NK-WP-0058-T07: conventions lint, server-side applies, one cluster route
All checks were successful
Bounded-execution contracts / contracts (push) Successful in 1m25s
CI Smoke / host-smoke (push) Successful in 0s
CI Smoke / container-smoke (push) Successful in 1s

tools/conventions/lint.py (make conventions-lint, contract suite): server-side-apply, one-wait-condition,
one-cluster-route, tool-numbered-receipts, managing-controller (Argo-owned paths from railiance-platform),
with reasoned exceptions. 21 runbook applies made server-side. portal-client-rollout.py (and so
keycape-client.py and the rollout tools) and identity-provisioner-reconcile.py use NK_KUBECONFIG,
default ~/.kube/config-railiance01, without a second kubeconfig or --server override.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
This commit is contained in:
tegwick 2026-11-06 00:34:00 +01:00
parent 80d280a340
commit cd265b84a4
13 changed files with 226 additions and 28 deletions

View file

@ -212,6 +212,9 @@ bounded-execution-admission-check: ## Validate the example admissions (read-only
execution-attribution-receipt-test: ## Receipt schema v0.1 validator; fixtures match the executor (NK-WP-0056-T15)
python3 -m pytest tools/execution-attribution-receipt/tests
conventions-lint: ## Runbook and tooling conventions from the 2026-11-04/05 pilot (NK-WP-0058-T07), read-only
python3 tools/conventions/lint.py
contract-test: ## Bounded-execution contract suite against the owner checkouts next to this repo; a skip fails (NK-WP-0058-T02)
bash tools/contract-tests/run.sh

View file

@ -22,7 +22,7 @@ native scope/auth/namespace denials and reader revocation. It invokes this
repository's `sso-mfa/k8s/keycape/approval-clients-rollout.py` for image/config
replacement and service acceptance. Do not run the platform child directly.
On this workstation the verified Railiance kubeconfig is
`/home/worsch/.kube/config-railiance01`; the default config targets another endpoint.
`~/.kube/config-railiance01` (`NK_KUBECONFIG` overrides it); the default config targets another endpoint.
Use `activate` only for a first provision with absent custody and delivery.
`resume-activate --prior-receipt <metadata receipt>` requires a completed compatible

View file

@ -25,10 +25,10 @@ chmod +x create-secrets.sh
./create-secrets.sh
# 2. Apply manifests (order matters)
kubectl apply -f pvc.yaml
kubectl apply -f configmap.yaml
kubectl apply -f deployment.yaml
kubectl apply -f ingress.yaml
kubectl apply --server-side -f pvc.yaml
kubectl apply --server-side -f configmap.yaml
kubectl apply --server-side -f deployment.yaml
kubectl apply --server-side -f ingress.yaml
# 3. Wait for pod to be ready
# The startup probe allows 90 s for the initial LLDAP connection.

View file

@ -30,7 +30,7 @@ is not in its file. Applying the whole file would therefore be a broader change
K="kubectl --kubeconfig $HOME/.kube/config-railiance01"
cd ~/approval-engine && git pull --ff-only
python3 -c "import yaml,sys; yaml.safe_dump([d for d in yaml.safe_load_all(open('deploy/networkpolicies.yaml')) if d and d['metadata']['name']=='approval-engine-bounded-executor'][0], sys.stdout)" > /tmp/ae-bx.yaml
$K apply --dry-run=server -f /tmp/ae-bx.yaml && $K apply -f /tmp/ae-bx.yaml
$K apply --server-side --dry-run=server -f /tmp/ae-bx.yaml && $K apply --server-side -f /tmp/ae-bx.yaml
```
audit-core: run steps 1 and 2 of audit-core `docs/bounded-executor-receipts.md`, "Founder runbook —
@ -86,7 +86,7 @@ C=$(mktemp); python3 sso-mfa/k8s/keycape/keycape-client.py client-from-case secu
B=$HOME/.local/share/key-cape/verified-bin/e2b5fa1 # check-config binary from the deployed image (T09 runbook)
test -x "$B/keycape" || { mkdir -p "$B"; cid=$(docker create forgejo.coulomb.social/coulomb/key-cape@sha256:9e082dca78f65a4af3e92d3de3705b99e49ca9b298d9562ad0d37511c4306f8d); docker cp "$cid:/keycape" "$B/keycape"; docker rm "$cid"; }
# keycape-client.py uses its built-in kubeconfig through 127.0.0.1:16444, exactly as in the T09 runbook;
# keycape-client.py uses the same route as $K (NK_KUBECONFIG, default ~/.kube/config-railiance01; NK-WP-0058-T07);
# it pins the cluster identity and refuses another cluster.
$K -n sso rollout pause deployment/keycape
$K -n sso patch deployment keycape --type=json -p '[

View file

@ -8,6 +8,7 @@ import copy
import importlib.util
import json
from pathlib import Path
import os
import subprocess
import yaml
@ -91,9 +92,13 @@ def replacement(secret):
return base64.b64encode(updated.encode()).decode(), True
# The runbooks' cluster route (NK-WP-0058-T07): NK_KUBECONFIG, else the admin kubeconfig they name. No second file and
# no --server override: until 2026-11-06 this used a second kubeconfig file with a server override to the same API.
KUBECONFIG = os.environ.get('NK_KUBECONFIG', os.path.expanduser('~/.kube/config-railiance01'))
def kubectl(arguments, payload=None):
args = ['kubectl', '--kubeconfig', '/home/worsch/.kube/config-hosteurope',
'--server', 'https://127.0.0.1:16444', '--request-timeout=20s', *arguments]
args = ['kubectl', '--kubeconfig', KUBECONFIG, '--request-timeout=20s', *arguments]
result = subprocess.run(args, input=None if payload is None else json.dumps(payload),
text=True, capture_output=True, timeout=30)
require(result.returncode == 0, 'kubernetes_operation_failed')

View file

@ -23,7 +23,7 @@ receipts under `~/nk54/`; they are value-free and go to `docs/evidence/` afterwa
```bash
cd ~/net-kingdom && git pull --ff-only
bridge up k3s-api-railiance01
K="kubectl --kubeconfig $HOME/.kube/config-hosteurope --server https://127.0.0.1:16444"
K="kubectl --kubeconfig $HOME/.kube/config-railiance01" # one route for runbooks and tools (NK-WP-0058-T07)
# The Secret scripts call plain kubectl; the default config is the limited ops-mason-builder.
export KUBECONFIG=$HOME/.kube/config-railiance01
kubectl auth can-i create secrets -n sso # must print yes
@ -94,7 +94,7 @@ python3 sso-mfa/k8s/lldap/bind-split.py verify --endpoint https://lldap.coulomb.
```bash
bash sso-mfa/k8s/identity-provisioner/create-secrets.sh "$CUSTODY"
$K apply -f sso-mfa/k8s/identity-provisioner/deployment.yaml
$K apply --server-side -f sso-mfa/k8s/identity-provisioner/deployment.yaml
$K -n sso rollout status deploy/identity-provisioner --timeout=180s
```
Its readiness includes a directory login, so `rollout status` succeeding means `svc-provisioner`
@ -104,13 +104,13 @@ works. Rollback: `$K -n sso rollout undo deploy/identity-provisioner`.
```bash
bash sso-mfa/k8s/authelia/create-secrets.sh "$CUSTODY" --rotate-key LLDAP_SVC_AUTHELIA_PASS
$K apply -f sso-mfa/k8s/authelia/configmap.yaml
$K apply --server-side -f sso-mfa/k8s/authelia/configmap.yaml
$K -n sso rollout restart deploy/authelia && $K -n sso rollout status deploy/authelia --timeout=180s
```
Check: sign in at `https://auth.coulomb.social` as yourself (password and OTP).
Rollback (the `admin` password still works until step 10):
```bash
git show 145ca8e~1:sso-mfa/k8s/authelia/configmap.yaml | $K apply -f -
git show 145ca8e~1:sso-mfa/k8s/authelia/configmap.yaml | $K apply --server-side -f -
read -rsp 'admin password: ' P; echo
printf '[{"op":"replace","path":"/data/ldap_password","value":"%s"}]' "$(printf %s "$P" | base64 -w0)" \
| $K -n sso patch secret authelia-secrets --type=json --patch-file=/dev/stdin; unset P
@ -148,7 +148,7 @@ automated rollback; the old bind is the `admin` account, which still works, so t
```bash
bash sso-mfa/k8s/lldap/create-secrets.sh "$CUSTODY" --rotate-key LLDAP_LDAP_USER_PASS
$K apply -f sso-mfa/k8s/lldap/deployment.yaml
$K apply --server-side -f sso-mfa/k8s/lldap/deployment.yaml
$K -n sso rollout status deploy/lldap --timeout=180s
```
The configured super-admin is now `ld-admin`. Check: sign in to the LLDAP UI as `ld-admin` once, then

View file

@ -24,10 +24,10 @@ chmod +x create-secrets.sh
./create-secrets.sh
# 3. Apply manifests (order matters)
kubectl apply -f pvc.yaml
kubectl apply -f middleware.yaml
kubectl apply -f deployment.yaml
kubectl apply -f ingress.yaml
kubectl apply --server-side -f pvc.yaml
kubectl apply --server-side -f middleware.yaml
kubectl apply --server-side -f deployment.yaml
kubectl apply --server-side -f ingress.yaml
# 4. Wait for pod to be ready
kubectl rollout status deployment/lldap -n sso --timeout=120s

View file

@ -8,14 +8,16 @@ from __future__ import annotations
import argparse
import base64
import getpass
import os
import json
import re
import subprocess
import sys
CLUSTER_UID = 'a553c742-0115-43d4-99a4-a5ca56fe0786'
KUBE = ['kubectl', '--kubeconfig', '/home/worsch/.kube/config-hosteurope',
'--server', 'https://127.0.0.1:16444', '--request-timeout=20s']
# The runbooks' cluster route (NK-WP-0058-T07): NK_KUBECONFIG, else the admin kubeconfig they name; no override.
KUBE = ['kubectl', '--kubeconfig', os.environ.get('NK_KUBECONFIG', os.path.expanduser('~/.kube/config-railiance01')),
'--request-timeout=20s']
NAMESPACE = 'sso'
# The provisioner's own directory account (NK-WP-0054-T05); needs the image with LLDAP_ADMIN_USERNAME.
SECRET = 'identity-provisioner-ldap'

View file

@ -54,7 +54,7 @@ kubectl create secret generic net-kingdom-pg-privacyidea-app \
### 3. Deploy the cluster
```bash
kubectl apply -f cluster.yaml
kubectl apply --server-side -f cluster.yaml
```
Wait for cluster to become ready (this provisions PVCs and runs initdb — allow 2–3 minutes):
@ -100,7 +100,7 @@ kubectl create secret generic net-kingdom-pg-backup-s3 \
Apply the updated cluster.yaml, then:
```bash
kubectl apply -f scheduled-backup.yaml
kubectl apply --server-side -f scheduled-backup.yaml
```
### 6. Run the restore drill

View file

@ -34,11 +34,11 @@ PI_SQLALCHEMY_DATABASE_URI).
```bash
# From sso-mfa/k8s/privacyidea/
kubectl apply -f pvc.yaml
kubectl apply -f configmap.yaml
kubectl apply -f middleware.yaml
kubectl apply -f deployment.yaml
kubectl apply -f ingress.yaml
kubectl apply --server-side -f pvc.yaml
kubectl apply --server-side -f configmap.yaml
kubectl apply --server-side -f middleware.yaml
kubectl apply --server-side -f deployment.yaml
kubectl apply --server-side -f ingress.yaml
```
**Wait for the pod to reach Running state** (DB migrations run on first start —
@ -185,7 +185,7 @@ Update `middleware.yaml` `privacyidea-admin-allowlist.spec.ipWhiteList.sourceRan
to your actual VPN/office CIDRs and reapply:
```bash
kubectl apply -f middleware.yaml
kubectl apply --server-side -f middleware.yaml
```
### Self-service portal

View file

@ -15,6 +15,7 @@ PYTHONDONTWRITEBYTECODE=1 python3 -m pytest -q -rs -p no:cacheprovider \
tools/execution-attribution-receipt/tests \
sso-mfa/k8s/bounded-executor \
tools/contract-tests/tests \
tools/conventions/tests \
"$@" | tee "$out"
if grep -q '^SKIPPED' "$out"; then
echo "contract-test: skipped tests above; every contract must run (are the owner checkouts present?)" >&2

116
tools/conventions/lint.py Normal file
View file

@ -0,0 +1,116 @@
#!/usr/bin/env python3
"""Runbook and tooling conventions from the 2026-11-04/05 pilot, as checks (NK-WP-0058-T07). Read-only.
python3 tools/conventions/lint.py [ROOT ...] # default: this repository; the rapp CI passes its own root
server-side-apply every `kubectl apply` in a runbook or script is server-side. A client-side apply leaves a
last-applied annotation; ESO copied it and the Secret admission guard refused (2026-11-04).
one-wait-condition no `kubectl wait` with more than one --for: they are ANDed, and the completed run hung it
(2026-11-05). Poll the status instead.
one-cluster-route tools and runbooks reach the cluster through one route: NK_KUBECONFIG, default
~/.kube/config-railiance01. No other kubeconfig file, absolute home path or --server override.
keycape-client.py used a second, hidden route (2026-11-04).
tool-numbered-receipts runbooks never number receipts by hand (`--receipt .../NN-...`); the tool numbers them.
managing-controller no runbook applies a path that an Argo CD Application owns (railiance-platform argocd/):
a manual write there is reverted or drifts (informed-decision, 2026-11-05).
Records (evidence, history, proposals, workplans, wiki) are not instructions and are not checked. A known exception is
listed in EXCEPTIONS with its reason. Exit 0 when clean; otherwise one line per finding: file:line rule.
"""
import re
import sys
from pathlib import Path
import yaml
HERE = Path(__file__).resolve()
RECORDS = re.compile(r'(^|/)(docs/evidence|docs/proposals|history|workplans|wiki|\.git)/|WORK-RECORDS|HISTORICAL')
RUNBOOK = {'.md', '.sh', '.mk'}
TOOL = {'.py', '.sh'}
KUBECTL = r'(kubectl|\$K\b|\$\{K\}|\$\(KUBECTL\)|\$KUBECTL|\$\{KUBECTL\})'
APPLY = re.compile(KUBECTL + r'[^|;&\n]*\sapply(\s|$)')
WAIT = re.compile(KUBECTL + r'[^|;&\n]*\swait\s')
ROUTE = re.compile(r'config-hosteurope|/etc/rancher/k3s/k3s\.yaml|/home/[^/\s\'"]+/\.kube/|'
r'--server[ =,\'"]+[\'"]?https?://127\.0\.0\.1:(16443|16444|6443)')
RECEIPT = re.compile(r'--receipt[ =]+["\']?[^ "\']*/\d{2}-')
SOURCE = re.compile(r'(?:~|\$HOME|/home/[^/\s]+)/([a-z0-9-]+)/([^\s"\']+)')
# (path relative to the root, rule): reason. Keep this short; each entry is a decision.
EXCEPTIONS = {
('sso-mfa/k8s/tenant-engine/README.md', 'server-side-apply'):
"quotes tenant-engine's own deploy command; the owner's runbook, not ours",
('sso-mfa/WORKPLAN.md', 'server-side-apply'): 'legacy checklist from the first SSO setup, not a runbook',
('sso-mfa/k8s/keycape/openbao-negative-session.py', 'one-cluster-route'):
"runs on railiance01 itself: /etc/rancher/k3s/k3s.yaml is the node's own cluster config",
('tools/bootstrap-recovery/install-identity-backup.py', 'one-cluster-route'):
"installs the node's backup job on railiance01: the node's own cluster config",
('sso-mfa/k8s/keycape/openbao-negative-node-route.md', 'one-cluster-route'):
'the command runs on railiance01 itself (the node route of openbao-negative-session.py)',
('docs/bounded-execution-contract-table.md', 'one-cluster-route'):
'the interface inventory describes the old routes on purpose',
('sso-mfa/k8s/bounded-executor/T10-LANES-RUNBOOK.md', 'tool-numbered-receipts'):
'one-time lane setup, completed 2026-10-31 (NK-WP-0056-T10)',
('sso-mfa/k8s/lldap/BIND-SPLIT-RUNBOOK.md', 'tool-numbered-receipts'):
'one-time bind split, completed (NK-WP-0054)',
('sso-mfa/k8s/synthetic-journey/T09-RUNBOOK.md', 'tool-numbered-receipts'):
'one-time journey lane setup, completed (NK-WP-0056-T09)',
}
def argo_paths(platform):
"""(repository, path) pairs that an Argo CD Application in railiance-platform deploys."""
owned = set()
for app in sorted((platform / 'argocd').rglob('*.yaml')) if platform.is_dir() else []:
for doc in yaml.safe_load_all(app.read_text()):
if not isinstance(doc, dict) or doc.get('kind') != 'Application':
continue
for source in [doc['spec'].get('source')] + list(doc['spec'].get('sources') or []):
if source and source.get('path'):
repo = source['repoURL'].rstrip('/').rsplit('/', 1)[-1].removesuffix('.git')
owned.add((repo, source['path'].strip('/')))
return owned
def check_file(rel, text, owned=frozenset()):
findings, suffix = [], Path(rel).suffix or Path(rel).name
runbook = suffix in RUNBOOK or Path(rel).name == 'Makefile'
tool = suffix in TOOL and '/tests/' not in rel and not Path(rel).name.startswith('test_')
for n, full in enumerate(text.splitlines(), 1):
line = full.split(' #')[0] # a trailing shell comment is prose, not part of the command
if runbook and APPLY.search(line) and '--server-side' not in line:
findings.append((rel, n, 'server-side-apply'))
if WAIT.search(line) and line.count('--for') > 1:
findings.append((rel, n, 'one-wait-condition'))
if (tool or runbook) and ROUTE.search(line):
findings.append((rel, n, 'one-cluster-route'))
if suffix == '.md' and RECEIPT.search(line):
findings.append((rel, n, 'tool-numbered-receipts'))
if runbook and APPLY.search(line):
for repo, path in SOURCE.findall(line):
if any(repo == r and (path == p or path.startswith(p + '/')) for r, p in owned):
findings.append((rel, n, 'managing-controller'))
return [f for f in findings if (f[0], f[2]) not in EXCEPTIONS]
def lint(root, owned):
findings = []
for path in sorted(root.rglob('*')):
rel = path.relative_to(root).as_posix()
if not path.is_file() or RECORDS.search(rel) or path == HERE or HERE.parent / 'tests' in path.parents:
continue
if path.suffix in RUNBOOK | TOOL or path.name == 'Makefile':
findings += check_file(rel, path.read_text(errors='replace'), owned)
return findings
def main(argv=None):
roots = [Path(a).resolve() for a in (argv if argv is not None else sys.argv[1:])] or [HERE.parents[2]]
owned = argo_paths(roots[0].parent / 'railiance-platform')
findings = [(root.name, *f) for root in roots for f in lint(root, owned)]
for repo, rel, n, rule in findings:
print(f'{repo}/{rel}:{n} {rule}')
return 1 if findings else 0
if __name__ == '__main__':
sys.exit(main())

View file

@ -0,0 +1,71 @@
"""tools/conventions/lint.py (NK-WP-0058-T07): each rule on synthetic lines, and this repository is clean."""
import importlib.util
from pathlib import Path
import pytest
import yaml
REPO = Path(__file__).resolve().parents[3]
spec = importlib.util.spec_from_file_location('conventions_lint', REPO / 'tools/conventions/lint.py')
lint = importlib.util.module_from_spec(spec)
spec.loader.exec_module(lint)
def rules(rel, line, owned=frozenset()):
return [rule for _, _, rule in lint.check_file(rel, line + '\n', owned)]
@pytest.mark.parametrize('line, found', [
('kubectl apply -f deployment.yaml', ['server-side-apply']),
('$K apply --dry-run=server -f x.yaml && $K apply -f x.yaml', ['server-side-apply']),
('$(KUBECTL) apply -f manifests/rbac.yaml', ['server-side-apply']),
('kubectl apply --server-side -f deployment.yaml', []),
('kubectl get pods # apply nothing', []),
])
def test_server_side_apply(line, found):
assert rules('runbooks/x.md', line) == found
def test_one_wait_condition():
assert rules('x.sh', 'kubectl wait --for=condition=complete --for=condition=failed job/x') == [
'one-wait-condition']
assert rules('x.sh', 'kubectl wait --for=condition=available deploy/x') == []
@pytest.mark.parametrize('line, found', [
("args = ['kubectl', '--kubeconfig', '/home/worsch/.kube/config-hosteurope']", True),
("args = ['kubectl', '--server', 'https://127.0.0.1:16444']", True),
("KUBE = '/etc/rancher/k3s/k3s.yaml'", True),
("KUBECONFIG = os.environ.get('NK_KUBECONFIG', os.path.expanduser('~/.kube/config-railiance01'))", False),
])
def test_one_cluster_route(line, found):
assert ('one-cluster-route' in rules('tools/x.py', line)) is found
assert rules('tools/tests/test_x.py', line) == [] # tests may build synthetic routes
def test_tool_numbered_receipts():
assert rules('runbooks/x.md', 'python3 lane.py setup --receipt "$E/01-setup.json"') == ['tool-numbered-receipts']
assert rules('runbooks/x.md', 'python3 bx.py run') == []
def test_managing_controller(tmp_path):
app = tmp_path / 'railiance-platform/argocd/apps/x.yaml'
app.parent.mkdir(parents=True)
app.write_text(yaml.safe_dump({'kind': 'Application', 'spec': {'source': {
'repoURL': 'https://forgejo.coulomb.social/coulomb/informed-decision.git', 'path': 'deploy/lifecycle'}}}))
owned = lint.argo_paths(tmp_path / 'railiance-platform')
assert owned == {('informed-decision', 'deploy/lifecycle')}
line = 'kubectl apply --server-side -f ~/informed-decision/deploy/lifecycle/review-groups.yaml'
assert rules('runbooks/x.md', line, owned) == ['managing-controller']
assert rules('runbooks/x.md', 'kubectl apply --server-side -f ~/approval-engine/deploy/np.yaml', owned) == []
def test_records_and_exceptions_are_not_findings():
assert lint.RECORDS.search('docs/evidence/2026-11-05-x/README.md')
assert rules('sso-mfa/WORKPLAN.md', 'kubectl apply -f a.yaml') == [] # a listed exception with its reason
assert all(reason for reason in lint.EXCEPTIONS.values())
def test_this_repository_is_clean():
findings = lint.lint(REPO, lint.argo_paths(REPO.parent / 'railiance-platform'))
assert findings == [], findings