Close RAIL-HO-WP-0009 declared-state gaps; leave live 6443 prune gated

Make the k3s API tunnel-only (ADR-005), stop declaring Flannel VXLAN
open to Anywhere, tag the base role so firewall can be scoped, and
schedule the Goss declared-vs-live check. CoulombCore sets ufw_manage
false so a converge cannot enable UFW there. T02 still needs operator
approval for make converge-firewall HOST=Railiance01.

Makefile

@@ -131,21 +131,21 @@ backup: ## Backup S1 OS config to /opt/backup/railiance/infra/ (age-encrypted, r
 
 # ---- Ansible ----
 ansible-bootstrap: ## Run base bootstrap play (users, ssh, ufw, sops-agent, custodian-agent)
-	cd ansible && ansible-playbook playbooks/bootstrap.yaml -u admin
+	cd ansible && ansible-playbook playbooks/bootstrap.yaml $(ANSIBLE_USER_FLAG)
 
 provision-custodian-agent: ## Deploy custodian agent SSH key to all managed hosts
 	@python3 -c "import yaml; d=yaml.safe_load(open('ansible/inventory/group_vars/all.yaml')); k=d.get('custodian_agent_pubkey',''); exit(0 if k else 1)" \
 	  || (echo "ERROR: custodian_agent_pubkey is empty. Run: cd ~/the-custodian && make custodian-keygen"; exit 1)
-	cd ansible && ansible-playbook playbooks/custodian-agent.yaml -u $(SSH_USER)
+	cd ansible && ansible-playbook playbooks/custodian-agent.yaml $(ANSIBLE_USER_FLAG)
 
 provision-custodian-agent-host: ## Deploy custodian agent key to one host: make provision-custodian-agent-host HOST=Railiance01
 	@test -n "$(HOST)" || (echo "Usage: make provision-custodian-agent-host HOST=Railiance01"; exit 1)
-	cd ansible && ansible-playbook playbooks/custodian-agent.yaml -u $(SSH_USER) \
+	cd ansible && ansible-playbook playbooks/custodian-agent.yaml $(ANSIBLE_USER_FLAG) \
 	  --limit "$(HOST)"
 
 bootstrap-ssh-ca: ## Deploy OpenBao SSH CA trust + auth_principals: make bootstrap-ssh-ca SSH_CA_PUBKEY=/path/to/ca_user.pub
 	@test -n "$(SSH_CA_PUBKEY)" || (echo "Usage: make bootstrap-ssh-ca SSH_CA_PUBKEY=/path/to/ca_user.pub [HOST=Railiance01]"; exit 1)
-	cd ansible && ansible-playbook playbooks/bootstrap-ssh-ca.yaml -u $(SSH_USER) \
+	cd ansible && ansible-playbook playbooks/bootstrap-ssh-ca.yaml $(ANSIBLE_USER_FLAG) \
 	  -e ssh_ca_pubkey_path="$(SSH_CA_PUBKEY)" \
 	  $(if $(HOST),--limit "$(HOST)",)
 
@@ -167,7 +167,7 @@ deploy-stack: ## Print the full S1→S5 ordered deploy sequence (operator follow
 	@echo "PRE-CONDITIONS"
 	@echo "  [ ] SSH key:    ~/.ssh/id_ops"
 	@echo "  [ ] SOPS key:   ~/.config/sops/age/keys.txt (or SOPS_AGE_KEY)"
-	@echo "  [ ] ops-bridge: bridge up state-hub-coulombcore k3s-api-coulombcore"
+	@echo "  [ ] ops-bridge: bridge up state-hub-coulombcore k3s-api-coulombcore k3s-api-railiance01"
 	@echo ""
 	@echo "S1 — Infrastructure Substrate  (this repo)"
 	@echo "  make tf-plan && make tf-apply    # provision server (skip if exists)"
@@ -222,7 +222,10 @@ remote-set: ## Set origin to your Gitea repo (GITEA/OWNER/REPO vars)
 ANS_DIR := ansible
 INV_SCRIPT := $(ANS_DIR)/inventory_from_yaml.py
 PLAY := $(ANS_DIR)/playbooks/bootstrap.yaml
-SSH_USER ?= admin
+# Inventory servers.yaml sets ansible_user. Override only when needed:
+#   make converge SSH_USER=tegwick
+SSH_USER ?=
+ANSIBLE_USER_FLAG := $(if $(SSH_USER),-u $(SSH_USER),)
 
 # Load your SOPS key for decryption when running playbooks (optional if you use keys.txt)
 export SOPS_AGE_KEY := $(shell cat ~/.config/sops/age/keys.txt 2>/dev/null)
@@ -234,50 +237,66 @@ ansible-help: ## Show common Ansible commands
 	@echo "  make converge                 # run baseline convergence on all hosts"
 	@echo "  make converge-host HOST=web-01# run on a single host"
 	@echo "  make converge-tags TAGS=base  # run only tagged tasks"
+	@echo "  make converge-firewall HOST=Railiance01  # UFW only (RAIL-HO-WP-0009)"
 	@echo "  make converge-check           # dry-run (check mode)"
 	@echo "  make converge-diff            # show config diffs"
+	@echo "  make verify-host HOST=Railiance01"
+	@echo "  make goss-status              # last on-host timer result"
 
 ansible-inventory: ## Print the dynamic inventory Ansible will use
 	cd $(ANS_DIR) && ansible-inventory --list | head -200
 
 ansible-ping: ## Quick connectivity check (SSH + Python availability)
-	cd $(ANS_DIR) && ansible all -u $(SSH_USER) -m ping
+	cd $(ANS_DIR) && ansible all $(ANSIBLE_USER_FLAG) -m ping
 
 status: ## Show live security state of all hosts (UFW, fail2ban, SSH hardening)
 	@echo "=== Connectivity ==="
-	cd $(ANS_DIR) && ansible all -u $(SSH_USER) -m ping
+	cd $(ANS_DIR) && ansible all $(ANSIBLE_USER_FLAG) -m ping
 	@echo "=== UFW ==="
-	cd $(ANS_DIR) && ansible all -u $(SSH_USER) -m shell -a "ufw status" --become
+	cd $(ANS_DIR) && ansible all $(ANSIBLE_USER_FLAG) -m shell -a "ufw status" --become
 	@echo "=== fail2ban ==="
-	cd $(ANS_DIR) && ansible all -u $(SSH_USER) -m shell -a "systemctl is-active fail2ban"
+	cd $(ANS_DIR) && ansible all $(ANSIBLE_USER_FLAG) -m shell -a "systemctl is-active fail2ban"
 	@echo "=== SSH hardening ==="
-	cd $(ANS_DIR) && ansible all -u $(SSH_USER) -m shell -a "grep -iE '^(PermitRootLogin|PasswordAuthentication)' /etc/ssh/sshd_config" --become
+	cd $(ANS_DIR) && ansible all $(ANSIBLE_USER_FLAG) -m shell -a "grep -iE '^(PermitRootLogin|PasswordAuthentication)' /etc/ssh/sshd_config" --become
 	@echo ""
 	@echo "--- Hint: run 'make verify' for a structured pass/fail report ---"
 
 verify: ## Run Goss test suite against all hosts, commit TAP reports — exits non-zero on failure
 	@echo "Running Goss baseline assertions..."
-	@cd $(ANS_DIR) && ansible-playbook playbooks/verify.yaml -u $(SSH_USER) || \
+	@cd $(ANS_DIR) && ansible-playbook playbooks/verify.yaml $(ANSIBLE_USER_FLAG) || \
 	  (echo "One or more assertions FAILED — see reports/ for TAP output." && exit 1)
 	@echo "All assertions passed."
 	@git add reports/ && \
 	  git diff --cached --quiet && echo "No new reports to commit." || \
 	  git commit -m "chore: Goss verification reports $$(date -u +%Y-%m-%dT%H%M%SZ)"
 
+verify-host: ## Run Goss against one host: make verify-host HOST=Railiance01
+	@test -n "$(HOST)" || (echo "Usage: make verify-host HOST=Railiance01"; exit 1)
+	@echo "Running Goss baseline assertions on $(HOST)..."
+	@cd $(ANS_DIR) && ansible-playbook playbooks/verify.yaml $(ANSIBLE_USER_FLAG) -l $(HOST) || \
+	  (echo "One or more assertions FAILED — see reports/ for TAP output." && exit 1)
+
+goss-status: ## Fetch last on-host Goss timer result (fails if FAILED flag present)
+	cd $(ANS_DIR) && ansible-playbook playbooks/goss-status.yaml $(ANSIBLE_USER_FLAG)
+
 converge: ## Converge all hosts to the baseline (idempotent)
-	cd $(ANS_DIR) && ansible-playbook $(PLAY) -u $(SSH_USER)
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG)
 
 converge-host: ## Converge a single host: make converge-host HOST=core-01
 	@test -n "$(HOST)" || (echo "Usage: make converge-host HOST=<name>"; exit 1)
-	cd $(ANS_DIR) && ansible-playbook $(PLAY) -u $(SSH_USER) -l $(HOST)
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG) -l $(HOST)
 
 converge-tags: ## Run only certain tags: make converge-tags TAGS="base,ufw"
 	@test -n "$(TAGS)" || (echo "Usage: make converge-tags TAGS=tag1,tag2"; exit 1)
-	cd $(ANS_DIR) && ansible-playbook $(PLAY) -u $(SSH_USER) --tags "$(TAGS)"
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG) --tags "$(TAGS)"
+
+converge-firewall: ## Apply only UFW tasks: make converge-firewall HOST=Railiance01
+	@test -n "$(HOST)" || (echo "Usage: make converge-firewall HOST=Railiance01"; exit 1)
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG) -l $(HOST) --tags firewall
 
 converge-check: ## Dry-run (no changes), great for previews
-	cd $(ANS_DIR) && ansible-playbook $(PLAY) -u $(SSH_USER) --check
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG) --check
 
 converge-diff: ## Show file/templating diffs while applying changes
-	cd $(ANS_DIR) && ansible-playbook $(PLAY) -u $(SSH_USER) --diff
+	cd $(ANS_DIR) && ansible-playbook $(PLAY) $(ANSIBLE_USER_FLAG) --diff
 

SCOPE.md

@@ -86,12 +86,10 @@ higher layer (Kubernetes, platform, etc.) can run.
   higher-layer and future reef planning. `railiance-hosts` is **superseded** by
   this repo and carries a banner saying so; its retirement is pending in
   `railiance-master`
-- **Open security work**: `RAIL-HO-WP-0009` — the base role declared the k3s API
+- **Open security work**: `RAIL-HO-WP-0009` — k3s API is declared tunnel-only
-  open to Anywhere while the live host was source-restricted by hand, so
+  (ADR-005). Live `Railiance01` still has three public 6443 grants until an
-  converging would have exposed the Kubernetes API. The allowlist is now
+  operator-approved `make converge-firewall HOST=Railiance01`. CoulombCore
-  declarative (`k3s_api_allowed_sources` / `k3s_api_revoked_sources`) but
+  UFW stays unmanaged (`ufw_manage: false`)
-  **has not yet been converged**, so the live host still carries two stale
-  grants to rotated operator addresses
 
 ---
 

WORK-RECORDS.md

@@ -29,7 +29,8 @@
 | task | RAIL-HO-WP-0008-T02 | todo | — | workplans/RAIL-HO-WP-0008-railiance01-resource-and-commercial-evidence.md |
 | task | RAIL-HO-WP-0008-T03 | todo | — | workplans/RAIL-HO-WP-0008-railiance01-resource-and-commercial-evidence.md |
 | task | RAIL-HO-WP-0009-T01 | done | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
-| task | RAIL-HO-WP-0009-T02 | progress | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
+| task | RAIL-HO-WP-0009-T02 | wait | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
-| task | RAIL-HO-WP-0009-T03 | todo | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
+| task | RAIL-HO-WP-0009-T03 | done | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
-| task | RAIL-HO-WP-0009-T04 | todo | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
+| task | RAIL-HO-WP-0009-T04 | done | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
-| task | RAIL-HO-WP-0009-T05 | todo | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
+| task | RAIL-HO-WP-0009-T05 | done | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |
+| task | RAIL-HO-WP-0009-T06 | done | — | workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md |

ansible/inventory/group_vars/all.yaml

@@ -26,29 +26,37 @@ ops_bridge_user: tegwick
 ops_bridge_pubkey: "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIKQmXbAVlEa8dzGx8Hk2S7AITpz6sMWdCN0MeMOzL82u ops-bridge@custodian"
 
 # ---------------------------------------------------------------------------
-# k3s API firewall allowlist (6443/tcp)
+# k3s API firewall (6443/tcp) — tunnel only (ADR-005)
 #
-# The k3s API is NOT world-reachable. Only these sources may connect.
+# Public allowlist is empty. Reach the API over ops-bridge:
-# Added 2026-08-11 after discovering the live host was source-restricted by hand
+#   bridge up k3s-api-railiance01    # local 16444
-# while this repo still declared 6443 open to Anywhere — meaning a convergence
+#   bridge up k3s-api-coulombcore    # local 16443
-# run would have exposed the Kubernetes API to the internet.
 #
-# These are operator egress addresses, not secrets. They are dynamic: when your
+# Revoked addresses are pruned on a firewall-tagged converge so rotated or
-# ISP rotates the lease, MOVE the old entry to k3s_api_revoked_sources rather
+# retired grants do not remain standing. Do not add new public sources here
-# than deleting the line, so convergence prunes the stale grant instead of
+# without amending ADR-005.
-# leaving it standing for whoever the address gets reassigned to.
+k3s_api_allowed_sources: []
-#
-# Durable alternative worth considering: reach the API over the ops-bridge SSH
-# tunnel instead (see railiance-infra/docs/deploy-stack.md, the
-# k3s-api-coulombcore pattern) and allow no public sources at all.
-k3s_api_allowed_sources:
-  - address: "89.244.90.236"
-    comment: "k3s-api-operator-current"
-  - address: "89.244.90.255"
-    comment: "k3s-api-operator-current"
 
 k3s_api_revoked_sources:
+  - address: "89.244.90.248"
+    comment: "hand grant added 2026-08-12/15; retired by ADR-005"
+  - address: "89.244.90.236"
+    comment: "rotated ISP lease; retired by ADR-005"
+  - address: "89.244.90.255"
+    comment: "rotated ISP lease; retired by ADR-005"
   - address: "89.244.90.246"
     comment: "rotated ISP lease, superseded 2026-08-11"
   - address: "85.132.220.102"
-    comment: "historic operator address, retired"
+    comment: "historic operator address; also the 2026-08-15 workstation lease"
+
+# Single-node clusters need no public VXLAN grant. Set peer addresses here
+# only when a second node must exchange Flannel frames (RAIL-BS-WP-0007).
+flannel_vxlan_allowed_sources: []
+
+# HostEurope Nydus agent — provider dashboard, root-password reset, backups.
+# Required by the VPS platform (hosteurope/260308-dependency-nydus.md).
+# Source-restricting it would break the provider; Anywhere is intentional.
+ufw_extra_allowed:
+  - port: "2224"
+    proto: tcp
+    comment: "nydus-ex-api dashboard agent"

ansible/inventory/host_vars/CoulombCore.yml

@@ -1,6 +1,12 @@
 # Host-specific variables for CoulombCore (92.205.130.254)
 # k3s single-node cluster host — HostEurope
 
+# Do not enable or rewrite UFW on this host. Live filter is iptables INPUT
+# DROP with a Plesk-era accept list (UFW status: inactive). Enabling UFW
+# here would take down 80/443 and the rest of the accepted surface unless
+# every live accept is declared first. RAIL-HO-WP-0009-T03.
+ufw_manage: false
+
 # Swapfile (T01)
 swap_size_gb: 4
 swap_swappiness: 10

ansible/inventory_from_yaml.py

@@ -18,9 +18,21 @@ def load_tf_outputs():
         return {}
 
 def load_host_vars(name):
-    """Load host_vars/<name>.yml if it exists."""
+    """Load host_vars/<name>.yml if it exists.
+
+    The inventory script is ansible/inventory_from_yaml.py. Ansible does not
+    auto-load a host_vars directory next to a script inventory, so this has
+    to emit hostvars itself. Look in ansible/inventory/host_vars first (where
+    CoulombCore.yml actually lives), then the unused repo-root path.
+    """
     script_dir = os.path.dirname(__file__)
-    path = os.path.join(script_dir, '..', 'inventory', 'host_vars', f'{name}.yml')
+    candidates = [
+        os.path.join(script_dir, 'inventory', 'host_vars', f'{name}.yml'),
+        os.path.join(script_dir, 'inventory', 'host_vars', f'{name}.yaml'),
+        os.path.join(script_dir, '..', 'inventory', 'host_vars', f'{name}.yml'),
+        os.path.join(script_dir, '..', 'inventory', 'host_vars', f'{name}.yaml'),
+    ]
+    for path in candidates:
         if os.path.exists(path):
             with open(path) as f:
                 return yaml.safe_load(f) or {}

ansible/playbooks/bootstrap.yaml

@@ -5,8 +5,13 @@
     - ../inventory/group_vars/secrets.sops.yaml
   roles:
     - role: base
+      tags: [base]
     - role: sops_agent
+      tags: [sops]
     - role: custodian_agent   # injects ~/.ssh/id_custodian_agent.pub into authorized_keys
+      tags: [custodian_agent]
     - role: swapfile          # provisions swap file (size + swappiness from host_vars)
+      tags: [swap]
     - role: resource_limits   # nproc PAM caps + systemd user slice memory limits
+      tags: [resource_limits]
     # - role: wireguard   # enable if you configure WireGuard variables

ansible/playbooks/goss-status.yaml

@@ -0,0 +1,32 @@
+---
+# Fetch the last on-host Goss timer result. Does not run a new check.
+# Usage: ansible-playbook ansible/playbooks/goss-status.yaml
+#        make goss-status
+
+- hosts: all
+  become: true
+  gather_facts: false
+  tasks:
+    - name: Read last Goss timer status
+      ansible.builtin.slurp:
+        src: /var/lib/railiance/goss/last.status
+      register: goss_status
+      failed_when: false
+
+    - name: Show last Goss timer status
+      ansible.builtin.debug:
+        msg: >-
+          {{ inventory_hostname }}:
+          {{ goss_status.content | default('') | b64decode | trim
+             if goss_status.content is defined
+             else 'no timer result yet' }}
+
+    - name: Fail when the last on-host check reported FAILED
+      ansible.builtin.stat:
+        path: /var/lib/railiance/goss/FAILED
+      register: goss_failed
+
+    - name: Report failed hosts
+      ansible.builtin.fail:
+        msg: "Goss baseline last run failed on {{ inventory_hostname }}"
+      when: goss_failed.stat.exists | default(false)

ansible/roles/base/defaults/main.yml

@@ -1,6 +1,12 @@
 ---
 # Base role defaults.
 
+# When false, this role will not enable or rewrite UFW. Use that for hosts
+# whose live packet filter is not UFW (CoulombCore: iptables INPUT DROP with
+# a Plesk-era accept list). Enabling UFW there is an availability decision,
+# not a side effect of an unrelated converge.
+ufw_manage: true
+
 # Source addresses permitted to reach the k3s API (6443/tcp).
 #
 # WHY THIS EXISTS
@@ -10,21 +16,45 @@
 # config WEAKER than reality: re-running this role would have removed the
 # restriction and exposed the Kubernetes API to the internet. Found 2026-08-11.
 #
-# The allowlist is therefore declared here and converged, not hand-edited.
+# Operator addresses rotate (dynamic ISP leases). An allowlist is a treadmill:
+# each rotation is either an outage or a stale grant to whoever the ISP
+# reassigns the address to. RAIL-HO-WP-0009-T04 therefore keeps this list
+# empty. Reach the API over the ops-bridge SSH tunnel
+# (`k3s-api-railiance01`, local port 16444; `k3s-api-coulombcore`, 16443).
+# See docs/adr/ADR-005-k3s-api-tunnel-only.md.
 #
 # Deliberately empty by default. A host that sets no sources gets NO public
 # access to 6443 — which is the safe failure. SSH (22) is unaffected, so a host
 # converged with an empty list is always recoverable.
 #
-# Set the real values in inventory/group_vars/all.yaml. Each entry:
+# Each entry, if any:
 #   - address: "203.0.113.10"
 #     comment: "k3s-api-operator-workstation"
 k3s_api_allowed_sources: []
 
 # Source addresses whose k3s API access must be REMOVED on convergence.
-#
+# Move an address here when it stops being yours (or when the public
-# Operator addresses rotate (dynamic ISP leases). Without this, every rotation
+# allowlist is retired); convergence then prunes it.
-# leaves a standing grant to an address the ISP has since reassigned to someone
-# else. Move an address here when it stops being yours; convergence then prunes
-# it rather than leaving it to accumulate.
 k3s_api_revoked_sources: []
+
+# Source addresses permitted to send Flannel VXLAN (8472/udp).
+#
+# Empty by default. A single-node cluster does not need a public VXLAN
+# grant; adding an unrestricted 8472/udp allow would expose the pod network
+# to injection. Set this to the other nodes' addresses only when the cluster
+# becomes multi-node (RAIL-BS-WP-0007 / ThreePhoenix HA).
+#
+# Each entry:
+#   - address: "203.0.113.20"
+#     comment: "flannel-vxlan-peer"
+flannel_vxlan_allowed_sources: []
+
+# Extra UFW allows that are not k3s. Used for provider agents that must stay
+# reachable (HostEurope Nydus on 2224/tcp). Empty by default so a Hetzner
+# host does not inherit a HostEurope-only hole.
+#
+# Each entry:
+#   - port: "2224"
+#     proto: tcp
+#     comment: "nydus-ex-api dashboard agent"
+ufw_extra_allowed: []

ansible/roles/base/handlers/main.yml

@@ -1,4 +1,9 @@
 ---
+- name: Restart sshd
+  ansible.builtin.service:
+    name: ssh
+    state: restarted
+
 - name: Restart fail2ban
   ansible.builtin.service:
     name: fail2ban

ansible/roles/base/tasks/main.yml

@@ -1,5 +1,6 @@
 ---
 - name: Ensure base packages
+  tags: [base, packages]
   ansible.builtin.package:
     name:
       - apt-transport-https
@@ -15,6 +16,7 @@
     update_cache: true
 
 - name: Harden SSH
+  tags: [base, ssh]
   ansible.builtin.copy:
     dest: /etc/ssh/sshd_config.d/10-hardening.conf
     owner: root
@@ -24,13 +26,10 @@
       PasswordAuthentication no
       PermitRootLogin no
       PubkeyAuthentication yes
-
+  notify: Restart sshd
-- name: Restart sshd
-  ansible.builtin.service:
-    name: ssh
-    state: restarted
 
 - name: Ensure .ssh directory exists for ops_bridge_user
+  tags: [base, ssh]
   ansible.builtin.file:
     path: "/home/{{ ops_bridge_user | default('tegwick') }}/.ssh"
     state: directory
@@ -39,6 +38,7 @@
     mode: '0700'
 
 - name: Inject ops-bridge public key into authorized_keys
+  tags: [base, ssh]
   ansible.posix.authorized_key:
     user: "{{ ops_bridge_user | default('tegwick') }}"
     key: "{{ ops_bridge_pubkey }}"
@@ -47,27 +47,46 @@
   when: ops_bridge_pubkey is defined and ops_bridge_pubkey | length > 0
 
 - name: Configure UFW default incoming policy
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     state: enabled
     policy: deny
     direction: incoming
+  when: ufw_manage | bool
 
-- name: Allow UFW routing (required for k3s flannel pod networking)
+- name: Allow UFW routing when VXLAN peers are declared
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     policy: allow
     direction: routed
+  when: ufw_manage | bool and (flannel_vxlan_allowed_sources | length > 0)
 
 - name: Allow SSH in UFW
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     rule: allow
     name: OpenSSH
+  when: ufw_manage | bool
 
-# k3s API access is source-restricted. See roles/base/defaults/main.yml for why
+- name: Allow declared extra UFW ports
-# this is declared rather than hand-applied. Order matters below: grants are
+  tags: [base, firewall, ufw]
-# added BEFORE the blanket rule is removed, so convergence never opens a window
+  ansible.builtin.ufw:
-# in which the operator cannot reach the API.
+    rule: allow
+    port: "{{ item.port }}"
+    proto: "{{ item.proto | default('tcp') }}"
+    comment: "{{ item.comment | default('extra-allow') }}"
+  loop: "{{ ufw_extra_allowed }}"
+  loop_control:
+    label: "{{ item.port }}/{{ item.proto | default('tcp') }}"
+  when: ufw_manage | bool
+
+# k3s API access is source-restricted and empty by default (tunnel-only).
+# See roles/base/defaults/main.yml and docs/adr/ADR-005-k3s-api-tunnel-only.md.
+# Order matters: remaining grants (if any) are added BEFORE the blanket rule
+# is removed, so a non-empty allowlist never opens a window without API access.
 
 - name: Allow k3s API from approved operator sources only
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     rule: allow
     port: '6443'
@@ -77,15 +96,19 @@
   loop: "{{ k3s_api_allowed_sources }}"
   loop_control:
     label: "{{ item.address }}"
+  when: ufw_manage | bool
 
 - name: Remove blanket k3s API rule if present (must not be world-reachable)
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     rule: allow
     port: '6443'
     proto: tcp
     delete: true
+  when: ufw_manage | bool
 
 - name: Revoke k3s API access for retired operator sources
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     rule: allow
     port: '6443'
@@ -95,29 +118,49 @@
   loop: "{{ k3s_api_revoked_sources }}"
   loop_control:
     label: "{{ item.address }}"
+  when: ufw_manage | bool
 
 - name: Warn when no operator source is allowed to reach the k3s API
+  tags: [base, firewall, ufw]
   ansible.builtin.debug:
     msg: >-
       k3s_api_allowed_sources is empty, so 6443/tcp is closed to all external
-      sources on this host. This is the safe default, not an error. SSH is
+      sources on this host. Reach the API over the ops-bridge tunnel
-      unaffected and the host remains recoverable. Set the allowlist in
+      (k3s-api-railiance01 on 16444, k3s-api-coulombcore on 16443). SSH is
-      inventory/group_vars/all.yaml to restore API access.
+      unaffected and the host remains recoverable.
-  when: k3s_api_allowed_sources | length == 0
+  when: ufw_manage | bool and (k3s_api_allowed_sources | length == 0)
 
-- name: Allow Flannel VXLAN in UFW
+- name: Allow Flannel VXLAN from declared cluster peers only
+  tags: [base, firewall, ufw]
   ansible.builtin.ufw:
     rule: allow
     port: '8472'
     proto: udp
+    from_ip: "{{ item.address }}"
+    comment: "{{ item.comment | default('flannel-vxlan-peer') }}"
+  loop: "{{ flannel_vxlan_allowed_sources }}"
+  loop_control:
+    label: "{{ item.address }}"
+  when: ufw_manage | bool
+
+- name: Remove blanket Flannel VXLAN rule if present (must not be world-reachable)
+  tags: [base, firewall, ufw]
+  ansible.builtin.ufw:
+    rule: allow
+    port: '8472'
+    proto: udp
+    delete: true
+  when: ufw_manage | bool
 
 - name: Enable fail2ban
+  tags: [base, fail2ban]
   ansible.builtin.service:
     name: fail2ban
     state: started
     enabled: true
 
 - name: Configure fail2ban SSH jail
+  tags: [base, fail2ban]
   ansible.builtin.copy:
     dest: /etc/fail2ban/jail.d/sshd.conf
     owner: root
@@ -134,6 +177,7 @@
   notify: Restart fail2ban
 
 - name: Set HISTCONTROL to ignorespace
+  tags: [base, histcontrol]
   ansible.builtin.copy:
     dest: /etc/profile.d/histcontrol.sh
     owner: root
@@ -143,5 +187,6 @@
       export HISTCONTROL=ignorespace
 
 - name: Set timezone
+  tags: [base, timezone]
   community.general.timezone:
     name: "{{ timezone | default('UTC') }}"

ansible/roles/goss/files/goss-baseline-check.sh

@@ -0,0 +1,59 @@
+#!/bin/bash
+# Recurring declared-vs-live check. Installed by the goss role.
+# Writes TAP + a one-line status; notifies only on pass/fail transitions.
+set -euo pipefail
+
+GOSS_BIN="${GOSS_BIN:-/usr/local/bin/goss}"
+GOSS_FILE="${GOSS_FILE:-/etc/goss/baseline.yaml}"
+STATE_DIR="${STATE_DIR:-/var/lib/railiance/goss}"
+NOTIFY_URL="${RAILIANCE_GOSS_NOTIFY_URL:-}"
+HOST="$(hostname -s)"
+
+mkdir -p "${STATE_DIR}"
+chmod 0755 "${STATE_DIR}"
+
+if [[ ! -x "${GOSS_BIN}" || ! -f "${GOSS_FILE}" ]]; then
+  echo "skip: goss binary or baseline missing" | tee "${STATE_DIR}/last.status"
+  exit 0
+fi
+
+set +e
+"${GOSS_BIN}" -g "${GOSS_FILE}" validate --format tap > "${STATE_DIR}/last.tap"
+rc=$?
+set -e
+
+if [[ "${rc}" -eq 0 ]]; then
+  result=pass
+else
+  result=fail
+fi
+
+printf 'result=%s host=%s ts=%s rc=%s\n' \
+  "${result}" "${HOST}" "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${rc}" \
+  > "${STATE_DIR}/last.status"
+
+prev=""
+if [[ -f "${STATE_DIR}/last.result" ]]; then
+  prev="$(cat "${STATE_DIR}/last.result")"
+fi
+echo "${result}" > "${STATE_DIR}/last.result"
+
+logger -t railiance-goss "baseline ${result} on ${HOST} (rc=${rc})"
+
+if [[ "${result}" == "fail" ]]; then
+  touch "${STATE_DIR}/FAILED"
+else
+  rm -f "${STATE_DIR}/FAILED"
+fi
+
+if [[ -n "${NOTIFY_URL}" && "${result}" != "${prev}" ]]; then
+  payload=$(printf \
+    '{"summary":"Goss baseline %s on %s","event_type":"note","author":"railiance-goss-timer"}' \
+    "${result}" "${HOST}")
+  curl -sS -m 10 -X POST "${NOTIFY_URL}" \
+    -H 'Content-Type: application/json' \
+    -d "${payload}" >/dev/null || \
+    logger -t railiance-goss "notify failed for ${HOST} ${result}"
+fi
+
+exit "${rc}"

ansible/roles/goss/handlers/main.yml

@@ -0,0 +1,4 @@
+---
+- name: Reload systemd daemon
+  ansible.builtin.systemd:
+    daemon_reload: true

ansible/roles/goss/tasks/main.yml

@@ -32,6 +32,39 @@
     group: root
     mode: "0644"
 
+- name: Install recurring Goss check wrapper
+  ansible.builtin.copy:
+    src: goss-baseline-check.sh
+    dest: /usr/local/sbin/goss-baseline-check
+    owner: root
+    group: root
+    mode: "0755"
+
+- name: Install Goss baseline systemd service
+  ansible.builtin.template:
+    src: railiance-goss-baseline.service.j2
+    dest: /etc/systemd/system/railiance-goss-baseline.service
+    owner: root
+    group: root
+    mode: "0644"
+  notify: Reload systemd daemon
+
+- name: Install Goss baseline systemd timer
+  ansible.builtin.template:
+    src: railiance-goss-baseline.timer.j2
+    dest: /etc/systemd/system/railiance-goss-baseline.timer
+    owner: root
+    group: root
+    mode: "0644"
+  notify: Reload systemd daemon
+
+- name: Enable hourly Goss baseline timer
+  ansible.builtin.systemd:
+    name: railiance-goss-baseline.timer
+    enabled: true
+    state: started
+    daemon_reload: true
+
 - name: Run Goss assertions (TAP output)
   ansible.builtin.command:
     cmd: "{{ goss_bin }} -g {{ goss_dir }}/baseline.yaml validate --format tap"

ansible/roles/goss/templates/railiance-goss-baseline.service.j2

@@ -0,0 +1,10 @@
+[Unit]
+Description=Railiance declared-vs-live Goss baseline
+Documentation=file:///etc/goss/baseline.yaml
+
+[Service]
+Type=oneshot
+ExecStart=/usr/local/sbin/goss-baseline-check
+Nice=10
+# The wrapper records TAP even when assertions fail.
+SuccessExitStatus=0 1

ansible/roles/goss/templates/railiance-goss-baseline.timer.j2

@@ -0,0 +1,12 @@
+[Unit]
+Description=Hourly Railiance Goss baseline (RAIL-HO-WP-0009-T05)
+
+[Timer]
+OnBootSec=5min
+OnUnitActiveSec=1h
+RandomizedDelaySec=5min
+Persistent=true
+Unit=railiance-goss-baseline.service
+
+[Install]
+WantedBy=timers.target

ansible/roles/resource_limits/tasks/main.yml

@@ -10,6 +10,7 @@
 #   user_memory_swap_max:        systemd MemorySwapMax (default: 512M)
 
 - name: Set PAM nproc limits
+  tags: [resource_limits]
   ansible.builtin.template:
     src: nproc-limits.conf.j2
     dest: /etc/security/limits.d/60-nproc-{{ resource_limit_user | default('tegwick') }}.conf
@@ -18,6 +19,7 @@
     mode: '0644'
 
 - name: Ensure systemd user slice override directory
+  tags: [resource_limits]
   ansible.builtin.file:
     path: "/etc/systemd/system/user-{{ resource_limit_uid | default(1000) }}.slice.d"
     state: directory
@@ -26,6 +28,7 @@
     mode: '0755'
 
 - name: Set systemd user slice memory limits
+  tags: [resource_limits]
   ansible.builtin.template:
     src: user-slice-limits.conf.j2
     dest: "/etc/systemd/system/user-{{ resource_limit_uid | default(1000) }}.slice.d/limits.conf"

ansible/roles/sops_agent/tasks/main.yml

@@ -1,5 +1,6 @@
 ---
 - name: Install age
+  tags: [sops]
   ansible.builtin.shell: |
     set -euo pipefail
     if ! command -v age >/dev/null; then
@@ -9,12 +10,14 @@
     executable: /bin/bash
 
 - name: Install sops
+  tags: [sops]
   ansible.builtin.get_url:
     url: https://github.com/getsops/sops/releases/download/v3.9.0/sops-v3.9.0.linux.amd64
     dest: /usr/local/bin/sops
     mode: '0755'
 
 - name: Create SOPS age dir
+  tags: [sops]
   ansible.builtin.file:
     path: /root/.config/sops/age
     state: directory

ansible/roles/swapfile/tasks/main.yml

@@ -6,17 +6,20 @@
 #   swap_swappiness: vm.swappiness value (default: 10)
 
 - name: Check if swapfile exists with correct size
+  tags: [swap]
   ansible.builtin.stat:
     path: /swapfile
   register: swapfile_stat
 
 - name: Allocate swapfile (fallocate)
+  tags: [swap]
   ansible.builtin.command:
     cmd: "fallocate -l {{ (swap_size_gb | default(4)) | int }}G /swapfile"
     creates: /swapfile
   when: not swapfile_stat.stat.exists
 
 - name: Set swapfile permissions
+  tags: [swap]
   ansible.builtin.file:
     path: /swapfile
     owner: root
@@ -24,17 +27,20 @@
     mode: '0600'
 
 - name: Format swapfile
+  tags: [swap]
   ansible.builtin.command:
     cmd: mkswap /swapfile
   when: not swapfile_stat.stat.exists
 
 - name: Enable swapfile
+  tags: [swap]
   ansible.builtin.command:
     cmd: swapon /swapfile
   when: not swapfile_stat.stat.exists
   ignore_errors: true  # already active is not an error
 
 - name: Ensure swapfile in /etc/fstab
+  tags: [swap]
   ansible.builtin.lineinfile:
     path: /etc/fstab
     regexp: '^/swapfile'
@@ -42,6 +48,7 @@
     state: present
 
 - name: Set vm.swappiness at runtime
+  tags: [swap]
   ansible.posix.sysctl:
     name: vm.swappiness
     value: "{{ swap_swappiness | default(10) }}"
@@ -49,6 +56,7 @@
     reload: true
 
 - name: Persist vm.swappiness across reboots
+  tags: [swap]
   ansible.builtin.copy:
     dest: /etc/sysctl.d/60-swappiness.conf
     owner: root

docs/adr/ADR-005-k3s-api-tunnel-only.md

@@ -0,0 +1,58 @@
+# ADR-005 — k3s API is tunnel-only
+
+**Status:** Accepted
+**Date:** 2026-08-15
+**Deciders:** implementation of RAIL-HO-WP-0009-T04
+**Workplans:** `RAIL-HO-WP-0009`
+
+---
+
+## Context
+
+Operator addresses on this network rotate with the ISP lease. A public UFW
+allowlist for `6443/tcp` is therefore a treadmill:
+
+- miss a rotation and `kubectl` breaks mid-session
+- leave the old grant standing and it becomes a grant to a stranger
+- hand-add the new address and the declaration drifts again
+
+That cycle produced this workplan. The live allowlist drifted by hand *during
+the session that was fixing allowlist drift*, and again before the next
+session (`89.244.90.248` appeared undeclared). On 2026-08-15 the workstation
+egress address was `85.132.220.102` — already on the revoked list as a
+"historic" operator address.
+
+`docs/deploy-stack.md` already documents API access over the ops-bridge SSH
+tunnel for CoulombCore (`k3s-api-coulombcore`, local port 16443). The same
+tunnel already exists for Railiance01 (`k3s-api-railiance01`, local port
+16444). SSH itself stays public, so the host remains recoverable.
+
+## Decision
+
+The public k3s API allowlist is **empty**. Operator and agent `kubectl` access
+uses the ops-bridge tunnels:
+
+| Cluster     | Tunnel                  | Local port | Remote |
+|-------------|-------------------------|------------|--------|
+| CoulombCore | `k3s-api-coulombcore`   | 16443      | 6443   |
+| Railiance01 | `k3s-api-railiance01`   | 16444      | 6443   |
+
+```bash
+bridge up k3s-api-railiance01
+# kubeconfig server: https://127.0.0.1:16444
+```
+
+Trade: every operator `kubectl` action depends on ops-bridge. That is accepted.
+A rotating public allowlist is the worse dependency.
+
+Emergency break-glass remains SSH: `ssh railiance01 -- sudo k3s kubectl …`.
+Do not re-open `6443/tcp` to Anywhere.
+
+## Consequences
+
+- `k3s_api_allowed_sources` stays `[]`.
+- Former public grants live in `k3s_api_revoked_sources` so a firewall-tagged
+  converge deletes them.
+- Goss asserts the 6443 allowlist size is exactly the declared length (zero)
+  and that no revoked address remains.
+- Amending this ADR is required before adding any new public 6443 source.

docs/conformance-loop.md

@@ -0,0 +1,61 @@
+# Declared-vs-live conformance loop
+
+This is the S1 slice of the unowned **Q7 Governance and Change Management**
+gap recorded in `railiance-platform/ArchitectureBlueprint.md` §5.3. The
+family owner of the loop is **`railiance-master`**. This repo owns the host
+check and the first reaction path; it does not own rapp/rail/reef schema
+conformance (that is 10c, blocked on the family declaration validator).
+
+## What runs
+
+| Piece | Where | Cadence |
+| --- | --- | --- |
+| Rendered Goss baseline | `/etc/goss/baseline.yaml` from `goss/baseline.yaml.j2` | every `make verify` |
+| On-host timer | `railiance-goss-baseline.timer` | hourly, plus 5 minutes after boot |
+| Wrapper | `/usr/local/sbin/goss-baseline-check` | writes `/var/lib/railiance/goss/` |
+| Operator collect | `make goss-status` | on demand |
+| Operator run | `make verify` / `make verify-host HOST=…` | on demand; commits TAP to `reports/` |
+
+The firewall assertions are generated from the same inventory lists that
+converge UFW (`k3s_api_allowed_sources`, `k3s_api_revoked_sources`,
+`flannel_vxlan_allowed_sources`, `ufw_extra_allowed`). They assert:
+
+- each declared source is present
+- the **count** of non-Anywhere 6443/8472 allows equals the declared length
+  (extra hand grants fail)
+- no `ALLOW Anywhere` on 6443 or 8472
+- every revoked 6443 source is gone
+
+That is the check that would have caught `89.244.90.248` appearing on
+Railiance01 after T01.
+
+## Where failures go
+
+1. **journald** — `logger -t railiance-goss`
+2. **on-host flag** — `/var/lib/railiance/goss/FAILED` and `last.status`
+3. **optional State Hub POST** — set `RAILIANCE_GOSS_NOTIFY_URL` on the host
+   to a State Hub `/progress/` endpoint. The wrapper notifies only on a
+   pass↔fail transition, so a known-fail host (CoulombCore, UFW inactive)
+   does not spam.
+4. **git TAP** — `make verify` still commits `reports/goss-<host>-<ts>.tap`
+
+`make goss-status` fails the play if any host still has the `FAILED` flag.
+
+## Routing to Q7
+
+`railiance-master` owns the conformance loop as a family capability
+(ArchitectureBlueprint §5.3; RAILIANCE-WP-0016 item 10b). This file is the
+handoff:
+
+- S1 check: **exists and is scheduled** in this repo
+- S1 reaction: journald + flag + optional progress event
+- Family reaction (page a human, open a task, block a rollout): **not here**
+
+Do not build a second alert path in `railiance-telemetry` for the same
+"check failed → someone sees it" plumbing. Item 9b should reuse this.
+
+## Known expected fail
+
+`CoulombCore` has UFW inactive. The baseline asserts `Status: active`. The
+timer will fail there until an explicit decision enables UFW or the host is
+removed from the verify inventory. That failure is evidence, not noise.

docs/convergence.md

@@ -9,7 +9,9 @@ This process is called **convergence**.
 When you run `make converge`, Ansible connects to all declared hosts and applies baseline roles:
 
 - **User setup** → ensures the `admin` user exists with your SSH key and passwordless sudo
-- **Firewall** → configures `ufw` with sensible defaults (deny incoming, allow SSH)
+- **Firewall** → configures `ufw` with deny incoming, SSH, declared extras
+  (Nydus), and a tunnel-only k3s API (ADR-005). Use
+  `make converge-firewall HOST=Railiance01` to apply only those tasks.
 - **Hardening** → basic SSH daemon hardening, disable root login, disable password auth
 - **Tooling** → installs essential packages (htop, vim, git, curl, fail2ban, etc.)
 - **SOPS agent** → ensures decryption tooling (`age`, `sops`) is available on the host

docs/deploy-stack.md

@@ -120,7 +120,8 @@ After S2 is up, establish the persistent tunnels from the workstation:
 ```bash
 bridge up state-hub-coulombcore       # state-hub HTTP (port 18000 remote)
 bridge up state-hub-mcp-coulombcore   # state-hub MCP (port 18001 remote)
-bridge up k3s-api-coulombcore         # k3s API (port 16443 local)
+bridge up k3s-api-coulombcore         # CoulombCore k3s API (port 16443 local)
+bridge up k3s-api-railiance01         # Railiance01 k3s API (port 16444 local)
 ```
 
 Verify: `bridge status` shows all three connected.

docs/evidence/RAIL-HO-WP-0009-T03-declared-vs-live-audit.md

@@ -0,0 +1,141 @@
+# RAIL-HO-WP-0009-T03 — declared-vs-live audit
+
+Audited 2026-08-15 from the workstation against `Railiance01`
+(`92.205.62.239`) and `CoulombCore` (`92.205.130.254`). SSH as `tegwick`.
+This is a record of what was found; absence of drift is recorded too.
+
+## Method
+
+Compared `ansible/roles/base/tasks/main.yml`, sibling bootstrap roles, and
+`spec/server-baseline.yaml` to live `ufw status`, `sshd -T`, fail2ban,
+sudoers, listening sockets, systemd user-slice, nproc, swappiness, and
+timezone.
+
+## Matches declaration
+
+| Surface | Railiance01 | CoulombCore |
+| --- | --- | --- |
+| sshd drop-in `10-hardening.conf` | present; root/password off, pubkey on | same |
+| `sshd -T` effective | matches drop-in | matches drop-in |
+| fail2ban `sshd` jail | active | active |
+| `/etc/profile.d/histcontrol.sh` | `ignorespace` | `ignorespace` |
+| sudoers `tegwick` NOPASSWD | present | present |
+| OpenSSH UFW | ALLOW Anywhere | UFW inactive (see below) |
+
+## Divergences
+
+### 1. k3s API allowlist (Railiance01) — same defect class as T01
+
+Live 6443 grants on 2026-08-15:
+
+| From | Declared? |
+| --- | --- |
+| `89.244.90.236` | was declared; now revoked by ADR-005 |
+| `89.244.90.255` | was declared; now revoked by ADR-005 |
+| `89.244.90.248` | **never declared** — third hand grant |
+
+No blanket `Anywhere` rule. Stale grants from T02 (`.246`, `.102`) are gone.
+
+`.248` appeared after the 2026-08-12 session that was already fixing
+allowlist drift. Goss as of T01 would not have failed: it asserted declared
+sources present, not that the live set equalled the declared set.
+
+Workstation egress on this audit was `85.132.220.102` — already on the
+revoked list. That is the allowlist treadmill in one line.
+
+### 2. Flannel VXLAN declaration weaker than needed (T06)
+
+`8472/udp` listens on `0.0.0.0` on both hosts. Railiance01 has **no** UFW
+rule for it (default deny incoming covers it). The role used to declare
+`8472/udp` ALLOW with no source restriction, so a converge would have
+*opened* VXLAN to the internet. Fixed: empty `flannel_vxlan_allowed_sources`,
+blanket rule deleted if present.
+
+### 3. Port 2224/tcp Anywhere — intended, now declared
+
+Both hosts run `nydus-ex-api` on `0.0.0.0:2224`. Railiance01 UFW allows it
+Anywhere with comment `nydus-ex-api dashboard agent`. This is the HostEurope
+Nydus agent (dashboard, root-password reset, backups); see
+`hosteurope/260308-dependency-nydus.md`. Removing it would break the
+provider. Declared as `ufw_extra_allowed`. Source-restricting it is not
+possible without HostEurope's egress list.
+
+### 4. CoulombCore: UFW inactive, but not "no firewall"
+
+`ufw status` → `Status: inactive`. Live filter is `iptables` `INPUT DROP`
+with a Plesk-era accept list: 22, 80, 443, 2224, mail, 3306, 5432, 53,
+445/139, and **tcp 49152-65535**. 6443 and 8472 are **not** in that accept
+list, so they are dropped even with UFW off.
+
+Converging the base role would enable UFW, allow only SSH + (now empty) k3s
++ nydus, and take down 80/443 and the rest of the accepted surface.
+`ufw_manage: false` on CoulombCore prevents that. Enabling UFW there is its
+own decision and needs a declared accept list first.
+
+### 5. Resource limits and swap — applied on CoulombCore, never on Railiance01
+
+| | Railiance01 | CoulombCore |
+| --- | --- | --- |
+| `user-1000.slice` MemoryMax | absent | `1500M` / swap `512M` (Ansible) |
+| PAM nproc | no `limits.d` files | `60-nproc-tegwick.conf` |
+| swappiness | 60 (kernel default) | 10 |
+| timezone | `Etc/UTC` | `Etc/UTC` |
+
+The user-slice cap can OOM agent workloads (T02). Do **not** apply it to
+Railiance01 without a separate decision. Role tags now isolate it
+(`resource_limits`, `swap`, `timezone`).
+
+`inventory/group_vars/all.yaml` sets `timezone: Europe/Berlin` but that file
+is not loaded by the Ansible inventory script. The role default is UTC, which
+matches live. Leave it; do not start loading that file as a side effect.
+
+### 6. Inventory host_vars path bug
+
+`ansible/inventory_from_yaml.py` loaded `inventory/host_vars/<name>.yml`
+(does not exist). The real file is `ansible/inventory/host_vars/CoulombCore.yml`.
+Ansible does not auto-load host_vars next to a script inventory, so
+`ufw_manage`, swap, and resource-limit overrides would not reach the play.
+Fixed: the script now reads `ansible/inventory/host_vars/` first.
+
+### 7. sshd restart was an unconditioned task
+
+`Restart sshd` ran on every converge, which is why `--check` reported an
+sshd restart among the 11 changes. It is now a handler notified only when
+the hardening drop-in changes.
+
+### 8. Undeclared listening sockets (not opened in UFW)
+
+On Railiance01, not in this role, UFW default-deny so not currently public:
+
+- `*:6443` k3s-server (the API this workplan closes)
+- `92.205.62.239:2379` / `:2380` etcd on the public address — S2 concern
+- `*:10250` kubelet — S2 concern
+- `0.0.0.0:18081` Python reverse proxy to `127.0.0.1:18765` (issue-core),
+  started 2026-07-26. No UFW rule. Out of scope here; record only.
+
+On CoulombCore, `*:6443` and `0.0.0.0:8472` listen; iptables INPUT DROP
+without an accept for those ports is what currently protects them, not UFW.
+
+### 9. Routed policy
+
+Railiance01 UFW default routed is **deny**. The role used to set routed
+**allow** for Flannel. Routed allow is now applied only when
+`flannel_vxlan_allowed_sources` is non-empty (multi-node). Single-node
+converge will not flip it.
+
+## Decisions taken in this task
+
+- Declare Nydus 2224; do not remove it.
+- Do not enable UFW on CoulombCore (`ufw_manage: false`).
+- Do not apply Railiance01 resource limits / swappiness / timezone as part
+  of this workplan.
+- Tag the base role and sibling bootstrap roles so firewall can be scoped.
+- Fix the host_vars loader.
+
+## Still operator decisions (not this workplan)
+
+- Enable UFW on CoulombCore, with a declared accept list that includes 80/443.
+- Apply or drop the declared user-slice memory cap on Railiance01.
+- Whether etcd/kubelet listening on the public address should be rebound
+  (`railiance-cluster`).
+- What the `0.0.0.0:18081` issue-core proxy is for, and whether it stays.

docs/k3s-api-access.md

@@ -0,0 +1,48 @@
+# k3s API access
+
+The Kubernetes API is not on the public internet. See
+[ADR-005](adr/ADR-005-k3s-api-tunnel-only.md).
+
+## Day-to-day (workstation)
+
+```bash
+bridge up k3s-api-railiance01     # https://127.0.0.1:16444  →  Railiance01 :6443
+bridge up k3s-api-coulombcore     # https://127.0.0.1:16443  →  CoulombCore :6443
+bridge status
+```
+
+Point the kubeconfig `server` at the matching local port. The default
+`~/.kube/config` and `~/.kube/config-hosteurope` currently both target
+`https://127.0.0.1:16443` (CoulombCore). For Railiance01 use 16444.
+
+```bash
+# Railiance01
+KUBECONFIG=~/.kube/config-railiance01 kubectl get nodes
+# or, once:
+#   kubectl --kubeconfig ~/.kube/config-railiance01 config set-cluster default --server=https://127.0.0.1:16444
+```
+
+SSH is unaffected. If the tunnel is down, recover over SSH:
+
+```bash
+ssh tegwick@92.205.62.239 -- sudo k3s kubectl get nodes
+```
+
+## What not to do
+
+- Do not `ufw allow 6443/tcp` from the current public IP.
+- Do not add addresses to `k3s_api_allowed_sources` without amending ADR-005.
+- When an old grant must die, put it in `k3s_api_revoked_sources` and run
+  `make converge-firewall HOST=Railiance01`.
+
+## Converge the firewall only
+
+The base role is tagged. Firewall changes no longer drag sshd restarts or
+user-slice memory caps with them:
+
+```bash
+make converge-firewall HOST=Railiance01
+```
+
+CoulombCore sets `ufw_manage: false`. That host's live filter is iptables, not
+UFW; this target will skip UFW tasks there.

docs/verification.md

@@ -10,11 +10,13 @@ a reproducible, CI-friendly pass/fail verdict.
 of every managed node. It covers:
 
 - **Firewall** — UFW active, default deny inbound, required ports allowed
-  (SSH 22/tcp, Flannel VXLAN 8472/udp). The k3s API (6443/tcp) is
+  (SSH 22/tcp; HostEurope Nydus 2224/tcp). The k3s API (6443/tcp) is
-  **source-restricted**, not world-open: only addresses in
+  **tunnel-only** (ADR-005): `k3s_api_allowed_sources` is empty, and
-  `k3s_api_allowed_sources` may reach it, and `k3s_api_revoked_sources` is
+  `k3s_api_revoked_sources` is pruned on a firewall-tagged converge. Flannel
-  pruned on convergence. A host with an empty allowlist has 6443 closed to all
+  VXLAN (8472/udp) is omitted while the cluster is single-node; peer
-  external sources — the safe failure, recoverable over SSH.
+  addresses go in `flannel_vxlan_allowed_sources` when a second node appears.
+  Hosts with `ufw_manage: false` (CoulombCore) are not rewritten. A host with
+  an empty 6443 allowlist is recoverable over SSH.
 - **SSH daemon** — root login disabled, password auth disabled, pubkey auth enabled
 - **Services** — ufw, fail2ban, ssh.socket enabled and running
 - **Packages** — ufw, fail2ban, git, curl, vim, htop (age and sops installed as binaries)
@@ -56,7 +58,8 @@ you need a structured, automatable check.
 
 ## Goss test file
 
-`goss/baseline.yaml` contains one Goss assertion per spec item. The mapping is:
+`goss/baseline.yaml.j2` is rendered per host from the same inventory lists
+that converge UFW. The mapping is:
 
 | spec section | Goss resource |
 |---|---|
@@ -73,8 +76,11 @@ you need a structured, automatable check.
 
 1. Add the desired state to `spec/server-baseline.yaml`
 2. Add the Ansible task to `ansible/roles/base/tasks/main.yml`
-3. Add the Goss assertion to `goss/baseline.yaml`
+3. Add the Goss assertion to `goss/baseline.yaml.j2`
-4. Run `make converge && make verify` to confirm
+4. Run `make converge-firewall` and `make verify-host HOST=…` to confirm
+
+An hourly on-host timer (`railiance-goss-baseline.timer`) reruns the last
+rendered baseline. See `docs/conformance-loop.md`.
 
 ## Reports
 

goss/baseline.yaml.j2

@@ -4,9 +4,10 @@
 #
 # THIS IS A TEMPLATE. It is rendered per host from inventory variables, so the
 # firewall assertions below are generated from the SAME declaration that
-# converges the host (k3s_api_allowed_sources in group_vars/all.yaml). That is
+# converges the host (k3s_api_allowed_sources, flannel_vxlan_allowed_sources,
-# deliberate: a hand-written assertion drifts from the declaration it is meant
+# ufw_extra_allowed in group_vars). That is deliberate: a hand-written
-# to check, which is exactly how RAIL-HO-WP-0009 happened.
+# assertion drifts from the declaration it is meant to check, which is exactly
+# how RAIL-HO-WP-0009 happened.
 
 package:
   ufw:
@@ -60,16 +61,33 @@ command:
 {% for src in k3s_api_allowed_sources | default([]) %}
       - '/6443\/tcp\s+ALLOW\s+{{ src.address | regex_escape }}/'
 {% endfor %}
-      - /8472\/udp.*ALLOW/
+{% for src in flannel_vxlan_allowed_sources | default([]) %}
+      - '/8472\/udp\s+ALLOW\s+{{ src.address | regex_escape }}/'
+{% endfor %}
+{% for rule in ufw_extra_allowed | default([]) %}
+      - '/{{ rule.port }}\/{{ rule.proto | default("tcp") }}\s+ALLOW/'
+{% endfor %}
 
-  # The k3s API must never be reachable from Anywhere. The previous assertion
+  # Exact allowlist size: extra hand grants must fail, not only missing ones.
-  # matched /6443\/tcp.*ALLOW/, which passes identically whether the API is
+  # The previous assertion matched /6443\/tcp.*ALLOW/, which passes identically
-  # restricted to one operator address or open to the entire internet — it
+  # whether the API is restricted to one operator address or open to the entire
-  # asserted that the port was allowed, not from whom.
+  # internet — it asserted that the port was allowed, not from whom.
   "ufw status | grep -Ec '6443/tcp[[:space:]]+ALLOW[[:space:]]+Anywhere' || true":
     exit-status: 0
     stdout:
       - "0"
+  "ufw status | grep -E '6443/tcp[[:space:]]+ALLOW' | grep -vc 'Anywhere' || true":
+    exit-status: 0
+    stdout:
+      - "{{ k3s_api_allowed_sources | default([]) | length }}"
+  "ufw status | grep -Ec '8472/udp[[:space:]]+ALLOW[[:space:]]+Anywhere' || true":
+    exit-status: 0
+    stdout:
+      - "0"
+  "ufw status | grep -E '8472/udp[[:space:]]+ALLOW' | grep -vc 'Anywhere' || true":
+    exit-status: 0
+    stdout:
+      - "{{ flannel_vxlan_allowed_sources | default([]) | length }}"
 
 {% for src in k3s_api_revoked_sources | default([]) %}
   # Revoked operator source must not retain access: {{ src.comment | default('') }}

spec/server-baseline.yaml

@@ -6,7 +6,7 @@
 # When you change something here, update the Ansible roles AND the Goss tests.
 # Format: human-readable YAML, kept technology-neutral.
 
-version: "1.0"
+version: "1.1"
 applies_to: all  # override per node group if needed
 
 # ---------------------------------------------------------------------------
@@ -17,6 +17,8 @@ firewall:
   status: active
   default_incoming: deny
   default_outgoing: allow
+  # Hosts with ufw_manage: false (CoulombCore) keep their existing packet
+  # filter. Do not enable UFW there as a side effect of an unrelated converge.
   rules:
     - name: OpenSSH   # UFW app name; resolves to 22/tcp
       action: allow
@@ -24,10 +26,17 @@ firewall:
       port: 6443
       proto: tcp
       action: allow
+      sources: []   # tunnel-only (ADR-005); public allowlist stays empty
     - name: flannel-vxlan
       port: 8472
       proto: udp
       action: allow
+      sources: []   # omit while single-node; peer addresses only when multi-node
+    - name: nydus-ex-api
+      port: 2224
+      proto: tcp
+      action: allow
+      sources: anywhere   # HostEurope provider agent; required by the VPS platform
 
 # ---------------------------------------------------------------------------
 # SSH daemon

workplans/RAIL-HO-WP-0009-firewall-declared-state-and-api-exposure.md

@@ -8,7 +8,7 @@ status: active
 owner: codex
 topic_slug: railiance
 created: "2026-08-12"
-updated: "2026-08-12"
+updated: "2026-08-15"
 related_repos:
   - railiance-cluster
   - railiance-platform
@@ -98,7 +98,7 @@ session (`89.244.90.236` granted; `kubectl` verified, node Ready v1.35.1+k3s1).
 
 ```task
 id: RAIL-HO-WP-0009-T02
-status: progress
+status: wait
 priority: high
 state_hub_task_id: "7d91dfc2-481b-4943-90e9-bb1b814a23dc"
 ```
@@ -141,9 +141,22 @@ That is worth recording as evidence rather than as a footnote: the allowlist
 drifted again, by hand, *during the very session that was fixing allowlist
 drift*. It is the strongest available argument for T04.
 
+**Progress 2026-08-15.** Tags now exist (`make converge-firewall HOST=Railiance01`).
+Live 6443 grants on this date: `.236`, `.255`, and an undeclared `.248`. T04
+emptied the declared allowlist, so this task's remaining work is the
+operator-approved prune of those three grants. SSH stays; kubectl goes via
+`k3s-api-railiance01` (`:16444`). CoulombCore is not touched
+(`ufw_manage: false`).
+
+Waiting on operator approval. Command:
+
+```bash
+make converge-firewall HOST=Railiance01
+```
+
 ```task
 id: RAIL-HO-WP-0009-T03
-status: todo
+status: done
 priority: high
 state_hub_task_id: "3e835b96-736c-458c-91eb-04bfd4c7e0e7"
 ```
@@ -179,9 +192,24 @@ ever applied — an sshd restart, `MemoryMax=1500M` / `MemorySwapMax=512M` on
 cap in particular could OOM running agent workloads. The base role has **no
 tags**, so convergence cannot be scoped. Add tags as part of this task.
 
+**Done 2026-08-15.** Full audit:
+`docs/evidence/RAIL-HO-WP-0009-T03-declared-vs-live-audit.md`.
+
+- sshd / fail2ban / HISTCONTROL / sudoers match on both hosts.
+- 2224/tcp is HostEurope Nydus; declared as `ufw_extra_allowed`.
+- CoulombCore UFW is inactive; live filter is iptables INPUT DROP (6443/8472
+  not accepted). Set `ufw_manage: false` so a converge cannot enable UFW there.
+- Base role and sibling bootstrap roles now have tags (`firewall`, `ssh`,
+  `packages`, `fail2ban`, `swap`, `resource_limits`, `sops`). sshd restart is
+  a handler. Routed-allow applies only when VXLAN peers are declared.
+- Inventory script now loads `ansible/inventory/host_vars/` (it was looking at
+  a path that does not exist).
+- Not in this workplan: Railiance01 resource limits, CoulombCore UFW enable,
+  etcd/kubelet bind addresses, the `0.0.0.0:18081` issue-core proxy.
+
 ```task
 id: RAIL-HO-WP-0009-T04
-status: todo
+status: done
 priority: medium
 state_hub_task_id: "908630e8-e245-47f5-a060-3949250f522c"
 ```
@@ -195,10 +223,18 @@ the same pattern to `railiance01` and reduce the public allowlist to nothing.
 Decide explicitly rather than by default: this trades convenience for exposure,
 and the tunnel becomes a dependency of every operator action.
 
+**Done 2026-08-15.** Decision recorded in
+`docs/adr/ADR-005-k3s-api-tunnel-only.md`. `k3s_api_allowed_sources` is empty;
+former public grants (including the undeclared `.248` and the 2026-08-15
+workstation lease `.102`) are in `k3s_api_revoked_sources`. Operator access
+is `bridge up k3s-api-railiance01` (local 16444). Live prune of the three
+remaining grants is T02.
+
 ```task
 id: RAIL-HO-WP-0009-T06
-status: todo
+status: done
 priority: high
+state_hub_task_id: "764adf9e-cb60-4edd-86bf-f0c170e65cd2"
 ```
 
 **Flannel VXLAN is declared open to Anywhere.** Found 2026-08-12 by the newly
@@ -219,9 +255,15 @@ Note the sequencing risk: `RAIL-BS-WP-0007` (ThreePhoenix HA, multi-node) will
 need working VXLAN between nodes. Resolve this before that lands, or the fix
 will be made under delivery pressure.
 
+**Done 2026-08-15.** `flannel_vxlan_allowed_sources` defaults to empty; the
+role no longer adds a blanket 8472/udp allow and deletes one if present.
+Goss asserts 8472 is never `ALLOW Anywhere` and that the non-Anywhere count
+equals the declared peer list. When ThreePhoenix adds a second node, put its
+address in that list before converging firewall.
+
 ```task
 id: RAIL-HO-WP-0009-T05
-status: todo
+status: done
 priority: medium
 state_hub_task_id: "81af8470-286c-440a-bbe7-b47a17ee6750"
 ```
@@ -251,6 +293,21 @@ Result on `Railiance01`: 32 assertions, 31 pass, one genuine failure (T06). What
 remains for this task is the *loop* — scheduling it and routing failures
 somewhere they are seen.
 
+**Done 2026-08-15.** The check now also asserts exact allowlist *size*, so an
+extra hand grant fails. The loop:
+
+- hourly `railiance-goss-baseline.timer` on each host after `make verify`
+- TAP + `/var/lib/railiance/goss/last.status` + `FAILED` flag
+- optional State Hub POST on pass↔fail via `RAILIANCE_GOSS_NOTIFY_URL`
+- `make goss-status` / `make verify-host`
+
+Routed to `railiance-master` (ArchitectureBlueprint §5.3, Q7 owner) in
+`docs/conformance-loop.md`. Family-level reaction (page a human, block a
+rollout) is not this repo.
+
 ## Outcome
 
-Pending. T01 done; the live host is reachable but not yet converged.
+T01, T03, T04, T05, T06 done. T02 waits on operator approval to run
+`make converge-firewall HOST=Railiance01`, which will delete the three live
+6443 grants and make the host match ADR-005. CoulombCore UFW is intentionally
+not enabled.