Pin caller-auth digest in warn on independently rollable overlay pins

The sanctioned Helm chart could not promote ADR 0004 at all, and the
emergency manifests selected enforce. That made a FLEX-WP-0011 apply
either a no-op or a global 401. First production pin is now warn, per
consumer, on CI digest sha256:138aa347… . Enforce stays a later
per-consumer flip so USER-WP-0023-T03 can close without waiting on
tenant-engine.

charts/flex-auth/Chart.yaml

@@ -2,5 +2,5 @@ apiVersion: v2
 name: flex-auth
 description: Independently rollable flex-auth policy-decision Deployment (one consumer per release).
 type: application
-version: 0.1.0
+version: 0.2.0
 appVersion: "0.1.0"

charts/flex-auth/templates/_helpers.tpl

@@ -18,3 +18,18 @@ app.kubernetes.io/name: {{ include "flex-auth.name" . }}
 app.kubernetes.io/instance: {{ .Release.Name }}
 app.kubernetes.io/part-of: flex-auth
 {{- end -}}
+
+{{- define "flex-auth.callerAuth.mode" -}}
+{{- $mode := "disabled" -}}
+{{- if and (hasKey .Values "callerAuth") .Values.callerAuth (hasKey .Values.callerAuth "mode") .Values.callerAuth.mode -}}
+{{- $mode = .Values.callerAuth.mode -}}
+{{- end -}}
+{{- if not (has $mode (list "disabled" "warn" "enforce")) -}}
+{{- fail (printf "callerAuth.mode must be disabled, warn, or enforce; got %q" $mode) -}}
+{{- end -}}
+{{- $mode -}}
+{{- end -}}
+
+{{- define "flex-auth.callerAuth.enabled" -}}
+{{- if has (include "flex-auth.callerAuth.mode" .) (list "warn" "enforce") -}}true{{- else -}}false{{- end -}}
+{{- end -}}

charts/flex-auth/templates/deployment.yaml

@@ -15,6 +15,9 @@ spec:
         {{- include "flex-auth.labels" . | nindent 8 }}
     spec:
       automountServiceAccountToken: false
+      {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+      serviceAccountName: {{ include "flex-auth.name" . }}
+      {{- end }}
       securityContext:
         runAsNonRoot: true
         seccompProfile:
@@ -25,6 +28,15 @@ spec:
           imagePullPolicy: {{ .Values.image.pullPolicy }}
           args:
             {{- toYaml .Values.args | nindent 12 }}
+            {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+            {{- $binding := required "callerAuth.binding is required when callerAuth.mode is warn or enforce" .Values.callerAuth.binding }}
+            - --caller-auth-mode
+            - {{ include "flex-auth.callerAuth.mode" . | quote }}
+            - --caller-kubernetes-url
+            - {{ required "callerAuth.kubernetesURL is required when callerAuth.mode is warn or enforce" .Values.callerAuth.kubernetesURL | quote }}
+            - --caller-binding
+            - {{ $binding | quote }}
+            {{- end }}
           ports:
             - name: http
               containerPort: {{ .Values.service.port }}
@@ -46,3 +58,25 @@ spec:
               drop:
                 - ALL
             readOnlyRootFilesystem: true
+          {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+          volumeMounts:
+            - mountPath: {{ .Values.callerAuth.reviewer.mountPath }}
+              name: flex-auth-reviewer
+              readOnly: true
+          {{- end }}
+      {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+      volumes:
+        - name: flex-auth-reviewer
+          projected:
+            defaultMode: 0440
+            sources:
+              - serviceAccountToken:
+                  audience: {{ .Values.callerAuth.reviewer.audience | quote }}
+                  expirationSeconds: {{ .Values.callerAuth.reviewer.expirationSeconds }}
+                  path: token
+              - configMap:
+                  name: kube-root-ca.crt
+                  items:
+                    - key: ca.crt
+                      path: ca.crt
+      {{- end }}

charts/flex-auth/templates/networkpolicy.yaml

@@ -11,7 +11,16 @@ spec:
   policyTypes:
     - Ingress
     - Egress
+  {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+  egress:
+    - ports:
+        - port: 443
+          protocol: TCP
+        - port: 6443
+          protocol: TCP
+  {{- else }}
   egress: []
+  {{- end }}
   {{- if .Values.consumer.isolated }}
   ingress: []
   {{- else }}

charts/flex-auth/templates/rbac.yaml

@@ -0,0 +1,30 @@
+{{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: {{ include "flex-auth.name" . }}-tokenreviewer
+  labels:
+    {{- include "flex-auth.labels" . | nindent 4 }}
+rules:
+  - apiGroups:
+      - authentication.k8s.io
+    resources:
+      - tokenreviews
+    verbs:
+      - create
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: {{ include "flex-auth.name" . }}-tokenreviewer
+  labels:
+    {{- include "flex-auth.labels" . | nindent 4 }}
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: {{ include "flex-auth.name" . }}-tokenreviewer
+subjects:
+  - kind: ServiceAccount
+    name: {{ include "flex-auth.name" . }}
+    namespace: {{ .Release.Namespace }}
+{{- end }}

charts/flex-auth/templates/serviceaccount.yaml

@@ -0,0 +1,9 @@
+{{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }}
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: {{ include "flex-auth.name" . }}
+  labels:
+    {{- include "flex-auth.labels" . | nindent 4 }}
+automountServiceAccountToken: false
+{{- end }}

charts/flex-auth/values.yaml

@@ -29,6 +29,18 @@ consumer:
   namespace: ""
   podName: ""
 
+# disabled keeps the image bootable without TokenReview (isolated canary).
+# Production pins set warn, then enforce, independently per consumer.
+callerAuth:
+  mode: disabled
+  audience: flex-auth
+  kubernetesURL: https://kubernetes.default.svc
+  binding: ""
+  reviewer:
+    mountPath: /var/run/secrets/flex-auth-reviewer
+    audience: https://kubernetes.default.svc
+    expirationSeconds: 3600
+
 resources:
   requests:
     cpu: 25m

deploy/README.md

@@ -7,6 +7,12 @@ emergency kubectl path recovered on 2026-08-11 from live
 `last-applied-configuration`, kept so a rollback does not depend on a
 cluster annotation or on Helm history.
 
+Caller authentication (FLEX-WP-0015-T02) is pinned here in **`warn`**,
+matching `values/user-engine.yaml` and `values/tenant-engine.yaml`. Do not
+edit these files back to `enforce` and apply them as a shortcut: that is
+the hazard. Flip `callerAuth.mode` per consumer in the overlay after that
+consumer's warn logs are clean, then re-render or edit only that pin.
+
 Manifests for the two cluster-local flex-auth policy-decision services.
 
 | File | Deployment | Consumer | Service DNS |
@@ -86,10 +92,12 @@ last-known-good digest below.
 
 | Deployment | Last-known-good digest | Policy state |
 | --- | --- | --- |
-| `flex-auth-tenant-engine` | `sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc` | **live** — nine-action policy (FLEX-WP-0014), CI-built from `f304688` |
+| `flex-auth-tenant-engine` | `sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b` | **pin, not live** — caller-auth **warn** (FLEX-WP-0015-T02), CI `main-3de72fe` |
+| `flex-auth-tenant-engine` *(live until warn promote)* | `sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc` | nine-action policy (FLEX-WP-0014), CI-built from `f304688` |
 | `flex-auth-tenant-engine` *(previous)* | `sha256:9320df394a642eff24da8af4a0ee8886a7bb78b0f14d8ee1deeb30ea8eeeaba7` | seven-action policy, FLEX-WP-0013 restore; guardrail actions deny `unknown_action` |
 | `flex-auth-tenant-engine` *(rollback)* | `sha256:c25fc34a6cd7e64d955f8723ec70e176a583d5ae71d76280c4e2d89fba0fe0aa` | four-action policy; lifecycle actions deny `unknown_action` |
-| `flex-auth-user-engine` | `sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4` | **current** — FLEX-WP-0009-T04, nine fixtures (incl. registration-applicant) verified live 2026-08-16 |
+| `flex-auth-user-engine` | `sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b` | **pin, not live** — caller-auth **warn** (FLEX-WP-0015-T02), CI `main-3de72fe` |
+| `flex-auth-user-engine` *(live until warn promote)* | `sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4` | FLEX-WP-0009-T04, nine fixtures (incl. registration-applicant) verified live 2026-08-16 |
 | `flex-auth-user-engine` *(previous)* | `sha256:a31961c45215aa6baf3bc748c6741ab703c2c8325e61aa7983a355026195e51b` | FLEX-WP-0009-T03, six fixtures verified live 2026-08-10 |
 
 Rolling back the tenant-engine service to `c25fc34a…` restores fail-closed

deploy/flex-auth-tenant-engine.yaml

@@ -25,12 +25,12 @@ spec:
         - --policy
         - /opt/flex-auth/examples/tenant-engine/policy_package.md
         - --caller-auth-mode
-        - enforce
+        - warn
         - --caller-kubernetes-url
         - https://10.43.0.1
         - --caller-binding
         - tenant-engine=system:serviceaccount:tenant-engine:tenant-engine
-        image: forgejo.coulomb.social/coulomb/flex-auth@sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc
+        image: forgejo.coulomb.social/coulomb/flex-auth@sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
         livenessProbe:
           httpGet:
             path: /healthz

deploy/flex-auth-user-engine.yaml

@@ -25,12 +25,12 @@ spec:
         - --policy
         - /opt/flex-auth/examples/user-engine/policy_package.md
         - --caller-auth-mode
-        - enforce
+        - warn
         - --caller-kubernetes-url
         - https://10.43.0.1
         - --caller-binding
         - user-engine=system:serviceaccount:user-engine:user-engine
-        image: forgejo.coulomb.social/coulomb/flex-auth@sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4
+        image: forgejo.coulomb.social/coulomb/flex-auth@sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
         livenessProbe:
           httpGet:
             path: /healthz

docs/adr/0004-inbound-caller-authentication.md

@@ -38,9 +38,11 @@ cached across rotation.
 
 Three modes support promotion: `disabled`, `warn`, and `enforce`. Warn mode
 records the same authentication failures without logging credentials. It is a
-bounded migration aid, not a conformant steady state. The reviewed desired
+bounded migration aid, not a conformant steady state. The first production pin
-manifests select `enforce`; promotion still follows FLEX-WP-0011 and requires a
+is `warn` on each independently rollable Deployment; `enforce` is the end
-new immutable image digest plus caller rollout evidence.
+state and is flipped per consumer after that consumer's warn logs are clean.
+Promotion follows FLEX-WP-0011. The two production pins must not be flipped
+together: user-engine can enforce while tenant-engine stays in warn.
 
 ## Rejected alternatives
 

docs/tenancy-posture-review.md

@@ -271,7 +271,7 @@ volunteer.
 | Task | |
 |---|---|
 | T01 | Publish the posture vector and this review; reply to `rapp-postgres` |
-| T02 | **Close the A0**: TokenReview source/desired state done; immutable promotion pending |
+| T02 | **Close the A0**: TokenReview source + warn-first overlay pin done; operator promote-then-per-consumer-enforce pending |
 | T03 | Deleted the unused tenant-engine live-roles adapter |
 | T04 | AuthZEN endpoint — `wait`, with a written trigger |
 | T05 | `make verify-posture` guards declaration/source/deployment drift |

railiance/README.md

@@ -45,6 +45,32 @@ $RAILIANCE rollback . --apply --approval-id <state-hub-id> --revision <helm-rev>
 The canary Service is `flex-auth-canary.flex-auth.svc.cluster.local:8080`.
 Production consumers keep calling their own Service names.
 
+## Caller-auth promotion (FLEX-WP-0015-T02)
+
+The two production pins are independently rollable. Do not flip both to
+`enforce` in one apply. The first pin of digest
+`sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b`
+is **warn** on each consumer.
+
+```bash
+# Warn pin (safe if a caller still sends no token)
+helm upgrade --install flex-auth-user-engine charts/flex-auth \
+  --namespace flex-auth -f values/user-engine.yaml --wait --timeout 2m
+helm upgrade --install flex-auth-tenant-engine charts/flex-auth \
+  --namespace flex-auth -f values/tenant-engine.yaml --wait --timeout 2m
+
+# Confirm warn logs: "caller authentication warning:" must be absent for a
+# migrated caller. user-engine is migrated; tenant-engine live status is
+# unconfirmed. Isolated canary keeps callerAuth.mode=disabled.
+
+# Then set values/user-engine.yaml callerAuth.mode: enforce and upgrade only
+# that release. USER-WP-0023-T03 runs only after that flip. tenant-engine
+# stays warn until its logs are clean.
+```
+
+Emergency kubectl path is also warn. Applying it in enforce is the hazard
+the warn pin exists to prevent.
+
 ## Rolling a production pin
 
 Policy is baked into the image. Do not build images on a workstation.
@@ -76,8 +102,10 @@ or re-apply the last-known-good digest in `deploy/README.md`.
 
 | Deployment | Last-known-good digest | Policy state |
 | --- | --- | --- |
-| `flex-auth-tenant-engine` | `sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc` | **live** — nine-action policy (FLEX-WP-0014), CI-built from `f304688` |
+| `flex-auth-tenant-engine` | `sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b` | **pin, not live** — caller-auth warn (FLEX-WP-0015-T02), CI `main-3de72fe` |
+| `flex-auth-tenant-engine` *(live until warn promote)* | `sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc` | nine-action policy (FLEX-WP-0014), CI-built from `f304688` |
 | `flex-auth-tenant-engine` *(previous)* | `sha256:9320df394a642eff24da8af4a0ee8886a7bb78b0f14d8ee1deeb30ea8eeeaba7` | seven-action policy, FLEX-WP-0013 restore |
 | `flex-auth-tenant-engine` *(rollback)* | `sha256:c25fc34a6cd7e64d955f8723ec70e176a583d5ae71d76280c4e2d89fba0fe0aa` | four-action policy; lifecycle actions deny `unknown_action` |
-| `flex-auth-user-engine` | `sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4` | FLEX-WP-0009-T04, nine fixtures, live 2026-08-16 |
+| `flex-auth-user-engine` | `sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b` | **pin, not live** — caller-auth warn (FLEX-WP-0015-T02), CI `main-3de72fe` |
+| `flex-auth-user-engine` *(live until warn promote)* | `sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4` | FLEX-WP-0009-T04, nine fixtures, live 2026-08-16 |
 | `flex-auth-user-engine` *(previous)* | `sha256:a31961c45215aa6baf3bc748c6741ab703c2c8325e61aa7983a355026195e51b` | FLEX-WP-0009-T03, six fixtures |

tenancy.yaml

@@ -40,10 +40,11 @@ tenancy:
       own inputs. Not a defect and not a target for movement.
     A: >-
       The running immutable digest still authenticates no caller, so current
-      remains A0. Source and reviewed desired manifests implement A2 with an
+      remains A0. Source, overlay, and reviewed desired manifests implement
-      audience-scoped Kubernetes TokenReview choke point and exact
+      A2 with an audience-scoped Kubernetes TokenReview choke point and exact
-      protected-system-to-ServiceAccount bindings. Promotion and a live
+      protected-system-to-ServiceAccount bindings. The first production pin
-      unbound-request probe remain under FLEX-WP-0011/FLEX-WP-0015-T02.
+      is warn, independently per consumer, under FLEX-WP-0011/FLEX-WP-0015-T02.
+      Current moves to A2 only after enforce and a live unbound-request probe.
     E: >-
       No tenant data at rest. Tenant scoping in decisions runs through one
       choke point (internal/decision/engine.go normalizeRequest and the
@@ -84,6 +85,7 @@ evidence:
     - "internal/callerauth/auth.go"
     - "internal/callerauth/auth_test.go"
     - "cmd/flex-auth/main_test.go"
+    - "charts/flex-auth/templates/rbac.yaml"
     - "deploy/caller-auth-rbac.yaml"
     - "docs/adr/0004-inbound-caller-authentication.md"
-  deployment: "deploy/flex-auth-user-engine.yaml, deploy/flex-auth-tenant-engine.yaml"
+  deployment: "values/user-engine.yaml, values/tenant-engine.yaml, deploy/flex-auth-user-engine.yaml, deploy/flex-auth-tenant-engine.yaml"

tests/stage1.sh

@@ -12,4 +12,21 @@ for values in values/stage1.yaml values/stage2-canary.yaml values/stage3-product
   helm template flex-auth charts/flex-auth -f "${values}" --namespace flex-auth >/dev/null
 done
 
+user_render="$(helm template flex-auth-user-engine charts/flex-auth -f values/user-engine.yaml --namespace flex-auth)"
+echo "$user_render" | grep -q -- '--caller-auth-mode' || { echo "user-engine render omits caller-auth-mode" >&2; exit 1; }
+echo "$user_render" | grep -A1 -- '--caller-auth-mode' | grep -q warn || { echo "user-engine first pin must be warn" >&2; exit 1; }
+echo "$user_render" | grep -q tokenreviews || { echo "user-engine render omits TokenReview RBAC" >&2; exit 1; }
+echo "$user_render" | grep -q flex-auth-reviewer || { echo "user-engine render omits reviewer projection" >&2; exit 1; }
+
+tenant_render="$(helm template flex-auth-tenant-engine charts/flex-auth -f values/tenant-engine.yaml --namespace flex-auth)"
+echo "$tenant_render" | grep -A1 -- '--caller-auth-mode' | grep -q warn || { echo "tenant-engine first pin must be warn" >&2; exit 1; }
+echo "$tenant_render" | grep -q 'tenant-engine=system:serviceaccount:tenant-engine:tenant-engine' \
+  || { echo "tenant-engine render omits its exact binding" >&2; exit 1; }
+
+canary_render="$(helm template flex-auth-canary charts/flex-auth -f values/stage2-canary.yaml --namespace flex-auth)"
+if echo "$canary_render" | grep -q tokenreviews; then
+  echo "isolated canary must not create TokenReview RBAC" >&2
+  exit 1
+fi
+
 echo "stage1 overlay render ok"

tools/verify-posture.sh

@@ -26,14 +26,36 @@ fail() {
   exit 1
 }
 
+caller_mode_from_manifest() {
+  awk '
+    $0 ~ /--caller-auth-mode/ { getline; gsub(/^[[:space:]-]+/, ""); print; exit }
+  ' "$1"
+}
+
+caller_mode_from_values() {
+  awk '
+    $1 == "callerAuth:" { in_block=1; next }
+    in_block && /^[^[:space:]#]/ { exit }
+    in_block && $1 == "mode:" { print $2; exit }
+  ' "$1"
+}
+
 if grep -q 'authenticator.Authorize' "$main"; then
   [[ "${declared_a:-0}" -ge 2 ]] || fail "caller authentication exists but current/implemented A is below 2"
   for manifest in deploy/flex-auth-user-engine.yaml deploy/flex-auth-tenant-engine.yaml; do
     grep -q -- '--caller-auth-mode' "$root/$manifest" || fail "$manifest omits caller auth mode"
-    grep -q 'enforce' "$root/$manifest" || fail "$manifest does not select enforce mode"
+    mode="$(caller_mode_from_manifest "$root/$manifest")"
+    [[ "$mode" == "warn" || "$mode" == "enforce" ]] || fail "$manifest caller-auth-mode is ${mode:-empty}, not warn or enforce"
     grep -q -- '--caller-binding' "$root/$manifest" || fail "$manifest omits the exact system binding"
+    grep -q 'flex-auth-reviewer' "$root/$manifest" || fail "$manifest omits the reviewer token projection"
   done
   [[ -f "$root/deploy/caller-auth-rbac.yaml" ]] || fail "TokenReview RBAC manifest is absent"
+  grep -q 'tokenreviews' "$root/charts/flex-auth/templates/rbac.yaml" || fail "overlay TokenReview RBAC template is absent"
+  for values in values/user-engine.yaml values/tenant-engine.yaml; do
+    mode="$(caller_mode_from_values "$root/$values")"
+    [[ "$mode" == "warn" || "$mode" == "enforce" ]] || fail "$values callerAuth.mode is ${mode:-empty}, not warn or enforce"
+    grep -q 'binding:' "$root/$values" || fail "$values omits callerAuth.binding"
+  done
 else
   [[ "${declared_a:-0}" -lt 2 ]] || fail "A2 is declared without a caller-authentication choke point"
 fi

values/stage1.yaml

@@ -2,7 +2,7 @@
 name: flex-auth-local
 image:
   repository: forgejo.coulomb.social/coulomb/flex-auth
-  digest: sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4
+  digest: sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
 args:
   - serve
   - --addr

values/stage2-canary.yaml

@@ -3,7 +3,7 @@
 name: flex-auth-canary
 image:
   repository: forgejo.coulomb.social/coulomb/flex-auth
-  digest: sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4
+  digest: sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
 args:
   - serve
   - --addr

values/stage3-production.yaml

@@ -5,7 +5,7 @@
 name: flex-auth-canary
 image:
   repository: forgejo.coulomb.social/coulomb/flex-auth
-  digest: sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4
+  digest: sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
 args:
   - serve
   - --addr

values/tenant-engine.yaml

@@ -1,8 +1,10 @@
 # Production pin for the tenant-engine policy service. Independently rollable.
+# Warn is safe whether or not the live tenant-engine client already sends a
+# token. Do not flip this pin to enforce until its warn logs are clean.
 name: flex-auth-tenant-engine
 image:
   repository: forgejo.coulomb.social/coulomb/flex-auth
-  digest: sha256:1bf060e61122693ce98359c167cc5fe8bdafc84e097e090eaa71af94d0f27cbc
+  digest: sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
 args:
   - serve
   - --addr
@@ -11,6 +13,10 @@ args:
   - /opt/flex-auth/examples/tenant-engine/registry_snapshot.json
   - --policy
   - /opt/flex-auth/examples/tenant-engine/policy_package.md
+callerAuth:
+  mode: warn
+  kubernetesURL: https://10.43.0.1
+  binding: tenant-engine=system:serviceaccount:tenant-engine:tenant-engine
 consumer:
   isolated: false
   namespace: tenant-engine

values/user-engine.yaml

@@ -1,8 +1,10 @@
 # Production pin for the user-engine policy service. Independently rollable.
+# First caller-auth pin is warn (FLEX-WP-0015-T02). Flip mode to enforce only
+# after warn logs are clean of unauthenticated callers for this consumer.
 name: flex-auth-user-engine
 image:
   repository: forgejo.coulomb.social/coulomb/flex-auth
-  digest: sha256:1f5290376dc5fcf456dc7a785e394d8b90949dabecd1d3e856f38557149bb5f4
+  digest: sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b
 args:
   - serve
   - --addr
@@ -11,6 +13,10 @@ args:
   - /opt/flex-auth/examples/user-engine/registry_snapshot.json
   - --policy
   - /opt/flex-auth/examples/user-engine/policy_package.md
+callerAuth:
+  mode: warn
+  kubernetesURL: https://10.43.0.1
+  binding: user-engine=system:serviceaccount:user-engine:user-engine
 consumer:
   isolated: false
   namespace: user-engine

workplans/FLEX-WP-0015-tenancy-posture-conformance.md

@@ -14,7 +14,7 @@ related_workplans:
   - FLEX-WP-0004
   - FLEX-WP-0011
 created: "2026-08-17"
-updated: "2026-08-18"
+updated: "2026-08-19"
 state_hub_workstream_id: "31846b19-c2a3-428e-950b-5985bc9146eb"
 ---
 
@@ -110,38 +110,50 @@ Ship through the FLEX-WP-0011 staged-promotion path, not by direct apply.
 ServiceAccount TokenReview with a separately projected reviewer identity and
 exact `resource.system` → ServiceAccount bindings. Both check endpoints share
 the choke point; missing/mismatched identity fails 401/403 and reviewer outage
-fails 503. Unit and handler tests include the required unbound request. Desired
+fails 503. Unit and handler tests include the required unbound request. Source
-manifests select enforce mode and carry narrow TokenReview RBAC. The running
+committed as `1e1e077`.
-digest is unchanged, so `tenancy.current.A` honestly remains 0 while
-`tenancy.implemented.A` is 2. Source committed as `1e1e077`.
 
-**Remaining, in this order — the sequence is a constraint, not a preference.**
+2026-08-19 overlay outcome: the sanctioned Helm chart did not render caller
-`ops-warden` ruled on rollout (2026-08-17) and `user-engine` asked for the same:
+auth at all, so a FLEX-WP-0011 pin of the new digest would have stayed
+`disabled`. That is now wired. Desired emergency manifests and production
+values select **`warn`**, not `enforce` — applying the previous enforce
+manifests was the hazard. CI image `sha256:138aa3471c46bca6e814691fa1e6520aedda3dffd743e6b09141ab433afdb64b`
+(`main-3de72fe`, built 2026-08-18T13:25Z) contains the TokenReview flags.
+The running cluster digest is still unchanged, so `tenancy.current.A`
+remains 0 while `tenancy.implemented.A` is 2.
 
-1. Build and pin a new immutable digest carrying the caller-auth code.
+**Better path than a global enforce flip.** The two production Deployments
-2. Promote it in **`warn` mode**, not `enforce`. The reviewed desired manifests
+are independently rollable (FLEX-WP-0011). USER-WP-0023-T03 only probes
-   in `deploy/` select `enforce` because that is the end state — applying them
+`flex-auth-user-engine`. Serializing that probe on tenant-engine's caller
-   directly is the hazard. Warn authenticates and logs failures without
+status would 401 tenant-engine writes if we enforced them together, and
-   rejecting, so it is safe for an unmigrated caller.
+would delay user-engine A2 evidence for no safety gain. Sequence:
-3. Confirm the warn logs are clean of unauthenticated callers. `user-engine` is
-   migrated and deployed as of 2026-08-18 (image `sha256:c501aeb2…`, token at
-   `/var/run/secrets/flex-auth-caller/token`). `tenant-engine`'s caller status
-   is **unconfirmed — asked 2026-08-18**; promoting `enforce` while their
-   client sends no token 401s every check and blocks their write paths.
-   `ops-warden` adopts the calling side on its own schedule.
-4. Flip to `enforce`, then capture the live negative probe. `user-engine` has
-   the three assertions written and ready in their
-   `docs/flex-auth-caller-identity.md`; they deliberately have not run it,
-   because against an unenforced digest "no token returns 401" comes back as a
-   normal decision and would record as false evidence.
-5. Only then may `policy.enabled` flip anywhere — `ops-warden` names that the
-   real deadline, and it is the same gate as FLEX-WP-0007.
 
-**Operator gate.** Steps 1–4 need cluster credentials this session does not
+1. Pin the caller-auth digest in overlay + emergency manifests — **done
+   2026-08-19**, first mode `warn` on both consumers.
+2. Operator: promote each pin in warn via `helm upgrade --install
+   flex-auth-<consumer> charts/flex-auth --namespace flex-auth -f
+   values/<consumer>.yaml`. Isolated canary may boot the same digest with
+   caller-auth disabled. Do not `kubectl apply` an enforce manifest.
+3. Confirm warn logs per consumer. `user-engine` is migrated and deployed
+   as of 2026-08-18 (image `sha256:c501aeb2…`, token at
+   `/var/run/secrets/flex-auth-caller/token`). `tenant-engine` source and
+   desired manifests already project that token and send it; live digest
+   is still unconfirmed. Warn is safe either way. `ops-warden` adopts the
+   calling side on its own schedule.
+4. Flip **user-engine only** to `enforce` (`callerAuth.mode: enforce` in
+   `values/user-engine.yaml`, same digest). Then capture the live negative
+   probe from a user-engine pod. The three assertions are already written
+   in user-engine `docs/flex-auth-caller-identity.md`; they must not run
+   against warn, because "no token returns 401" would still be a decision.
+5. Flip tenant-engine to enforce only after its warn logs are clean.
+6. Only then may `policy.enabled` flip anywhere — `ops-warden` names that
+   the real deadline, and it is the same gate as FLEX-WP-0007.
+
+**Operator gate.** Steps 2–5 need cluster credentials this session does not
 have: `kubectl` returns `Unauthorized` and the context is `default`, which
 `tenant-engine` documented on 2026-08-16 as indistinguishable from a
-wrong-cluster KUBECONFIG. Promotion follows FLEX-WP-0011 staged promotion and
+wrong-cluster KUBECONFIG. Promotion follows FLEX-WP-0011 and the CI image,
-the CI image build, not a hand-built image.
+not a hand-built image.
 
 `RISK-F-0001` (risk-nexus) tracks this A0 externally. Their NetworkPolicy
 question was answered 2026-08-18: both Deployments carry an ingress policy
@@ -218,6 +230,7 @@ correct-looking responses, exactly like the two silent pin rollbacks the
 estate has already been bitten by.
 
 Completed 2026-08-18. `make verify-posture` checks the current/implemented
-distinction, both authenticated handlers, enforce-mode desired manifests,
+distinction, both authenticated handlers, warn-or-enforce desired manifests
-TokenReview RBAC, absence of a latent tenant-engine caller, and the stateless
+and overlay pins, TokenReview RBAC, absence of a latent tenant-engine caller,
-deployment assumptions behind `R: n/a`.
+and the stateless deployment assumptions behind `R: n/a`. First production
+pin is warn; enforce is the end state, flipped per consumer.