up after `update committed`; or the release procedure sends the bundle job only after the commit. **Earlier state:** **VERIFY** — fixed on main 2026-10-25 night (agent `abcc1e6`), waits for the agent release + the vouch: the agent runs the wrapper's `bundle-policy` pass after every install (red-proved, `redproof-r1285.txt`). **Live on the rig in ONE signed run** (BYO host, staged as upgraded from 0.164.0: the new agent by hand, the old 0.164.0 wrapper on disk, our drop-in planted; one `agent_config_update` of a test bundle): the old wrapper installed the bundle and replaced itself, then „BUNDLE POLICY REMOVED … (ours, on a BYO host)", `sshd -T` the owner's own `permitrootlogin yes` (`rig/r1285-live.txt`); the test package deleted from the registry after. **Earlier state:** OPEN | — | either the wrapper plans again after it replaced itself (one agent change + a test that a 0.164.0→0.165.0 bundle on a BYO record removes our copy in ONE run, red first), or the release procedure sends a BYO box's config bundle twice; until then a BYO box upgraded from ≤ 0.164.0 keeps our drop-in until its next bundle | CC |
| **R-1432** | Box system & updates | P4 | **Agent and host hardening, low risk each (security review 2): the root wrapper's used-number file reads as empty when it is damaged; a signed update has no „not older than what runs” rule; the PBS helper grants before it checks the storage type; the poke listener is bound to an address, not to the tunnel interface; the agent trusts any public certificate for the hub's name; two restart races report a wrong job ending; the `pct` grants take any guest number; a local mount needs no `nosuid,nodev`; snapshots have no count limit; the power-off „once per boot” key moves with the clock; a plan file is checked, then opened by name; the pairing code file stays world-readable on the host after enrolment (measured on demo-hp).** Fix each in its own small change. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-28; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19 | — | Batch into hardening work | CC |
## Monitoring & notifications — 29 rows (P2 3, P3 18, P4 8)
@@ -343,7 +353,7 @@ stopping line that lies.
| **R-1396** | Monitoring & notifications | P3 | **The agent now names WHY its trust in Proxmox is wrong, and the hub does not tell the operator yet.** Found 2026-11-14 night (release 2b, built). The agent's host report carries a `proxmox_trust` stanza (`mode`, `problem`, `detail`, the CA copy's end date; `03-host-agent.md` §12a). Two gaps: (1) the hub does not read the stanza — a box whose CA copy does not verify, ends within 30 days, or whose move failed is said only in the box's journal; (2) a box that is fully blind (neither the pinned certificate nor the CA copy works) sends NO host report, because the report starts with a call to Proxmox — that case reaches the operator only as „host stale” after 45 minutes, with no cause. With release 2b a pinned box heals itself at the first failed call, so (2) needs both ways to fail at once. | NARROWED — **2026-11-15 evening: the list that gap (2) waited for is made** — every reader of a host report in the hub, 29 of them, with what a short report would do at each (`audits/build-release-4b-2026-11-15/F/host-report-readers.md`). The result changes the design: keep the short report OUT of the report table, and only one handler branch, the stale check and three pages must change. Not built. Earlier: NARROWED — **2026-11-15 (build of release 4): gap (2) has a design, not built** (`03-host-agent.md` §12a item 8): a short report with the cause, sent only to a hub that says it understands it; the hub skips every reader but last-seen, the clock, the trust check and one new alarm. Not built because it changes what a report means for the hub: every reader of a host report in the hub must be listed first. The test that settles it: `pveproxy` stopped for 20 minutes on demo-hp. Gap (1) is live since hub 0.182.0. Earlier: NARROWED — **2026-11-15: gap (1) is LIVE and was seen from a real event** (hub 0.182.0): Tester 1 reported `ca_copy_unavailable` for the minutes between its agent update and its root-files update; the hub mailed the operator only, named the cause, and said „no problem any more” at the next report. That alarm was FALSE (the pin worked) and agent 0.175.1 no longer raises it. **LEFT: gap (2), the fully blind box** — a box whose trust does not work sends no report, so a staged wrong CA copy can raise no alarm (it also repairs itself at the agent's start, seen on demo-hp). Earlier: NARROWED — **2026-11-14 night: gap (1) is BUILT, not deployed** — the hub reads `proxmox_trust` and raises one operator-only `proxmox_trust_problem` per (box, cause), and `proxmox_trust_ok` at the recovery (hub `main` after v0.181.0; the wire-contract gate asked for the receiver in the same push as the sender). **LEFT: gap (2), the fully blind box.** Earlier: OPEN — filed 2026-11-14 | — | hub: an operator-only event from `proxmox_trust.problem` (raise once per cause, clear when it is empty), and the mode on the System page; agent: send a report without the Proxmox part when the failure is one of trust (a design choice: what the hub does with a report that has no guests) | CC |
| **R-1377** | Monitoring & notifications | P4 | **Two endings tell nobody outside the page.** Found 2026-11-10 (the endings list, `documentation/audits/endings-list-2026-11-10/TABLE.md` endings EC03, EC15; read in source). **Orphan delete** of an app sends the hub a fresh report and no 'app removed' event (`internal/api/router.go:1150-1178`); the normal remove sends one. **Wipe of a drive** sends no event to the hub (`internal/web/storage_handlers.go:911-1000`). Not a member: retiring a drive — the hub does not read that field by ruling (decision 130, closed R-888). **Members (answers marked N):** EC03-6, EC15-6. | OPEN — filed 2026-11-10 (endings list) | — | Controller: send the 'app removed' event from orphan delete; decide in `08-alarm-ladder.md` whether a wipe is an operator event. | CC |
## Hub & operator — 18 rows (P2 1, P3 5, P4 12)
## Hub & operator — 21 rows (P2 1, P3 6, P4 14)
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
@@ -365,7 +375,9 @@ stopping line that lies.
| **R-1376** | Hub & operator | P3 | **Records of a thing that is gone stay in six places.** Found 2026-11-10 (the endings list, `documentation/audits/endings-list-2026-11-10/TABLE.md` endings EH00, EH02, EH21, EA00, EA11, EA27; read in source, `documentation/audits/endings-list-2026-11-10/evidence/hub-delete-statements.txt`; no hub database was read). (1) **Hub:** thirteen tables have no delete statement anywhere. Deleting a box leaves its update settings, its update reports and its kernel notices; deleting a household leaves its off-site key rows, off-site window rows and abandon requests. Most are small; the update reports of each box grow with time. (2) **Host uninstall** leaves the mount units and host mounts of household drives (said when R-179 closed; no open row carried it). (3) **A drive wiped by a signed job** keeps a record and a mount unit that name the old filesystem (agent `internal/signedjobs/wipe.go:61-110`). (4) **The token of a destroyed guest** stays valid until a new guest takes its number: the token store has no remove (agent `internal/localapi/tokenstore.go:127-210`). **Members (answers marked N):** EH00-4, EH02-4, EH21-4, EA00-4, EA11-4, EA27-4. | NARROWED — **2026-11-15 evening (second build): D86 is BUILT, not released** (hub `main` after 0.182.0; `05` §14.3): at a deleted household's year-end, in one transaction with the audit rows, the journal rows of that deleted life lose their ids (the empty string; the rows, times and step states stay). The first real run is on 2027-09-25. What it does not do is R-1403. The privacy sentence is not published. Earlier: NARROWED — **2026-11-15 evening — RULED: D86 = remove (decision 314).** Fix shape: one more statement in the one-year prune of a deleted household, in the same transaction — the id columns of `host_deletions` and `customer_resets` of that household hold no id after it (`05` §14.3, option A, its six rules and its test list). Earlier: NARROWED — **2026-11-15: two more hub marks are not removed at a household's delete** — `offsite_snapshot_listing:<household>` and `offsite_window_grant:<household>` in `hub_settings` (found while building D72; the same shape as the alarm marks). To add to the delete and to the sweep. Earlier: NARROWED — **2026-11-15: part 1 is LIVE** (hub 0.182.0). The first sweep at the start deleted 26 rows (old update reports of 13 boxes deleted earlier), the same count as the preview on a copy; the ids are kept off git. A box delete by the operator at 11:29 ran with the new code. **One thing the sweep does not cover:** the hub's own alarm marks of a box deleted BEFORE the release (4 marks of one such box were read today) — a delete now removes them, the sweep of old ones does not. Small; to add to the sweep. Earlier: NARROWED — part 1 (the hub, D73) is built, not released. **2026-11-14 (night build):** deleting a box removes its update settings, update reports, kernel notices and the hub's alarm marks that name it; deleting a household removes its off-site key, window and abandon rows; the operator's action log of a deleted household goes one year after the deletion with the other audit rows (decided by CC unattended — the operator may reverse); the delete journal stays; a daily sweep removes what earlier deletes left (`05` §14.2). The privacy sentence is proposed in the report, not published. LEFT: the agent part (a destroyed guest's token), parts (2) and (3), and an age limit for a LIVE box's update reports (needs a design: the kernel lane reads old reports). OPEN — filed 2026-11-10 (endings list) **2026-11-10 15:19 — RULED (D73 = yes, with one exception; `09` §3 decision 293).** Not built. | — | Hub (D73): add the tables with no delete to the two purges and to the daily clean-up; the delete journal (`host_deletions`, `customer_resets`) stays on purpose, and the privacy text says so. Agent: remove a guest's token when the guest is destroyed. Parts (2) and (3) have no ruling of their own. | CC |
| **R-1403** | Hub & operator | P4 | **A household whose delete never ended keeps its ids in the delete journal with no end, and no page names it.** Found 2026-11-15 (D86, `05` §14.3 rule 4). The year after which the journal loses the ids starts when the delete cascade is COMPLETE. If one step failed and nobody pressed Resume, the journal rows, the events, the mail log and the action log of that household stay for ever; the System page does not show such a household. Not measured: whether one exists on the live hub. | OPEN — filed 2026-11-15 | — | one line on the System page for a delete that started and did not end, with its age; then decide what ends it | CC |
| **R-1386** | Hub & operator | P3 | **Timed jobs with no time limit and no measured duration (25 answers).** Found 2026-11-11 (the Clock list, class W4, `documentation/audits/clock-list-2026-11-11/TABLE.md`; read in source; the hub was not measured, by ruling). The hub's 60-second sweep runs its checks one after the other with no limit for each: one slow check delays all the others, the liveness check too (CH19-3, CH19-8, CH01-3, CH08-3, CH08-8, CH10-3, CH13-8, CH18-3, CH06-8, CH02-3); the 07:10 audit has 10 minutes, measured with three boxes only (CH11-3, CH11-8). On the box: the interval jobs have no limit of their own (CC11-3, CC12-3, CC13-3, CC14-3, CC15-3, CC16-3, CC20-3, CA06-3, CA10-3, CA10-8, CA13-8, CA15-3, CA19-3). Every measured duration is from a small test store. | OPEN — filed 2026-11-11 (clock list) | — | Hub: log each sweep step's duration and alarm when one sweep takes longer than its interval; then read a week of numbers and set limits. Box: the same for the controller's interval jobs. | CC |
| **R-1426** | Hub & operator | P3 | **The hub has no limit on what one box may send, its limits by visitor address rest on an ingress setting that is not in git, its country filter lets every `10.` address in, and one ring-0 test box alone decides a package of the automatic OS release.** One taken-over box can fill the hub's disk or hold its connections (reports and alarms then stop for all); whether the address limits of the bind page and the appliance register hold could not be told from source; a taken-over test box can put a package line into every ring-1 box's list (the box re-checks the name and version shape). Static; own second read. Fix: per-key limits on the report, event and off-site routes; read and record the ingress setting, key on the hop the ingress wrote; narrow the `10.` rule; weigh a two-box agreement for ring 0. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-19, S2-20; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Measure the ingress first (one request, then the hub's own access log); then a fix task | CC |
| **R-1393** | Hub & operator | P4 | **For about 8 minutes after a hub restart its database statements take 1 to 32 seconds; the System page took 12.7 s and a box's enrolment 11.4 s.** Measured 2026-11-14 at the 0.181.0 deploy (`audits/release-1-dates-2026-11-14/hub/05-watch.txt`): 24 `slow statement` lines from 09:58 to 10:06 UTC, none in the 20 minutes after. The database file is 678 MB; the slow ones read `host_reports`, `os_reports` and `events` (whole report JSON per row). One 32-second statement ran while CC copied the database out of the pod for a read (CC's own load). No report was lost that CC could see; not measured: whether a box's report timed out in that window. | OPEN — filed 2026-11-14 | — | measure a restart with no copy running; then decide: an index or a shorter keep on `host_reports`/`os_reports`, or accept | CC |
| **R-1430** | Hub & operator | P4 | **Hub hardening, low risk each (security review 2): no security headers on any hub page (measured: no frame, content-type or script policy on the sign-in page); a link written by a box is shown on the operator's customer page; the bind page's try counter is read before the check and its address table never forgets; a break-glass reveal for a host with no customer leaves no timeline row; a hosts token can read app log tails; unknown and wrong ids answer differently on four routes; two reads fail open on a database error; the route tests pin the switch, not what each handler checks.** Fix each in its own small change. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-26; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19 | — | Batch into hardening work | CC |
## Business & legal — 14 rows (P2 4, P3 2, P4 8)
@@ -386,7 +398,7 @@ stopping line that lies.
| **R-990** | Business & legal | P4 | **Operator: read `documentation/audits/forum-research-2026-10-12/FELHOM.md` §5 and rule on each recommendation.** READ-ONLY research 2026-10-12: three years (#7943–#11092, 2,908 posts, 195 people, 82 beginner threads) of the public PROHARDVER! topic „Home server / házi szerver építése”, tagged and counted by person (`FINDINGS.md`, method and spot-check in `METHOD.md`; no username or post text in git). Ten ranked recommendations, none acted on; the top three: a „which machine works” page (8 of 33 beginners who name their hardware own only a Synology, a router disk or a Raspberry Pi — the installer cannot run there; the same gap is R-989's first recommendation from a different source, and the shipped-page half is R-991), „you do not choose an operating system — we do” (the biggest beginner question, 20 of 82 threads), and „bring your old office PC” as the cheap entry. Catalog gap (≥ 3 people): a DLNA server (7), an ad/DNS blocker (5), a camera recorder (5). | **WAITING-ON-OPERATOR** — filed 2026-10-12 | the operator's reading | Operator: rule on each of the ten (yes / no / later). If nothing is done: nothing changes; the research ages | operator |
| **R-1035** | Business & legal | P4 | **Did the „felhő tárhely” page earn anything? Two dated checks.** **Baseline (reported by the operator from Search Console, 2026-10-14, last 3 months):** 1 click, 48 impressions, CTR 2.1 %, average position **73.6**; „felhő tárhely” 19 impressions, „felhőtárhely” 5, „felhő alapú tárhely” 3. CC measured separately that the phrase „felhő tárhely” appeared **0 times** on the home, FAQ and backup pages — Google was showing the home page for those queries by accident. `/sajat-felho-tarhely` + `/en/own-cloud-storage` published **2026-10-14**; full baseline, price sources and the power provenance in `audits/seo-2026-10-14/BASELINE.md`. **What to measure, both times:** impressions and average position for „felhő tárhely” and „saját felhő tárhely”, plus the new page's own impressions, clicks and position. **A null result at +4 weeks means nothing on its own** — a new page on a site at position 73.6 takes weeks to months to move, and „felhő tárhely” alone is held by Microsoft, Google and the big list sites; the page aims at the two longer phrases. That is why there are two checks and not one, and why neither should be read as a verdict on the copy. **There is no Search Console connector in this project, so CC cannot read these numbers** — the operator reads them, CC records and compares. **Blocked until the operator asks for indexing** (Search Console → URL-ellenőrzés → the URL → „Indexelés kérése”): until that happens, or Google finds the page itself, both checks read zero for a reason that is not the page. **This is a slow channel and does not replace talking to people.** **2026-11-11:** the 4-week check came due today in a run with no operator present; only the operator can read Search Console, so the date moved to 2026-11-18 (the gate's rule allows a move with a reason). The 8-week check keeps its date. | **WATCHING** — 2026-11-19: the 4-week read moved to 2026-11-26 (it is the operator's read from Search Console; the release run of this day cannot take it, and asks in its note). | operator: request indexing, then supply the Search Console read | 2026-11-11: first read (+4 weeks); 2026-12-09: second (+8 weeks), compared with the first | operator + CC |
## Process & tooling — 36 rows (P3 9, P4 27)
## Process & tooling — 39 rows (P3 10, P4 29)
| ID | Category | Sev | What | State | Blocked on | Next action | Owner |
|---|---|---|---|---|---|---|---|
@@ -437,6 +449,7 @@ stopping line that lies.
<!-- DUE-CHECKS-END -->
| **R-1297** | Process & tooling | P3 | **CI waits dominate a working day: felhom.eu's CI run takes ~13 min and the ONE runner makes every repo's job wait behind it (up to 10–17 min queued).** Measured 2026-10-26 (operator: „these CI waiting times are getting out of hand"): since 10-25 felhom.eu runs median 793 s, the controller/agent/catalog 57–102 s; queue median 5 s, max 548–1023 s. In run 2366, ~10 of 13 minutes are four suites — `test_gate_decoys.py` 232 s, `test_due_checks_gate.py` 133 s, `test_repo_gates.py` 132 s (it runs the whole gate set a SECOND time), `dooplex-offsite/test_dooplex_offsite.py` 69 s — plus `hub-gofmt` ~110 s (a cold Go cache per run). **Ruled 2026-10-26 by the operator: BOTH** — (A) trim felhom.eu CI: the heavy suites only when their own files change, no duplicate full run, a Go cache for hub-gofmt; (B) a second CI runner on DooPlex. | **OPEN — (A) BUILT 2026-10-26 evening, closes after the first nightly run (2026-10-27 03:20 UTC) is read.** Both runners proven at once (jobs 2374 + 2375 on `felhom-gates-runner` / `-2`, 17:11:41Z). felhom.eu CI, one push each: **(a) docs only** 810 s (old runner) → 351 s (2 CPUs) → see the close-out; **(b) a gate script** 351 → 203 s (`ecd82aba`, job 2376); **(c) a hub change** 351 → 118 s (`7e57e824`, job 2378). Heavy suites by path + tag + nightly; `test_repo_gates.py` and due-checks group G test the wiring (130 s each → under 3 s). Decoys through CI's entry point with the filter ON: a docs-only register break → red; an inert gate → GREEN at first (a decoy hole, fixed in `e33ec821`) → red. No Go cache: CI has no Go (the 110 s was security-scan, now 27 s). `audits/r1297-ci-trim-2026-10-26/`. (B) DONE by the operator (`09` §3 decision 267). | — | (A) CC, a run of its own (a CI scope change, ruled); (B) a DooPlex change — a brief or the homelab session; and CC waits on CI only before a release or at the end of a run | CC |
| **R-1402** | Process & tooling | P3 | **The catalog's decoy suite (`scripts/test_gate_decoys.py`, gate `gate-decoys`) is red on catalog main: 17 decoys fail.** Measured 2026-11-15 on a CLEAN clone of `f4c5c79` (32 failures there, 15 of them only because the clone has no felhom.eu sibling) and on the working tree with the Securo change (17, all also in the clean run — none new). The failing ones are the move gate's and the security-fix checks': e.g. „GENUINE: a proven entry whose digest the registry serves" answers 1 (expected 0) — the planted range reads „adds NO update_ladder entry" for navidrome; „a re-test whose digest the registry no longer serves" answers 0 (expected 1). The suite is not in `--fast` and CI does not run it, so nothing stops a push; but a decoy that no longer fails or passes as planned means the `test-record-move` gate is not shown to work. Evidence: the session scratch outputs were not kept — reproduce with `python3 scripts/test_gate_decoys.py` in the catalog. | OPEN | — | Find why the planted ladder entry is not seen (the plant or the gate drifted since 2026-10-27, the last change to either file: `06e603e`); fix, run the suite green, then decide whether CI should run it. If nothing is done: the move gate may refuse or pass the wrong thing and no test says so | CC |
| **R-1429** | Process & tooling | P3 | **One CI runner serves all four repositories in host mode with one shared file system, and the nightly security-watch job holds a write token for `felhom.eu` `main`.** Whoever can push to any one of the four repositories (today: the operator's account only) can reach the others' CI secrets and push to `main`, whose manifests ArgoCD applies. Static; who can push, and whether `main` is protected, was not read live. Fix: a runner registration and secrets per repository, a fresh work folder per job, no write token in a job that runs fetched code, a narrow ArgoCD project for the hub's manifests. Security review 2, 2026-11-19 (`audits/security-review-2-2026-11-19/FINDINGS.md`, S2-25; detail off-git, `/mnt/5_hdd/felhom.eu/evidence/security-review-2-2026-11-19-PRIVATE/`). | OPEN — filed 2026-11-19, technical severity Medium | — | Read Gitea's live settings (who can push, branch protection); then a homelab task | CC |
| **R-1381** | Process & tooling | P4 | **No document says when eleven things end, or where to read the date (11 answers).** Found 2026-11-11 (the End-dates list, class T3, `documentation/audits/end-dates-list-2026-11-11/TABLE.md`; search S18 found no document that lists end dates). OS support for Debian 13 and Proxmox 9 (DB17-2, DB17-3), the tunnel client's oldest allowed version (DB19-2), the installer ISO (DH19-3), and the outside accounts and tokens (DH23-2, DH24-2, DH27-2, DH28-2, DH29-2, DH31-2, DH32-2). This run could not read these dates (no account call, by its fences). The operator reads them once: the table „Only you can read these dates” in `documentation/audits/end-dates-list-2026-11-11/README.md`. | OPEN — filed 2026-11-11 (end-dates list) | — | When the operator has read the dates: write them in one place (a short table in `documentation/operations/`), and give the due-checks block the ones inside a year. | CC |
| **R-1387** | Process & tooling | P4 | **A document and the source disagree on a time, in four places (4 answers).** Found 2026-11-11 (the Clock list, class W5, `documentation/audits/clock-list-2026-11-11/TABLE.md`). (1) `documentation/architecture/03-host-agent.md` names `Restart=always` and a watchdog for the agent's unit; the unit has `Restart=on-failure` and no watchdog (CA18-6, CA18-8). Which one is the design is not stated, so this session did not edit either. (2) The hub-database copy alarm: 13 hours in one place, 26 in another (CH23-7; a DooPlex procedure, listed for the operator, not edited). (3) Old „24 h / 02:30” text for the hub's daily job (CH17-6). **Seen beside the lists, same kind, not in the count:** (4) the comment at `felhom-controller/controller/cmd/controller/main.go:1412` says 14 days where the code keeps restore copies 30 days (D62); (5) `documentation/architecture/07-backup-architecture.md:455-457` cites `main.go` lines that have moved; (6) `documentation/operations/tailscale.md` says „both nodes” where three test hosts exist, and Tailscale was not running on demo-hp and Tester 1 on 2026-11-11 (not known if that is meant); (7) the threat model and the recovery inventory do not say the same about how often the hub bearer is rotated. | OPEN — filed 2026-11-11 (clock list) | — | Operator or a task: say which side is right for (1) and (2); then one docs commit. (3) is a text fix in the same commit. | CC |
| **R-1389** | Process & tooling | P4 | **From the operator's home network the installer address gives 404.** Found 2026-11-11 (`documentation/tests/golden-0.352.0-2026-11-11/fresh-install/installer-address-404-on-lan.txt`). The home router's DNS (192.168.0.250) answers `felhom.eu` with 192.168.0.192, the site copy on DooPlex that was retired on 2026-11-07 (R-1228). So `https://felhom.eu/scripts/…` gives 404 for every machine that uses that DNS; from DooPlex (public DNS) it gives 200. A household is not touched. A test box on the home network, and Tester 2 if it sits there, cannot get the installer by its public address. | OPEN — filed 2026-11-11 (golden fresh install) | — | The operator removes the `felhom.eu` override from the home DNS (it is his router), then one `curl -sI https://felhom.eu/scripts/` from a machine on that network reads 200. | operator |