docs(adr): renumber the second ADR-0041 to 0042, and unbreak main

[Gerrrt]

main is red, and has been since 23:28Z. Every open PR inherits it.

README.md:95 claims 41 ADRs; the repository has 42
README.md:96 claims 26 runbooks; the repository has 27
ADR-0041 is claimed by 2 files: 0041-run-the-crs326-…, 0041-terminate-the-remote-path-…
  — renumber the one that landed second

How

#491 gave ADR-0041 to Run the
CRS326 on RouterOS at 23:27:57Z.
#495 gave the same number to
Terminate the remote path on the lab at 23:28:52Z. Fifty-five seconds
apart, each green against a main that did not yet hold the other.

check_docs.py's ADR-numbering assertion is precisely the check for this. It is
not at fault and it needs no change — it fired the moment both files were in one
tree, which is the first moment the collision existed. A pre-merge branch cannot
see an ADR number that is still sitting in another branch.

What changed

The remote-path ADR landed second, so it becomes ADR-0042 — the file
check_docs.py names. git records it as a rename, so its history follows it.

Filename and its own H1	0042-terminate-the-remote-path-on-the-lab-and-route-it.md
Links retargeted	11
Bare ADR-0041 mentions renumbered	13
Files touched	12

Every remaining ADR-0041 in the tree is the CRS326 one, and each was
classified by hand rather than by pattern — the whole difficulty here is that
two different ADRs shared one number, so a search-and-replace would have
renumbered both. The survivors are in roadmap.md, hardware.md, ADR-0018's
amendment and swap-the-switch.md, and all four are about the switch.

Two of the renumbered mentions sit inside other ADRs' amendment blockquotes
(ADR-0011, ADR-0014). Updating a pointer to a renumbered file is not revising a
decision, so ADR-0001's immutability is not in question — and the alternative is
leaving a reference to a number that now means a different ADR.

open-the-remote-path.md's pfSense rule descriptions move with it
(WireGuard peers — ADR-0042), which is safe because nothing has been typed
into a firewall yet: that runbook is gated on the jumpbox of
#436, and
restore-the-firewall.md still counts the tripwires as though the tunnel were
unbuilt — which it is.

The two README numbers are the arithmetic of the same two merges: one ADR and
two runbooks arrived, and the sentence that counts them did not move.

Blast radius

• No change to network segmentation or firewall rules
• No new port published to a VLAN that could not already reach the service
• No credential added outside secrets/*.sops.yaml

Documentation only — a file rename and reference updates. No decision text is
altered, and no ADR's content changes.

Verification

make check-docs  docs OK — 82 Prometheus + 18 Loki rules, 7 dashboards, 141 panels, 10 assertions
make lint        PASS  (yamllint, markdownlint-cli2, shellcheck, actionlint, editorconfig-checker)

check-docs went from three disagreements to zero, which is the whole point of
this PR. Verified there is no leftover 0041-terminate reference anywhere in the
tree, including outside markdown.

Note

#502 is rebased onto the broken
main and will stay red until this merges. Nothing is wrong with it.

Refs #491 #495

🤖 Generated with Claude Code

[Gerrrt]

Already landed. Closing as a duplicate of
#503, merged just now as
107256f — main is green again.

Three sessions produced this fix independently and all three were
byte-identical: #503 at 23:36:31Z, this one at 23:37:07Z (36 seconds later),
and #509 at 01:05:22Z. Same 12
files, same +27/−27, same 54 content lines once blob hashes and hunk offsets
are set aside — the same renumber of the remote-path ADR to 0042, the same
split of the bare ADR-0041 mentions between the tunnel and the CRS326, and
the same two README counts. #509 was closed for the same reason. #503 won only
on being first.

Nothing was wrong with this PR. A red main is visible to every session at
once, so every session picks it up at once; the duplication is structural, not
a mistake anyone made. Worth noting that check_docs.py naming the remedy —
"renumber the one that landed second" — is probably why all three answers
were identical rather than merely equivalent.

One thing none of the three fixed, and that is still open: README.md:210
says "39 architecture decision records", now against 42. check_docs.py's
counted-claims pattern only matches {COUNT} ADRs, so line 210 is unguarded
and has been drifting silently. Fixing the number without widening the check
would just hide it again — worth its own issue.