docs(sync): harmonize documentation after the lab merge

[Soushi888]

SoushAI analysis. Drafted by Soushi's AI assistant, reviewed and posted by @Soushi888.

No linked issue: this is the documentation sync after #67 and #68 brought the lab session into main.

What and why

After the lab merge, the docs disagreed with each other and with the tree. This PR fixes 22 of those disagreements in 14 files, one commit. When a claim could be checked against the code, the code decided. Otherwise the most recently edited line won, and the higher-level doc won when neither could be checked.

• README. It now counts fourteen VM tests: thirteen built in the nix flake check job, plus vmTestHoloportInstall in its own job. The README used to say ten. The tests table adds vmTestServices, vmTestServices-noBootstrap, vmTestMossNode and vmTestHoloportInstall, and names the checks that need no VM.
  • The status line says sensorica-holoport-01 was installed on 2026-09-27, while the five-machine fleet (Hardware: one Holoport boots vanilla NixOS from the workshop ISO (ADR-003 gate) #8 to Hardware: facilitator guide reviewed with Tibi, preflight sent seven days out #12) is still open.
  • The modules table lists all seven modules in nixosModules, adding holochain-moss-node and sensorica-event-node.
  • The repository tree now points at docs/architecture.md for the file-by-file list instead of repeating an old copy of it.
  • "Documentation site" is ticked (feat: the lab session into main (Moss node service, Sensorica Holoport fleet, Grafana home page, mdBook) #67).
• CHANGELOG. [Unreleased] stopped at chore(example): lock nixos-holochain to main after the stack #23. It now covers every PR merged into main through chore(fleet): pin nixos-holochain to main after the lab merge #68: new options, packages, modules and checks, the *_total counter change, and the fleet renames.
• ADR-017. The "Later record" now includes holoport-install and vmTestHoloportInstall (feat(install): Holoport install script, runbook and a SeaBIOS VM check #61), and the first real Holoport install (feat: the lab session into main (Moss node service, Sensorica Holoport fleet, Grafana home page, mdBook) #67). The UEFI and Secure Boot question is still open.
• Stale names and counts. hosts/edgenode-XX becomes hosts/sensorica-holoport-0N. There are five dashboards, not four, and none is titled "Holochain Fleet" any more (deployment guide, #fleet template). docs/architecture.md lists every file under modules/ and packages/, plus the Moss node's units. The introduction says seven modules, and SECURITY.md lists the bootstrap and Moss modules.
• Workshop material.
  • nixos-rebuild --rollback becomes nixos-rebuild switch --rollback. The 26.05 nixos-rebuild requires an action and only prints its help without one.
  • There is no .#minimal host, so the facilitator guide now names the fleet host instead.
  • Wind Tunnel was presented as the source of the Grafana traffic, but it feeds nothing to Grafana (architecture.md § 3). The traffic comes from the fleet's three hApps.
  • The failure-modes table has the right unit name and the right place for operator keys. It no longer says the lair keystore needs an init step, or that .happ files belong in happs/.
  • colmena apply gets --impure.
  • The pre-flight checklist now checks for the three hApps the fleet actually runs.
  • "August 2026" is removed, because Hardware: fix the workshop date #7 has not fixed the date.
• happs/README.md. The workshop-bundle table named Wind Tunnel and Moss. It is replaced by a pointer to the fleet README, where the hApp list lives.
• holochain-nixos-plan.md. This is the May plan. A "Superseded" note now says the tree wins wherever the two differ. The body is unchanged.

Left alone on purpose:

• ADR-005 and ADR-011 quote hosts/edgenode-0* as Design: Workshop 2026 milestone (PM/Builder binôme record) #1 recorded it.
• docs/images/holochain-fleet-dashboard-2026-08-28.png shows the dashboard before feat(grafana): four dashboards, each titled with its reader's question #65, and replacing it needs a new capture.
• Two Nix comments still name the "Holochain Fleet" dashboard, and modules/sensorica-event-node.nix still says "the other four modules". Those are code, outside this docs pass.
• The 2:15 workshop segment now reads "Observe", for the reason above. The agenda itself waits for the facilitator review with Tibi (Hardware: facilitator guide reviewed with Tibi, preflight sent seven days out #12).

Commands run and what they printed

$ nix run nixpkgs#alejandra -- .
(no .nix file changed)

$ nix flake check --no-build --all-systems --accept-flake-config
exit 0

$ cd examples/sensorica-fleet && nix flake check --no-build --accept-flake-config --override-input nixos-holochain "path:$(git rev-parse --show-toplevel)"
exit 0

$ diff <(cat "$(nix build .#options-doc --print-out-paths)") docs/module-options.md
(no output: in sync)

$ mdbook build
INFO HTML book written to `book` (no warning)

The drvPaths of all 28 root checks were compared between this branch and main at d335034, and 27 are identical. The exception is moss-dashboard, which reads the source tree.

VM tests built

All fourteen vmTest* checks and moss-dashboard were built in one nix build --keep-going run on a KVM laptop. It exited 0 with 14 vm-test-run-* outputs. The VM test derivations are the ones main already passed, and moss-dashboard was built new.

• vmTest, vmTest-0_6, vmTestWithHapp, vmTestWithHapp-0_6, vmTestConductorMetrics-0_6, vmTestGrafana, vmTestGateway, vmTestWindtunnel, vmTestWdocker, vmTestMossNode, vmTestBootstrap, vmTestServices, vmTestServices-noBootstrap, vmTestHoloportInstall: passed

Option reference

• No option was added, changed or removed.
• Options changed, and docs/module-options.md was regenerated with cp "$(nix build .#options-doc --print-out-paths)" docs/module-options.md in the same commit.

Documentation

• docs/ (and README.md, if it says anything about this) updated in this PR, or nothing there describes the changed behaviour.