diff --git a/Makefile b/Makefile index 9cc60b4..26ae34f 100644 --- a/Makefile +++ b/Makefile @@ -64,6 +64,10 @@ ci: vet lint test build overlay-render verify-posture: @bash tools/verify-posture.sh +.PHONY: verify-emission-cadence +verify-emission-cadence: + python3 tools/verify_emission_cadence.py + overlay-render: @tests/stage1.sh diff --git a/WORK-RECORDS.md b/WORK-RECORDS.md index b189bfd..245548a 100644 --- a/WORK-RECORDS.md +++ b/WORK-RECORDS.md @@ -35,7 +35,7 @@ | workplan | FLEX-WP-0024 | finished | — | workplans/FLEX-WP-0024-decision-envelope-authenticity.md | | workplan | FLEX-WP-0025 | finished | — | workplans/FLEX-WP-0025-fact-versus-assertion.md | | workplan | FLEX-WP-0026 | finished | — | workplans/FLEX-WP-0026-openrouter-native-contract.md | -| workplan | FLEX-WP-0027 | blocked | — | workplans/FLEX-WP-0027-t03-human-review.md | +| workplan | FLEX-WP-0027 | finished | — | workplans/FLEX-WP-0027-t03-human-review.md | | workplan | FLEX-WP-0028 | finished | — | workplans/FLEX-WP-0028-compact-sitting-review.md | | workplan | FLEX-WP-0029 | finished | — | workplans/FLEX-WP-0029-stance-register-second-edition.md | | workplan | FLEX-WP-0030 | finished | — | workplans/FLEX-WP-0030-boundary-declaration-cleanup.md | @@ -165,7 +165,7 @@ | task | FLEX-WP-0026-T02 | done | — | workplans/FLEX-WP-0026-openrouter-native-contract.md | | task | FLEX-WP-0027-T01 | done | — | workplans/FLEX-WP-0027-t03-human-review.md | | task | FLEX-WP-0027-T02 | done | — | workplans/FLEX-WP-0027-t03-human-review.md | -| task | FLEX-WP-0027-T03 | wait | — | workplans/FLEX-WP-0027-t03-human-review.md | +| task | FLEX-WP-0027-T03 | done | — | workplans/FLEX-WP-0027-t03-human-review.md | | task | FLEX-WP-0028-T01 | done | — | workplans/FLEX-WP-0028-compact-sitting-review.md | | task | FLEX-WP-0028-T02 | done | — | workplans/FLEX-WP-0028-compact-sitting-review.md | | task | FLEX-WP-0029-T01 | done | — | workplans/FLEX-WP-0029-stance-register-second-edition.md | @@ -184,8 +184,10 @@ | task | FLEX-WP-0031-T02 | wait | — | workplans/FLEX-WP-0031-decision-record-emission.md | | task | FLEX-WP-0031-T03 | done | — | workplans/FLEX-WP-0031-decision-record-emission.md | | task | FLEX-WP-0031-T04 | wait | — | workplans/FLEX-WP-0031-decision-record-emission.md | -| task | FLEX-WP-0031-T05 | wait | — | workplans/FLEX-WP-0031-decision-record-emission.md | +| task | FLEX-WP-0031-T05 | done | — | workplans/FLEX-WP-0031-decision-record-emission.md | | task | FLEX-WP-0031-T06 | wait | — | workplans/FLEX-WP-0031-decision-record-emission.md | +| task | FLEX-WP-0031-T07 | done | — | workplans/FLEX-WP-0031-decision-record-emission.md | +| task | FLEX-WP-0031-T08 | wait | — | workplans/FLEX-WP-0031-decision-record-emission.md | | task | FLEX-WP-0032-T01 | done | — | workplans/FLEX-WP-0032-informed-decision-list-action.md | | task | FLEX-WP-0032-T02 | done | — | workplans/FLEX-WP-0032-informed-decision-list-action.md | | task | FLEX-WP-0032-T03 | done | — | workplans/FLEX-WP-0032-informed-decision-list-action.md | diff --git a/cadence.yaml b/cadence.yaml index f0aa2b1..a9fab9e 100644 --- a/cadence.yaml +++ b/cadence.yaml @@ -36,10 +36,11 @@ form: heartbeat-or-reconciliation rate_monitoring: forbidden # class-level exception stated on `allow` only # STATE. The classification below is published and in force as a declaration. -# The emission pipeline is not built: no sender named flex-auth or -# access-engine is registered with audit-core, no token lane, no scope row, and -# the decision record reaches consumers in the /v1/check response rather than -# into any custody. This file is therefore the declaration §11 requires and not +# The outbox, disabled-by-default heartbeat/drain and optional PVC are built +# (FLEX-WP-0031 T03/T05/T07). No pin enables them in production, no sender is +# registered with audit-core, and token/scope/observer admission is pending. +# The decision record still reaches consumers in the /v1/check response rather +# than into custody. This file is therefore the declaration §11 requires and not # a claim that the guarantee is delivered today. The delivery gap is G2 in # docs/conformance/security-layer-conformance.md, with owner, blocker and # review date, and it fails CLOSED on its distinguishing case: a stream with no @@ -150,10 +151,10 @@ classes: removed. An allow carries a TTL lifetime, so a suppressed allow expires into a denial of service rather than into a standing grant. heartbeat: - action: flex-auth.decision.heartbeat + action: audit-core.heartbeat evidence_class: operational rarity: scheduled - note: The positive assertion above; not a decision. + note: Per-class data.class assertion above; not a decision. Wire declaration is docs/contracts/decision-emission-cadence.json. bound: >- Heartbeat and reconciliation detect loss, outage, drain failure and accident. diff --git a/charts/flex-auth/templates/deployment.yaml b/charts/flex-auth/templates/deployment.yaml index 55b0ff1..8bccd6c 100644 --- a/charts/flex-auth/templates/deployment.yaml +++ b/charts/flex-auth/templates/deployment.yaml @@ -5,6 +5,13 @@ metadata: labels: {{- include "flex-auth.labels" . | nindent 4 }} spec: + {{- if .Values.outbox.enabled }} + {{- if ne (int .Values.replicaCount) 1 }} + {{- fail "outbox requires replicaCount=1: one writer per source and PVC" }} + {{- end }} + strategy: + type: Recreate + {{- end }} replicas: {{ .Values.replicaCount }} selector: matchLabels: @@ -31,6 +38,12 @@ spec: imagePullPolicy: {{ .Values.image.pullPolicy }} args: {{- toYaml .Values.args | nindent 12 }} + {{- if .Values.outbox.enabled }} + - --outbox-dir + - {{ .Values.outbox.mountPath | quote }} + - --emission-source + - {{ required "outbox.source is required when outbox.enabled=true" .Values.outbox.source | quote }} + {{- end }} {{- 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 @@ -65,14 +78,26 @@ spec: drop: - ALL readOnlyRootFilesystem: true - {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }} + {{- if or .Values.outbox.enabled (eq (include "flex-auth.callerAuth.enabled" .) "true") }} volumeMounts: + {{- if .Values.outbox.enabled }} + - mountPath: {{ .Values.outbox.mountPath | quote }} + name: flex-auth-outbox + {{- end }} + {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }} - mountPath: {{ .Values.callerAuth.reviewer.mountPath }} name: flex-auth-reviewer readOnly: true + {{- end }} {{- end }} - {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }} + {{- if or .Values.outbox.enabled (eq (include "flex-auth.callerAuth.enabled" .) "true") }} volumes: + {{- if .Values.outbox.enabled }} + - name: flex-auth-outbox + persistentVolumeClaim: + claimName: {{ default (printf "%s-outbox" (include "flex-auth.name" .)) .Values.outbox.existingClaim | quote }} + {{- end }} + {{- if eq (include "flex-auth.callerAuth.enabled" .) "true" }} - name: flex-auth-reviewer projected: defaultMode: 0440 @@ -88,4 +113,5 @@ spec: items: - key: ca.crt path: ca.crt + {{- end }} {{- end }} diff --git a/charts/flex-auth/templates/outbox-pvc.yaml b/charts/flex-auth/templates/outbox-pvc.yaml new file mode 100644 index 0000000..7219f77 --- /dev/null +++ b/charts/flex-auth/templates/outbox-pvc.yaml @@ -0,0 +1,19 @@ +{{- if and .Values.outbox.enabled (not .Values.outbox.existingClaim) }} +apiVersion: v1 +kind: PersistentVolumeClaim +metadata: + name: {{ include "flex-auth.name" . }}-outbox + annotations: + helm.sh/resource-policy: keep + labels: + {{- include "flex-auth.labels" . | nindent 4 }} +spec: + accessModes: + - ReadWriteOnce + {{- if .Values.outbox.storageClassName }} + storageClassName: {{ .Values.outbox.storageClassName | quote }} + {{- end }} + resources: + requests: + storage: {{ .Values.outbox.size | quote }} +{{- end }} diff --git a/charts/flex-auth/values.yaml b/charts/flex-auth/values.yaml index f0dc11d..9f3d517 100644 --- a/charts/flex-auth/values.yaml +++ b/charts/flex-auth/values.yaml @@ -9,6 +9,15 @@ image: replicaCount: 1 +# Inert until enabled for an approved pin. The outbox is single-writer. +outbox: + enabled: false + source: "" + mountPath: /var/lib/flex-auth/outbox + size: 1Gi + storageClassName: "" + existingClaim: "" + args: - serve - --addr diff --git a/cmd/flex-auth/main.go b/cmd/flex-auth/main.go index 49d078a..3355872 100644 --- a/cmd/flex-auth/main.go +++ b/cmd/flex-auth/main.go @@ -12,6 +12,7 @@ import ( "net/http" "os" "strings" + "time" "gopkg.in/yaml.v3" @@ -329,6 +330,8 @@ func runServe(args []string, stdout, stderr io.Writer) int { logPath := fs.String("log", "", "optional JSONL decision log path") outboxDir := fs.String("outbox-dir", "", "durable emission outbox directory (FLEX-WP-0031); exclusive with --log") emissionSource := fs.String("emission-source", "", "audit-core source name for emitted decision events, e.g. flex-auth.tenant-engine") + auditURL := fs.String("audit-url", "", "admitted audit-core /v1/events endpoint; empty disables drain and heartbeats") + auditTokenFile := fs.String("audit-token-file", "", "mounted sender token file, never a token value") callerAuthMode := fs.String("caller-auth-mode", "disabled", "disabled, warn, or enforce") callerAudience := fs.String("caller-audience", "flex-auth", "required caller token audience") callerKubernetesURL := fs.String("caller-kubernetes-url", "https://kubernetes.default.svc", "Kubernetes API base URL for TokenReview") @@ -348,6 +351,17 @@ func runServe(args []string, stdout, stderr io.Writer) int { fmt.Fprintln(stderr, "serve --outbox-dir requires --emission-source and excludes --log") return 64 } + delivery := emission.DeliveryConfig{URL: *auditURL, TokenFile: *auditTokenFile} + if *auditURL != "" || *auditTokenFile != "" { + if *outboxDir == "" { + fmt.Fprintln(stderr, "audit delivery requires --outbox-dir") + return 64 + } + if err := delivery.Validate(); err != nil { + fmt.Fprintln(stderr, err) + return 64 + } + } engine, err := buildEngine(context.Background(), *registryPath, *policyPath, *logPath) if err != nil { @@ -380,6 +394,15 @@ func runServe(args []string, stdout, stderr io.Writer) int { if outbox != nil { mountEmission(mux, outbox) } + if *auditURL != "" { + ctx, cancel := context.WithCancel(context.Background()) + done := make(chan struct{}) + go func() { + defer close(done) + runEmissionDelivery(ctx, outbox, delivery, stderr) + }() + defer func() { cancel(); <-done }() + } fmt.Fprintf(stderr, "flex-auth serving on http://%s\n", *addr) if err := http.ListenAndServe(*addr, mux); err != nil { @@ -388,6 +411,31 @@ func runServe(args []string, stdout, stderr io.Writer) int { return 0 } +func runEmissionDelivery(ctx context.Context, outbox *emission.Outbox, config emission.DeliveryConfig, stderr io.Writer) { + ticker := time.NewTicker(time.Minute) + defer ticker.Stop() + for { + if ctx.Err() != nil { + return + } + if err := outbox.AppendDueHeartbeats(); err != nil { + fmt.Fprintf(stderr, "emission heartbeat commit: %v\n", err) + } + n, err := outbox.Drain(ctx, config, 100) + if err != nil && ctx.Err() == nil { + fmt.Fprintf(stderr, "emission drain: %v\n", err) + } + if err == nil && n == 100 { + continue // Drain the next bounded batch without imposing a rate cap. + } + select { + case <-ctx.Done(): + return + case <-ticker.C: + } + } +} + func newServeMux(engine *decisioncore.Engine) *http.ServeMux { return newServeMuxWithCallerAuth(engine, callerauth.Disabled()) } diff --git a/docs/blocked-task-dependencies-2026-10-04.md b/docs/blocked-task-dependencies-2026-10-04.md new file mode 100644 index 0000000..e439938 --- /dev/null +++ b/docs/blocked-task-dependencies-2026-10-04.md @@ -0,0 +1,40 @@ +# Blocked task review — 2026-10-04 + +Review scope: all nonterminal task blocks in access-engine workplans. Changes +are confined to this repository and its State Hub projection. Owner source +records were inspected read-only. This record establishes acceptance +requirements; it records no new live approval, credential, identity or window. + +| Our task | Owner dependency and record | Required return / acceptance evidence | +| --- | --- | --- | +| FLEX-WP-0031-T02 | audit-core, accepted AUDIT-IN-0006 | Exact sender rows for `flex-auth.ops-warden`, `.tenant-engine`, `.user-engine`, `.secrets-engine`, `.informed-decision-t03`, `.informed-decision-sitting`; `may_write: true`, `may_read: false`, `tenants: ["*"]` with the accepted deny-for-not-yet-created-tenant justification, load-bearing evidence kind, redact policy, four heartbeat classes at 86400s, and the source-owned wire cadence declaration. Return owning revision, registration references, token custody/mount references, and native acceptance status; prepared rows alone do not establish live admission. The new fetch pin needs separate admission if it is added to the emission rollout; the six-pin intake does not cover it. | +| FLEX-WP-0031-T02 | gate-house, GH-WP-0004-T12 (todo in owner source) | A published decision disposing of FLEX-DEC-2026-018: deny/redact/not_applicable may leave after outbox failure while allow/audit_only are withheld. audit-core's accepted intake is not a statutory waiver. State whether the direction-limited exception is admitted, its declaration/detection conditions, or the required replacement semantics. | +| FLEX-WP-0031-T02 / T08 | railiance-platform custody; ops-warden routing | AUDIT-IN-0006 names the attended OpenBao sender mint into `platform/data/workloads/audit-core/senders`, using the platform attended helper. Supply the governed lane and per-pin token mount/rotation references plus owner admission receipts; never send token values to Git, State Hub or chat. Before any credential request, run `warden route find` and `warden route show` as AGENTS.md requires. No credential request or mint was made by this review. | +| FLEX-WP-0031-T06 | audit-core, reconciliation contract | Current `audit_core/postgres_backend.py:event_counts` and `sqlite_backend.py:event_counts` count **accepted_at** in `[since, until)`. Our log has **CommittedAt/OccurredAt**. Ingest returns only status/reference, with no acceptance timestamp. Return an approved matching basis: counts by source event time, or stable per-event acceptance timestamps (including duplicates) that our sender can persist. Define bounds, clock/late-delivery handling, source/tenant scope and retry/crash treatment. Identical bounds alone cannot reconcile these two clocks; missing admission is not the only blocker. | +| FLEX-WP-0031-T08 | operator, platform/network owners and audit-core observer | Approve exact immutable image, enabled pin/source, PVC size/node capacity, storage growth/retention, governed token mount and ingress/egress changes. Return native 202/duplicate-200 delivery, restart/cursor and lag evidence, plus an audit-core silence finding in a separately authorized non-production exercise. This review sets no test window or engagement. The optional chart and worker remain disabled on every pin. | +| FLEX-WP-ADHOC-2026-10-03-T02 | operator decision 877c268b-2089-42f1-a811-d98f549c1cb2, ops-bridge, ops-warden WARDEN-WP-0039-T03 | Decision is still open. Approve the exact CI digest and live install; record headroom/dry-run/rollout, Service IP and ops-bridge tunnel on 19091. ops-warden's unread return 26992e49 confirms fetch URL separation is implemented at 7d4381e/88c22d4; its task now needs configured `policy.fetch_flex_auth_url`, native health/allow/wrong-owner/fail-closed receipts. This is deployment acceptance, not missing evaluator code. | +| FLEX-WP-0020-T11 | sbom-nexus, open SBOM-IN-0001 | Owner receipt of the scheduled catch-up ingest: `access-engine` snapshot, frozen source SHA, repository UUID `fda8ad85-a7d7-4055-8f21-902a533e59df`. Owner source still shows open; elapsed time is not an ingest receipt. No manual production ingest was attempted. | +| FLEX-WP-0020-T11 | secrets-engine; railiance-fabric | secrets-engine still has `../flex-auth` in `docs/exec-owner-binding.md:89`; return the source-coordinate correction and existing handoff completion reference. railiance-fabric must return the owning residual record for its historical `repo:flex-auth` discovery node and explicitly state historical retention vs canonical mapping. No receiving task ID or completion is invented for either return. | +| FLEX-WP-0020-T11 | operator decision e071b34e-c4d4-49d0-824d-2c58bb0ae496 | Decision remains open. Name the soak bounds, observed deployment/decision/telemetry/package/automation evidence and residual owner receipts, then approve the exact obsolete checkout to remove. No soak success or cleanup is inferred. The protected flex-auth alias remains retained. | + +The three remaining root workplans are `blocked`, with task status `wait` and +explicit dependency/human gates. Completed local work is not blocked on remote +registration: FLEX-WP-0031-T05 now covers the implementation, T07 the optional +PVC and profile validation, and T08 the remaining native delivery acceptance. +T06 retains the valid-window comparator and native acceptance requirements. + +FLEX-WP-0027 is finished on the original task scope. secrets-engine's terminal +SECRETS-WP-0010-T03 and hashed native completion receipt establish human +acceptance/consumption for the September 16 version-2 cycle, actor `worsch`. +See `evidence/2026-10-04-t03-owner-completion-review.json`. The version-3 +records now in access-engine are the separate IR-WP-0005-T02 cycle; consumed +approvals cannot be replayed. The old open human decision remains recorded as +history; this review does not answer a fresh human decision for the operator. + +The whitehat return from this session also remains a gap: no recorded E2 +tenant-A identity establishes absence of tenant-B authority. Reply +`ec96e1ba-27bd-4faa-8773-55ffeaba095c` on thread +`bd8cf37e-80e8-4df9-9181-1441b31baa30` returns that gap as access-engine. E2 and +ASM targets remain pending; terminal engagements were not reused and no test +packet was sent. A real tenant-scoped identity/authority receipt must come from +its identity/surface owner before access-engine can confirm it. diff --git a/docs/conformance/security-layer-conformance.md b/docs/conformance/security-layer-conformance.md index 4ab14c6..4e45f4d 100644 --- a/docs/conformance/security-layer-conformance.md +++ b/docs/conformance/security-layer-conformance.md @@ -56,7 +56,7 @@ review date. | # | Gap | Owner | Blocked on | Review | | --- | --- | --- | --- | --- | | G1 | Registry-snapshot digest absent from decision provenance. §9.7.2 promotes it to a conformance prerequisite: a decision that turned on registry content must be replayable from its own record. | `flex-auth` | implementation | `FLEX-WP-0019` | -| G2 | **Emission guarantee declared, not delivered.** Ruled 2026-09-21 (`GH-DEC-2026-018`): flex-auth is the source; custody is never source; `audit-core` confirmed its half and declined the role (`AUDIT-IN-0005`). The per-event-class inventory and classification are **published** in [`cadence.yaml`](../../cadence.yaml) and named from `INTENT.md` (`source_of_evidence: true`, `emission_guarantee`). Five decision classes are load-bearing; `deny`, `redact`, `not_applicable`, `audit_only` are **rare** — heartbeat **and** reconciliation, rate monitoring forbidden; `allow` is **volume** — expected-rate **and** reconciliation. Nothing emits yet: no sender is registered with `audit-core`, no outbox, no heartbeat. | `flex-auth` | outbox + heartbeat implementation, and an `audit-core` sender registration (intake + token lane, per `AUDIT-IN-0002`/`0003`) — `FLEX-WP-0031` | **2026-10-19** | +| G2 | **Emission guarantee declared, not delivered.** Ruled 2026-09-21 (`GH-DEC-2026-018`): flex-auth is the source; custody is never source; `audit-core` confirmed its half and declined the role (`AUDIT-IN-0005`). The per-event-class inventory and classification are **published** in [`cadence.yaml`](../../cadence.yaml) and named from `INTENT.md` (`source_of_evidence: true`, `emission_guarantee`). Five decision classes are load-bearing; `deny`, `redact`, `not_applicable`, `audit_only` are **rare** — heartbeat **and** reconciliation, rate monitoring forbidden; `allow` is **volume** — expected-rate **and** reconciliation. No production pin emits yet. Durable outbox, per-class heartbeat/drain and optional PVC are implemented and locally tested; sender admission, compatible reconciliation windows and native observer evidence remain pending. | `flex-auth` | audit-core `AUDIT-IN-0006` admission/token delivery, gate-house `GH-WP-0004-T12` ruling, compatible count windows and approved native heartbeat/reconciliation/silence-finding evidence — `FLEX-WP-0031` | **2026-10-19** | | G3 | Published stance-register review is stale: written at two register rows, §13.1 now carries five, and its Finding 1 was ruled by v0.8 §6.4 obligation 3. | `flex-auth` | second edition | `FLEX-WP-0029` | **G2 fails closed.** A stream with no delivered cadence produces no silence @@ -96,3 +96,10 @@ heartbeat and reconciliation and forbidding rate monitoring. That is the same pr praised in `ops-warden`'s stance map: the published artifact is asserted equal to the shipped one by test, rather than merely written down. It does **not** assert the contents below the declaration — a reviewer still has to read this file. + +2026-10-04 implementation review: source-owned +`docs/contracts/decision-emission-cadence.json` passes the supplied InfoTechCanon +schema and NetKingdom security profile. This is declaration validity, not native +emission evidence. `FLEX-DEC-2026-018` supplies the local release rule; +`GH-WP-0004-T12` still owes the statutory disposition of its failure-path +exception. G2 remains open and no atomicity waiver is inferred. diff --git a/docs/contracts/decision-emission-cadence.json b/docs/contracts/decision-emission-cadence.json new file mode 100644 index 0000000..975f174 --- /dev/null +++ b/docs/contracts/decision-emission-cadence.json @@ -0,0 +1,178 @@ +{ + "schema_version": "0.1", + "declaration_id": "flex-auth-decision-emission-cadence-v0.1", + "source": "flex-auth", + "sources": [ + { + "source_id": "flex-auth:deny", + "event_class": "flex-auth.decision.deny", + "form": "heartbeat-or-reconciliation", + "reconciliation": { + "compare_local": "committed outbox counts per class, tenant and explicit since/until window, with pending deliveries counted as lag", + "compare_observed": "audit-core GET /v1/reconciliation for the exact pin source and tenant, once receiver acceptance-time windows are compatible with local counts; time-basis contract pending", + "divergence": "finding", + "undrained_local": "lag-not-divergence" + }, + "extensions": { + "netkingdom": { + "evidence_class": "load-bearing", + "rate_monitoring": "forbidden" + } + }, + "heartbeat": { + "event_class": "audit-core.heartbeat", + "interval_seconds": 86400, + "assertion": "nothing-to-report", + "missing": "finding", + "extensions": { + "flex-auth": { + "asserted_class": "flex-auth.decision.deny", + "observer_registration_status": "pending" + } + } + } + }, + { + "source_id": "flex-auth:redact", + "event_class": "flex-auth.decision.redact", + "form": "heartbeat-or-reconciliation", + "reconciliation": { + "compare_local": "committed outbox counts per class, tenant and explicit since/until window, with pending deliveries counted as lag", + "compare_observed": "audit-core GET /v1/reconciliation for the exact pin source and tenant, once receiver acceptance-time windows are compatible with local counts; time-basis contract pending", + "divergence": "finding", + "undrained_local": "lag-not-divergence" + }, + "extensions": { + "netkingdom": { + "evidence_class": "load-bearing", + "rate_monitoring": "forbidden" + } + }, + "heartbeat": { + "event_class": "audit-core.heartbeat", + "interval_seconds": 86400, + "assertion": "nothing-to-report", + "missing": "finding", + "extensions": { + "flex-auth": { + "asserted_class": "flex-auth.decision.redact", + "observer_registration_status": "pending" + } + } + } + }, + { + "source_id": "flex-auth:not_applicable", + "event_class": "flex-auth.decision.not_applicable", + "form": "heartbeat-or-reconciliation", + "reconciliation": { + "compare_local": "committed outbox counts per class, tenant and explicit since/until window, with pending deliveries counted as lag", + "compare_observed": "audit-core GET /v1/reconciliation for the exact pin source and tenant, once receiver acceptance-time windows are compatible with local counts; time-basis contract pending", + "divergence": "finding", + "undrained_local": "lag-not-divergence" + }, + "extensions": { + "netkingdom": { + "evidence_class": "load-bearing", + "rate_monitoring": "forbidden" + } + }, + "heartbeat": { + "event_class": "audit-core.heartbeat", + "interval_seconds": 86400, + "assertion": "nothing-to-report", + "missing": "finding", + "extensions": { + "flex-auth": { + "asserted_class": "flex-auth.decision.not_applicable", + "observer_registration_status": "pending" + } + } + } + }, + { + "source_id": "flex-auth:audit_only", + "event_class": "flex-auth.decision.audit_only", + "form": "heartbeat-or-reconciliation", + "reconciliation": { + "compare_local": "committed outbox counts per class, tenant and explicit since/until window, with pending deliveries counted as lag", + "compare_observed": "audit-core GET /v1/reconciliation for the exact pin source and tenant, once receiver acceptance-time windows are compatible with local counts; time-basis contract pending", + "divergence": "finding", + "undrained_local": "lag-not-divergence" + }, + "extensions": { + "netkingdom": { + "evidence_class": "load-bearing", + "rate_monitoring": "forbidden" + } + }, + "heartbeat": { + "event_class": "audit-core.heartbeat", + "interval_seconds": 86400, + "assertion": "nothing-to-report", + "missing": "finding", + "extensions": { + "flex-auth": { + "asserted_class": "flex-auth.decision.audit_only", + "observer_registration_status": "pending" + } + } + } + }, + { + "source_id": "flex-auth:allow:rate", + "event_class": "flex-auth.decision.allow", + "form": "expected-rate", + "window_seconds": 3600, + "expected_min": 1, + "drop_below": "finding", + "extensions": { + "netkingdom": { + "evidence_class": "load-bearing", + "rate_monitoring": "permitted" + }, + "flex-auth": { + "reconciliation": { + "compare_local": "committed outbox counts per class, tenant and explicit since/until window, with pending deliveries counted as lag", + "compare_observed": "audit-core GET /v1/reconciliation for the exact pin source and tenant, once receiver acceptance-time windows are compatible with local counts; time-basis contract pending", + "divergence": "finding", + "undrained_local": "lag-not-divergence" + }, + "reason": "The generic expected-rate form forbids a reconciliation block; this source-owned extension preserves the additional reconciliation guarantee without a duplicate event class." + } + } + } + ], + "extensions": { + "flex-auth": { + "state": "declared-not-yet-emitting", + "gap": "G2", + "inventory": { + "load_bearing": [ + "flex-auth.decision.deny", + "flex-auth.decision.redact", + "flex-auth.decision.not_applicable", + "flex-auth.decision.audit_only", + "flex-auth.decision.allow" + ], + "rare_load_bearing": [ + "flex-auth.decision.deny", + "flex-auth.decision.redact", + "flex-auth.decision.not_applicable", + "flex-auth.decision.audit_only" + ], + "attributive": [] + }, + "contract": { + "owner": "info-tech-canon", + "schema_sha256": "6d52692eb1e4d1325e0d6062d95acded98beddaae5bcfe24c6f0e91f1b6be6ae" + }, + "lag_bound": { + "outbox_depth": 100, + "outbox_age": "1h", + "exceed": "finding" + }, + "bound": "Heartbeat and reconciliation detect loss, outage, drain failure and accident. Neither detects a compromised source suppressing an event and its own count together \u2014 \u00a79.6's stated residual, and \u00a716 puts the independent observer outside audit-core's scope. Conformance to this declaration is not a claim of stream completeness." + } + } +} diff --git a/docs/decision-emission-operations.md b/docs/decision-emission-operations.md new file mode 100644 index 0000000..fe6d5d6 --- /dev/null +++ b/docs/decision-emission-operations.md @@ -0,0 +1,70 @@ +# Decision emission preparation + +FLEX-WP-0031 implements a durable local outbox, daily per-class heartbeats, +asynchronous delivery and optional PVC. Every pin still defaults to no outbox +and no delivery worker. `cadence.yaml` remains declared-not-yet-emitting; G2 +does not close on local tests or chart rendering. + +After sender admission and the exact approved rollout, enable the chart's +`outbox.enabled`, choose a unique `outbox.source` for the pin, and set +`outbox.size` against measured capacity and expected cumulative log growth. +`outbox.existingClaim` can reuse its own retained claim. New claims carry Helm's keep policy so uninstall does not remove the +evidence log; any eventual claim cleanup needs its own reviewed action. +The chart enforces one +replica and Recreate; each source requires its own outbox. Never run two +processes against one outbox directory. The fsGroup permits the non-root +process to write the mount while keeping its root filesystem read-only. + +Delivery requires explicit serve flags: + +```text +--audit-url +--audit-token-file +``` + +Neither flag is enabled by the chart or any checked-in production pin. These +flags require `--outbox-dir`; the source is supplied by `--emission-source`. +The worker reads the token file per event to support rotation, refuses HTTP +redirects, and does not log tokens or receiver response bodies. Receiver +authentication/network admission and token mounts remain owner dependencies. + +The worker commits due heartbeats into the same log as decisions, then drains +bounded batches in order. The daily timestamp is recovered from the committed +log after restart. Each heartbeat has its own event/correlation ID and carries +`data.class`, committed count and the process's released-uncommitted count. +Known uncommitted releases produce `released-uncommitted-finding` instead of a +nothing-to-report assertion. This count is volatile across restart, as the +existing atomicity ruling states; the heartbeat does not strengthen that bound. + +Only documented 200 (duplicate) and 202 (accepted) responses advance +`drain.json`. The cursor is fsynced, atomically replaced and its directory +synced. A rejection, timeout or cursor-write failure retains the event and +replays its original idempotency key. A failing head event blocks subsequent +delivery rather than being dropped. HTTP occurs outside the decision lock. + +`GET /v1/emission` reports committed and released-uncommitted decision counts, +pending event depth, oldest pending commit time and `lag_exceeded` when depth +exceeds 100 or age exceeds 1h. Counts include no record payload. The drain +cursor advances over heartbeats as well as decisions; pending depth includes +both. Acknowledged entries are retained in the append-only log. There is no +compaction/retention worker, so operators must plan storage and monitoring for +cumulative log growth before rollout. + +`docs/contracts/decision-emission-cadence.json` is the schema-valid wire +declaration for receiver registration. It names all five classes explicitly; +allow's additional reconciliation obligation is in a source extension because +the generic expected-rate form prohibits a reconciliation block and the +security profile prohibits duplicate classes. audit-core's parser accepts the +declaration; that does not prove native sender admission or evaluation of the +source extension. + +Run `make verify-emission-cadence` against the sibling InfoTechCanon schema and +NetKingdom profile checker. The declaration pins the schema SHA-256 and the +check rejects drift. Go tests check inventory/heartbeat parity; chart checks +exercise auth on/off, existing/new PVCs and refusal of multiple replicas. + +A valid remote comparator remains blocked: audit-core windows use receiver +accepted_at, and local windows would use commit time. The current ingest +receipt contains no acceptance timestamp. Comparing these directly would +manufacture divergence after delayed delivery. Requirements and native +acceptance gates are in `blocked-task-dependencies-2026-10-04.md`. diff --git a/docs/evidence/2026-10-04-decision-emission-profile.json b/docs/evidence/2026-10-04-decision-emission-profile.json new file mode 100644 index 0000000..2036652 --- /dev/null +++ b/docs/evidence/2026-10-04-decision-emission-profile.json @@ -0,0 +1,33 @@ +{ + "assessment_scope": "supplied-inventory", + "conformant": true, + "contract_schema": "/home/worsch/info-tech-canon/infospace/schemas/emission-cadence.schema.yaml", + "contract_valid": true, + "declaration": "/home/worsch/access-engine/docs/contracts/decision-emission-cadence.json", + "findings": [], + "inventory": { + "attributive": [], + "load_bearing": [ + "flex-auth.decision.allow", + "flex-auth.decision.audit_only", + "flex-auth.decision.deny", + "flex-auth.decision.not_applicable", + "flex-auth.decision.redact" + ], + "rare_load_bearing": [ + "flex-auth.decision.audit_only", + "flex-auth.decision.deny", + "flex-auth.decision.not_applicable", + "flex-auth.decision.redact" + ] + }, + "profile": "netkingdom-emission-cadence-security-profile-v0.1", + "profile_assessed": true, + "summary": { + "must": 0, + "should": 0 + }, + "reviewed": "2026-10-04", + "declaration_sha256": "6a4af47ebeab0d387d7088adb9601a79fa8a1eeb5c581ec9e15f0ace2977ccc5", + "scope": "Source declaration/schema/profile only; no native audit-core delivery or silence finding claimed." +} diff --git a/docs/evidence/2026-10-04-t03-owner-completion-review.json b/docs/evidence/2026-10-04-t03-owner-completion-review.json new file mode 100644 index 0000000..22e98fc --- /dev/null +++ b/docs/evidence/2026-10-04-t03-owner-completion-review.json @@ -0,0 +1,47 @@ +{ + "reviewed": "2026-10-04", + "task": "FLEX-WP-0027-T03", + "owner_record": "SECRETS-WP-0010-T03", + "source_receipt": "secrets-engine/docs/evidence/2026-09-16-t03-completion.json", + "source_sha256": "f68b38a3968373fa0811a5906756b63d9dade40475b341b12937c3668b182b3f", + "scope": "Original three T03 reviews, renewed version-2 approvals completed on 2026-09-16; excludes later version-3 IR-WP-0005 run", + "owner_human_acceptance_record": "secrets-engine/workplans/SECRETS-WP-0010-openrouter-native-access.md, T03 completed with native delivery \u2014 2026-09-16", + "native_actions": [ + { + "action": "apply", + "actor": "worsch", + "result": "applied", + "record_id": "fb256f9f-303b-4a7f-a060-71b713e2c085", + "decision_id": "decision:dc57b3b9614aa15b", + "ts": "2026-09-16T00:05:09.827448+00:00", + "approval_id": "9935335c-8e9a-566e-a48e-6a5b5f4882eb", + "approval_consumed_at": "2026-09-16T00:05:09+00:00", + "request_digest": "sha256:c5c44adc715385640dffeb3305ff854fc69bd800929f8c31f93a536524089831" + }, + { + "action": "verify", + "actor": "worsch", + "result": "pass", + "record_id": "39f8084d-c3f4-443a-8272-2cda87c46c89", + "decision_id": "decision:b4fca706e4c55b9a", + "ts": "2026-09-16T00:05:12.565647+00:00", + "approval_id": "273d6882-6253-5dc9-ac54-544f92ef5e56", + "approval_consumed_at": "2026-09-16T00:05:11+00:00", + "request_digest": "sha256:a735896e9d208032ea08b073bcc819864b12116bdf60852fcdd2e54dad2e62bf" + }, + { + "action": "exec", + "actor": "worsch", + "result": "exit-0", + "record_id": "3823a903-1a13-4d85-aef0-cc2e05c477a0", + "decision_id": "decision:82cec64febcf635f", + "ts": "2026-09-16T00:08:30.937892+00:00", + "approval_id": "7ba0c13b-68cd-5b3e-9481-42ba9e385e68", + "approval_consumed_at": "2026-09-16T00:08:32+00:00", + "request_digest": "sha256:4082d2ccad78097a8be34ff0b0221ed2c4b6407fdbaf36e4a583927ae9acc820" + } + ], + "approval_state": "All three native claims inspected as consumed after completion", + "new_execution_performed": false, + "consumed_approvals_reused": false +} diff --git a/internal/emission/delivery.go b/internal/emission/delivery.go new file mode 100644 index 0000000..0f208ba --- /dev/null +++ b/internal/emission/delivery.go @@ -0,0 +1,233 @@ +package emission + +import ( + "bufio" + "bytes" + "context" + "encoding/json" + "errors" + "fmt" + "io" + "net/http" + "net/url" + "os" + "path/filepath" + "strings" + "time" +) + +const HeartbeatType = "audit-core.heartbeat" + +// RareClasses is the source-owned inventory in cadence.yaml. +func RareClasses() []string { + return []string{ClassPrefix + "deny", ClassPrefix + "redact", ClassPrefix + "not_applicable", ClassPrefix + "audit_only"} +} + +type heartbeatData struct { + Class string `json:"class"` + Assertion string `json:"assertion"` + Committed uint64 `json:"committed"` + ReleasedUncommitted uint64 `json:"released_uncommitted"` +} + +func (o *Outbox) noteHeartbeat(event Event) { + if event.Type != HeartbeatType { + return + } + b, err := json.Marshal(event.Data) + if err != nil { + return + } + var data heartbeatData + if json.Unmarshal(b, &data) != nil { + return + } + at, err := time.Parse(time.RFC3339Nano, event.OccurredAt) + if err == nil { + o.heartbeats[data.Class] = at + } +} + +// AppendDueHeartbeats commits one heartbeat per rare class at most every 24h. +// The timestamp is rebuilt from the log on restart, never from a process timer. +func (o *Outbox) AppendDueHeartbeats() error { + o.mu.Lock() + defer o.mu.Unlock() + now := o.clock().UTC() + var events []Event + for _, class := range RareClasses() { + if at, ok := o.heartbeats[class]; ok && now.Sub(at) < 24*time.Hour { + continue + } + id, err := eventID() + if err != nil { + return err + } + assertion := "nothing-to-report" + if o.uncommitted[class] > 0 { + assertion = "released-uncommitted-finding" + } + events = append(events, Event{ID: id, Type: HeartbeatType, Source: o.source, Subject: o.source, + Tenant: PlatformTenant, CorrelationID: id, OccurredAt: now.Format(time.RFC3339Nano), + Data: heartbeatData{Class: class, Assertion: assertion, Committed: o.committed[class], ReleasedUncommitted: o.uncommitted[class]}}) + } + if len(events) == 0 { + return nil + } + return o.appendEvents(now.Format(time.RFC3339Nano), events) +} + +type drainCursor struct { + Seq uint64 `json:"seq"` + Offset int64 `json:"offset"` +} + +func (o *Outbox) loadCursor() error { + b, err := os.ReadFile(filepath.Join(filepath.Dir(o.path), "drain.json")) + if errors.Is(err, os.ErrNotExist) { + return nil + } + if err != nil { + return fmt.Errorf("read drain cursor: %w", err) + } + if err := json.Unmarshal(b, &o.cursor); err != nil { + return fmt.Errorf("decode drain cursor: %w", err) + } + if o.cursor.Offset < 0 || (o.cursor.Seq == 0) != (o.cursor.Offset == 0) { + return errors.New("invalid drain cursor") + } + return nil +} + +// nextEntry reads only the committed prefix; mu protects the size snapshot. +func (o *Outbox) nextEntry() (entry, int64, error) { + f, err := os.Open(o.path) + if err != nil { + return entry{}, 0, err + } + defer f.Close() + r := bufio.NewReader(io.NewSectionReader(f, o.cursor.Offset, o.size-o.cursor.Offset)) + line, err := r.ReadBytes('\n') + if err != nil { + return entry{}, 0, err + } + var e entry + if err := json.Unmarshal(line, &e); err != nil { + return entry{}, 0, err + } + if e.Seq != o.cursor.Seq+1 || e.Event.Source != o.source { + return entry{}, 0, errors.New("inconsistent next outbox entry") + } + return e, o.cursor.Offset + int64(len(line)), nil +} + +func (o *Outbox) saveCursor(cursor drainCursor) error { + dir := filepath.Dir(o.path) + f, err := os.CreateTemp(dir, ".drain-*") + if err != nil { + return err + } + defer os.Remove(f.Name()) + b, _ := json.Marshal(cursor) + if _, err = f.Write(b); err == nil { + err = f.Sync() + } + err = errors.Join(err, f.Close()) + if err != nil { + return err + } + if err = os.Rename(f.Name(), filepath.Join(dir, "drain.json")); err != nil { + return err + } + d, err := os.Open(dir) + if err != nil { + return err + } + defer d.Close() + return d.Sync() +} + +// DeliveryConfig is supplied only after sender admission. Tokens are read from +// a mounted file for each event to support rotation; they never enter logs. +type DeliveryConfig struct { + URL string + TokenFile string +} + +func (c DeliveryConfig) Validate() error { + u, err := url.Parse(c.URL) + if err != nil || u.Host == "" || (u.Scheme != "http" && u.Scheme != "https") || u.User != nil || u.RawQuery != "" || u.Fragment != "" || u.Path != "/v1/events" { + return errors.New("audit URL must be an explicit http(s) /v1/events endpoint without credentials, query or fragment") + } + if c.TokenFile == "" { + return errors.New("audit delivery requires a token file") + } + return nil +} + +// Drain sends a bounded batch sequentially. Only 200/202 advance the durable +// cursor. Failed or ambiguous requests replay the same event/idempotency key. +// No receiver error payload or token is included in an error. +func (o *Outbox) Drain(ctx context.Context, config DeliveryConfig, limit int) (int, error) { + if err := config.Validate(); err != nil { + return 0, err + } + if limit < 1 { + return 0, errors.New("drain limit must be positive") + } + o.drainMu.Lock() + defer o.drainMu.Unlock() + client := &http.Client{Timeout: 10 * time.Second, CheckRedirect: func(_ *http.Request, _ []*http.Request) error { return http.ErrUseLastResponse }} + delivered := 0 + for delivered < limit { + o.mu.Lock() + if o.cursor.Seq == o.seq { + o.mu.Unlock() + return delivered, nil + } + e, offset, err := o.nextEntry() + o.mu.Unlock() + if err != nil { + return delivered, fmt.Errorf("read pending event: %w", err) + } + token, err := os.ReadFile(config.TokenFile) + if err != nil { + return delivered, errors.New("audit token file unavailable") + } + value := strings.TrimSpace(string(token)) + if value == "" || strings.ContainsAny(value, "\r\n") { + return delivered, errors.New("audit token file is empty or invalid") + } + body, err := json.Marshal(e.Event) + if err != nil { + return delivered, err + } + req, err := http.NewRequestWithContext(ctx, http.MethodPost, config.URL, bytes.NewReader(body)) + if err != nil { + return delivered, errors.New("construct audit request failed") + } + req.Header.Set("Content-Type", "application/json") + req.Header.Set("Authorization", "Bearer "+value) + req.Header.Set("Idempotency-Key", e.Event.ID) + response, err := client.Do(req) + if err != nil { + return delivered, errors.New("audit delivery failed; event retained") + } + response.Body.Close() + if response.StatusCode != http.StatusOK && response.StatusCode != http.StatusAccepted { + return delivered, fmt.Errorf("audit delivery HTTP %d; event retained", response.StatusCode) + } + o.mu.Lock() + cursor := drainCursor{Seq: e.Seq, Offset: offset} + err = o.saveCursor(cursor) + if err == nil { + o.cursor = cursor + } + o.mu.Unlock() + if err != nil { + return delivered, fmt.Errorf("persist drain cursor: %w", err) + } + delivered++ + } + return delivered, nil +} diff --git a/internal/emission/delivery_test.go b/internal/emission/delivery_test.go new file mode 100644 index 0000000..a38f9ea --- /dev/null +++ b/internal/emission/delivery_test.go @@ -0,0 +1,246 @@ +package emission_test + +import ( + "context" + "encoding/json" + "net/http" + "net/http/httptest" + "os" + "path/filepath" + "strings" + "testing" + "time" + + "github.com/netkingdom/flex-auth/internal/emission" + "github.com/netkingdom/flex-auth/pkg/api" +) + +func TestHeartbeatScheduleSurvivesRestart(t *testing.T) { + dir := t.TempDir() + now := time.Date(2026, 10, 4, 0, 0, 0, 0, time.UTC) + o, err := emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + o.SetClock(func() time.Time { return now }) + if err := o.AppendDueHeartbeats(); err != nil { + t.Fatal(err) + } + if err := o.AppendDueHeartbeats(); err != nil { + t.Fatal(err) + } + o.Close() + o, err = emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + defer o.Close() + o.SetClock(func() time.Time { return now }) + if err := o.AppendDueHeartbeats(); err != nil { + t.Fatal(err) + } + if got := len(readEvents(t, dir)); got != 4 { + t.Fatalf("heartbeats after restart = %d", got) + } + seen := map[string]bool{} + for _, e := range readEvents(t, dir) { + if e["type"] != emission.HeartbeatType || e["id"] != e["correlation_id"] { + t.Fatalf("invalid heartbeat: %v", e) + } + if _, err := time.Parse(time.RFC3339Nano, e["occurred_at"].(string)); err != nil { + t.Fatal(err) + } + data := e["data"].(map[string]any) + seen[data["class"].(string)] = true + } + for _, class := range emission.RareClasses() { + if !seen[class] { + t.Fatalf("no heartbeat for %s", class) + } + } + now = now.Add(24 * time.Hour) + o.NoteReleasedUncommitted(api.DecisionEffectDeny) + if err := o.AppendDueHeartbeats(); err != nil { + t.Fatal(err) + } + if got := len(readEvents(t, dir)); got != 8 { + t.Fatalf("daily heartbeats = %d", got) + } + for _, event := range readEvents(t, dir)[4:] { + data := event["data"].(map[string]any) + if data["class"] == "flex-auth.decision.deny" && data["assertion"] != "released-uncommitted-finding" { + t.Fatal("known emission failure concealed by nothing-to-report heartbeat") + } + } +} + +func TestDrainRetainsRejectionsAndResumesWithStableIDs(t *testing.T) { + for _, rejection := range []int{400, 401, 403, 500, 302, 204} { + t.Run(http.StatusText(rejection), func(t *testing.T) { + dir := t.TempDir() + tokenFile := filepath.Join(dir, "sender-token") + if err := os.WriteFile(tokenFile, []byte("test-token"), 0600); err != nil { + t.Fatal(err) + } + o, err := emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + if err := o.AppendBatch([]api.DecisionEnvelope{deny("decision:1", "tenant:a"), deny("decision:2", "tenant:a")}); err != nil { + t.Fatal(err) + } + var ids []string + status := rejection + server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) { + var event emission.Event + if err := json.NewDecoder(r.Body).Decode(&event); err != nil { + t.Error(err) + } + if r.URL.Path != "/v1/events" || r.Header.Get("Idempotency-Key") != event.ID || r.Header.Get("Authorization") != "Bearer test-token" { + t.Errorf("bad delivery headers/path") + } + ids = append(ids, event.ID) + w.Header().Set("Location", "/redirect-must-not-be-followed") + w.WriteHeader(status) + w.Write([]byte("sensitive receiver payload")) + })) + defer server.Close() + cfg := emission.DeliveryConfig{URL: server.URL + "/v1/events", TokenFile: tokenFile} + if n, err := o.Drain(context.Background(), cfg, 100); n != 0 || err == nil || strings.Contains(err.Error(), "sensitive") { + t.Fatalf("rejection advanced or exposed response: %d %v", n, err) + } + if o.Status().Pending != 2 { + t.Fatal("rejected event lost") + } + status = 202 + if n, err := o.Drain(context.Background(), cfg, 1); n != 1 || err != nil { + t.Fatalf("accepted: %d %v", n, err) + } + o.Close() + o, err = emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + defer o.Close() + if o.Status().Pending != 1 { + t.Fatal("cursor did not survive restart") + } + status = 200 + if n, err := o.Drain(context.Background(), cfg, 100); n != 1 || err != nil { + t.Fatalf("duplicate: %d %v", n, err) + } + if len(ids) != 3 || ids[0] != ids[1] || ids[1] == ids[2] { + t.Fatalf("unstable/reused event IDs: %v", ids) + } + if o.Status().Pending != 0 { + t.Fatal("pending after drain") + } + }) + } +} + +func TestDrainAmbiguousAcceptanceReplaysAndTokenRotates(t *testing.T) { + dir := t.TempDir() + tokenFile := filepath.Join(dir, "token") + os.WriteFile(tokenFile, []byte("first"), 0600) + o, err := emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + defer o.Close() + o.Append(deny("decision:1", "tenant:a")) + var ids, tokens []string + server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) { + ids = append(ids, r.Header.Get("Idempotency-Key")) + tokens = append(tokens, r.Header.Get("Authorization")) + w.WriteHeader(202) + })) + defer server.Close() + cfg := emission.DeliveryConfig{URL: server.URL + "/v1/events", TokenFile: tokenFile} + // Receiver accepts, but persisting the cursor fails. Retain and retry. + os.Mkdir(filepath.Join(dir, "drain.json"), 0700) + if n, err := o.Drain(context.Background(), cfg, 1); n != 0 || err == nil { + t.Fatalf("cursor failure: %d %v", n, err) + } + os.Remove(filepath.Join(dir, "drain.json")) + os.WriteFile(tokenFile, []byte("rotated"), 0600) + if n, err := o.Drain(context.Background(), cfg, 1); n != 1 || err != nil { + t.Fatalf("retry: %d %v", n, err) + } + if len(ids) != 2 || ids[0] != ids[1] || tokens[1] != "Bearer rotated" { + t.Fatalf("retry/rotation: %v %v", ids, tokens) + } +} + +func TestOutboxReportsLagAndRefusesSourceReuse(t *testing.T) { + dir := t.TempDir() + o, err := emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + now := time.Now().UTC() + o.SetClock(func() time.Time { return now }) + o.Append(deny("decision:1", "tenant:a")) + if o.Status().LagExceeded { + t.Fatal("fresh event treated as lag finding") + } + now = now.Add(2 * time.Hour) + if !o.Status().LagExceeded { + t.Fatal("overdue event not reported") + } + o.Close() + if other, err := emission.Open(dir, "flex-auth.other"); err == nil { + other.Close() + t.Fatal("source changed on existing outbox") + } +} + +func TestSlowReceiverDoesNotBlockDecisionCommit(t *testing.T) { + dir := t.TempDir() + tokenFile := filepath.Join(dir, "token") + if err := os.WriteFile(tokenFile, []byte("local-test-token"), 0600); err != nil { + t.Fatal(err) + } + o, err := emission.Open(dir, "flex-auth.test") + if err != nil { + t.Fatal(err) + } + defer o.Close() + if err := o.Append(deny("decision:1", "tenant:a")); err != nil { + t.Fatal(err) + } + arrived, release := make(chan struct{}), make(chan struct{}) + server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) { + close(arrived) + <-release + w.WriteHeader(202) + })) + defer server.Close() + drained := make(chan error, 1) + go func() { + _, err := o.Drain(context.Background(), emission.DeliveryConfig{URL: server.URL + "/v1/events", TokenFile: tokenFile}, 1) + drained <- err + }() + defer func() { + close(release) + if err := <-drained; err != nil { + t.Error(err) + } + }() + <-arrived + committed := make(chan error, 1) + go func() { committed <- o.Append(deny("decision:2", "tenant:a")) }() + select { + case err := <-committed: + if err != nil { + t.Fatal(err) + } + case <-time.After(3 * time.Second): + t.Fatal("remote HTTP holds the decision commit lock") + } + // Deferred release lets the drain finish even on test failure. Assert the + // pending count while HTTP is still blocked rather than waiting for it. + if o.Status().Pending != 2 { + t.Fatal("new commit lost during delivery") + } +} diff --git a/internal/emission/outbox.go b/internal/emission/outbox.go index e7f67e3..57061eb 100644 --- a/internal/emission/outbox.go +++ b/internal/emission/outbox.go @@ -61,7 +61,10 @@ type Status struct { ReleasedUncommitted map[string]uint64 `json:"released_uncommitted"` // ReleasedUncommitted is held in memory: a restart during an outbox outage // loses it (FLEX-DEC-2026-018, consequences). - Means string `json:"means"` + Means string `json:"means"` + Pending uint64 `json:"pending"` + OldestPendingAt string `json:"oldest_pending_at,omitempty"` + LagExceeded bool `json:"lag_exceeded"` } // Outbox is a durable, append-only, fsync-per-commit event log. @@ -76,6 +79,9 @@ type Outbox struct { size int64 // bytes committed; a failed commit is cut back to it committed map[string]uint64 uncommitted map[string]uint64 + drainMu sync.Mutex + cursor drainCursor + heartbeats map[string]time.Time } // Open opens or creates the outbox in dir and rebuilds committed counts from it. @@ -92,10 +98,17 @@ func Open(dir, source string) (*Outbox, error) { clock: time.Now, committed: map[string]uint64{}, uncommitted: map[string]uint64{}, + heartbeats: map[string]time.Time{}, + } + if err := o.loadCursor(); err != nil { + return nil, err } if err := o.load(); err != nil { return nil, err } + if o.cursor.Seq > o.seq || o.cursor.Offset > o.size { + return nil, errors.New("drain cursor is ahead of committed outbox") + } file, err := os.OpenFile(o.path, os.O_CREATE|os.O_APPEND|os.O_WRONLY, 0o640) if err != nil { return nil, fmt.Errorf("open outbox: %w", err) @@ -126,11 +139,21 @@ func (o *Outbox) load() error { if err := json.Unmarshal(data[good:good+end], &e); err != nil { return fmt.Errorf("outbox line %d is corrupt: %w", line, err) } + if e.Seq != o.seq+1 || e.Event.Source != o.source { + return fmt.Errorf("outbox line %d has inconsistent sequence or source", line) + } o.seq = e.Seq o.committed[e.Event.Type]++ good += end + 1 + if e.Seq == o.cursor.Seq && int64(good) != o.cursor.Offset { + return errors.New("drain cursor does not match outbox") + } + o.noteHeartbeat(e.Event) } o.size = int64(good) + if o.cursor.Seq > o.seq { + return errors.New("drain cursor is ahead of outbox") + } return nil } @@ -159,15 +182,24 @@ func (o *Outbox) AppendBatch(decisions []api.DecisionEnvelope) error { o.mu.Lock() defer o.mu.Unlock() now := o.clock().UTC().Format(time.RFC3339Nano) - var buf []byte - seq := o.seq + var events []Event for _, decision := range decisions { id, err := eventID() if err != nil { return err } + events = append(events, o.event(id, now, decision)) + } + return o.appendEvents(now, events) +} + +// appendEvents requires mu; heartbeats use the same durable commit as decisions. +func (o *Outbox) appendEvents(now string, events []Event) error { + var buf []byte + seq := o.seq + for _, event := range events { seq++ - line, err := json.Marshal(entry{Seq: seq, CommittedAt: now, Event: o.event(id, now, decision)}) + line, err := json.Marshal(entry{Seq: seq, CommittedAt: now, Event: event}) if err != nil { return fmt.Errorf("marshal outbox entry: %w", err) } @@ -181,8 +213,9 @@ func (o *Outbox) AppendBatch(decisions []api.DecisionEnvelope) error { } o.seq = seq o.size += int64(len(buf)) - for _, decision := range decisions { - o.committed[Class(decision.Effect)]++ + for _, event := range events { + o.committed[event.Type]++ + o.noteHeartbeat(event) } return nil } @@ -217,6 +250,16 @@ func (o *Outbox) Status() Status { "released_uncommitted is decisions released without a record " + "since this process started, and any non-zero value is a finding", } + status.Pending = o.seq - o.cursor.Seq + if status.Pending > 0 { + if e, _, err := o.nextEntry(); err == nil { + status.OldestPendingAt = e.CommittedAt + at, _ := time.Parse(time.RFC3339Nano, e.CommittedAt) + status.LagExceeded = status.Pending > 100 || o.clock().Sub(at) > time.Hour + } else { + status.LagExceeded = true + } + } for _, effect := range api.DecisionEffects() { class := Class(effect) status.Committed[class] = o.committed[class] diff --git a/internal/emission/outbox_test.go b/internal/emission/outbox_test.go index e2e7209..229fc66 100644 --- a/internal/emission/outbox_test.go +++ b/internal/emission/outbox_test.go @@ -4,6 +4,7 @@ import ( "os" "path/filepath" "sort" + "strings" "testing" "gopkg.in/yaml.v3" @@ -150,6 +151,11 @@ func TestCadenceClassifiesEveryEffect(t *testing.T) { Action string `yaml:"action"` EvidenceClass string `yaml:"evidence_class"` } `yaml:"classes"` + Heartbeat struct { + Classes []struct { + Class string `yaml:"class"` + } `yaml:"classes"` + } `yaml:"heartbeat"` } if err := yaml.Unmarshal(data, &cadence); err != nil { t.Fatal(err) @@ -177,4 +183,17 @@ func TestCadenceClassifiesEveryEffect(t *testing.T) { t.Fatalf("cadence.yaml load-bearing classes %v; want %v", declared, want) } } + var heartbeatClasses []string + for _, class := range cadence.Heartbeat.Classes { + heartbeatClasses = append(heartbeatClasses, class.Class) + } + rare := emission.RareClasses() + sort.Strings(heartbeatClasses) + sort.Strings(rare) + if strings.Join(heartbeatClasses, ",") != strings.Join(rare, ",") { + t.Fatalf("runtime heartbeat inventory %v differs from declared %v", rare, heartbeatClasses) + } + if cadence.Classes["heartbeat"].Action != emission.HeartbeatType { + t.Fatal("heartbeat wire class differs from declaration") + } } diff --git a/internal/policy/informed_decision_attribute_test.go b/internal/policy/informed_decision_attribute_test.go index 02b7a4e..b1bcaa7 100644 --- a/internal/policy/informed_decision_attribute_test.go +++ b/internal/policy/informed_decision_attribute_test.go @@ -95,14 +95,16 @@ func TestInformedDecisionAttributeAdmission(t *testing.T) { } } -func TestSittingImagePinUnchanged(t *testing.T) { +func TestSittingImagePinMatchesRecordedBuild(t *testing.T) { raw, err := os.ReadFile(filepath.Join(repoRoot(t), "values", "informed-decision-sitting.yaml")) if err != nil { t.Fatal(err) } text := string(raw) - if !strings.Contains(text, "sha256:f6fa2062f7095e66c7a91f20cd130914b5481d46d9d909339e028f7ecdf0065d") { - t.Fatal("sitting image digest changed") + // values/informed-decision-sitting.yaml records the v9 image of 103d8b6, + // pinned in 257a83e. The earlier v7 pin is no longer the baseline. + if !strings.Contains(text, "sha256:aa7ccbafb2a8f884c97119e295fd168fa4e67e974682a8c30b35149201469eff") { + t.Fatal("sitting image digest differs from the recorded v9 build") } if !strings.Contains(text, "/opt/flex-auth/examples/informed-decision-sitting/policy.md") { t.Fatal("sitting policy path changed") diff --git a/tests/stage1.sh b/tests/stage1.sh index 1bac5c9..0531d95 100755 --- a/tests/stage1.sh +++ b/tests/stage1.sh @@ -5,6 +5,7 @@ set -euo pipefail cd "$(dirname "${BASH_SOURCE[0]}")/.." python3 tests/validate_app_toml.py +python3 tests/test_outbox_chart.py for values in values/stage1.yaml values/stage2-canary.yaml values/stage3-production.yaml \ values/user-engine.yaml values/tenant-engine.yaml values/ops-warden.yaml; do diff --git a/tests/test_outbox_chart.py b/tests/test_outbox_chart.py new file mode 100644 index 0000000..6f506e0 --- /dev/null +++ b/tests/test_outbox_chart.py @@ -0,0 +1,50 @@ +#!/usr/bin/env python3 +"""Render the optional outbox with caller auth on/off, without cluster access.""" +import subprocess +import yaml + + +def render(*settings, succeeds=True): + command = ["helm", "template", "flex-auth", "charts/flex-auth", "-f", "values/stage2-canary.yaml"] + for setting in settings: + command += ["--set", setting] + result = subprocess.run(command, capture_output=True, text=True, check=False) + if not succeeds: + assert result.returncode != 0, "invalid multi-writer configuration was accepted" + return [] + assert result.returncode == 0, result.stderr + return list(yaml.safe_load_all(result.stdout)) + + +def deployment(documents): + return next(d for d in documents if d["kind"] == "Deployment") + + +base = render() +assert not any(d["kind"] == "PersistentVolumeClaim" for d in base) +assert "--outbox-dir" not in deployment(base)["spec"]["template"]["spec"]["containers"][0]["args"] + +enabled = ("outbox.enabled=true", "outbox.source=flex-auth.test") +for auth in ((), ("callerAuth.mode=enforce", "callerAuth.binding=test=system:serviceaccount:test:test")): + documents = render(*enabled, *auth) + pod = deployment(documents)["spec"]["template"]["spec"] + container = pod["containers"][0] + args = container["args"] + assert args[args.index("--outbox-dir") + 1] == "/var/lib/flex-auth/outbox" + assert args[args.index("--emission-source") + 1] == "flex-auth.test" + assert deployment(documents)["spec"]["strategy"]["type"] == "Recreate" + pvc = next(d for d in documents if d["kind"] == "PersistentVolumeClaim") + assert pvc["spec"]["accessModes"] == ["ReadWriteOnce"] + assert pvc["spec"]["resources"]["requests"]["storage"] == "1Gi" + assert any(v.get("persistentVolumeClaim", {}).get("claimName") == pvc["metadata"]["name"] for v in pod["volumes"]) + assert any(v["name"] == "flex-auth-outbox" and not v.get("readOnly", False) for v in container["volumeMounts"]) + if auth: + assert any(v["name"] == "flex-auth-reviewer" for v in pod["volumes"]) + +existing = render(*enabled, "outbox.existingClaim=retained-outbox") +assert not any(d["kind"] == "PersistentVolumeClaim" for d in existing) +pod = deployment(existing)["spec"]["template"]["spec"] +assert any(v.get("persistentVolumeClaim", {}).get("claimName") == "retained-outbox" for v in pod["volumes"]) +render(*enabled, "replicaCount=2", succeeds=False) +render("outbox.enabled=true", succeeds=False) +print("optional outbox chart checks passed") diff --git a/tools/verify_emission_cadence.py b/tools/verify_emission_cadence.py new file mode 100644 index 0000000..7dec95b --- /dev/null +++ b/tools/verify_emission_cadence.py @@ -0,0 +1,29 @@ +#!/usr/bin/env python3 +"""Verify the source declaration against the pinned owner schema/profile.""" +import argparse +import hashlib +import json +from pathlib import Path +import subprocess +import sys + +root = Path(__file__).resolve().parents[1] +parser = argparse.ArgumentParser() +parser.add_argument("--contract-schema", type=Path, default=root.parent / "info-tech-canon/infospace/schemas/emission-cadence.schema.yaml") +parser.add_argument("--checker", type=Path, default=root.parent / "net-kingdom/tools/emission-cadence-profile/emission_cadence_profile.py") +args = parser.parse_args() +declaration = root / "docs/contracts/decision-emission-cadence.json" +document = json.loads(declaration.read_text()) +extension = document["extensions"]["flex-auth"] +if not args.contract_schema.is_file() or not args.checker.is_file(): + parser.error("owner schema/profile is unavailable; supply explicit --contract-schema and --checker paths") +if hashlib.sha256(args.contract_schema.read_bytes()).hexdigest() != extension["contract"]["schema_sha256"]: + parser.error("owner schema changed; review and update the declared schema pin before assessment") +command = [sys.executable, str(args.checker), "--contract-schema", str(args.contract_schema)] +inventory = extension["inventory"] +for event in inventory["load_bearing"]: + command += ["--load-bearing", event] +for event in inventory["rare_load_bearing"]: + command += ["--rare-load-bearing", event] +command += [str(declaration)] +sys.exit(subprocess.run(command, check=False).returncode) diff --git a/workplans/ADHOC-2026-10-03.md b/workplans/ADHOC-2026-10-03.md index 9016945..368cfb0 100644 --- a/workplans/ADHOC-2026-10-03.md +++ b/workplans/ADHOC-2026-10-03.md @@ -91,3 +91,14 @@ Steps, none run yet: Close-down checklist row 1 (the-custodian) is updated by message, not by edit. 2026-10-04 CUST-WP-0075 owner routing: `FLEX-WP-ADHOC-2026-10-03-T02` remains a human wait under open decision `877c268b-2089-42f1-a811-d98f549c1cb2`. Exact scoped human act and native receipts remain required; this metadata records no approval or execution. + +2026-10-04 blocked-task review: ops-warden message 26992e49 confirms its +fetch/sign separation is implemented at 7d4381e (sync 88c22d4). The evaluator +code is already complete; do not repeat that implementation or fall back to +the signing URL. Remaining requirements are the exact CI digest/live install, +ops-bridge 19091 tunnel, configured policy.fetch_flex_auth_url and native +health/positive/wrong-owner/fail-closed receipts under WARDEN-WP-0039-T03. +Decision 877c268b-2089-42f1-a811-d98f549c1cb2 was checked and remains open. +This task and workplan stay blocked on the approved rollout; no live change was +made. Requirements are recorded in +`docs/blocked-task-dependencies-2026-10-04.md`. diff --git a/workplans/FLEX-WP-0020-repository-identity-migration.md b/workplans/FLEX-WP-0020-repository-identity-migration.md index 0f1b400..dafad7a 100644 --- a/workplans/FLEX-WP-0020-repository-identity-migration.md +++ b/workplans/FLEX-WP-0020-repository-identity-migration.md @@ -479,7 +479,8 @@ generic error is treated as proof of rollback. id: FLEX-WP-0020-T11 status: wait needs_human: true -blocking_reason: "Soak window after T10; destructive cleanup needs separate human approval." +blocked_on: message-from:sbom-nexus +blocking_reason: "Needs SBOM-IN-0001 catch-up receipt, secrets-engine coordinate correction and railiance-fabric residual disposition, observed soak evidence, and exact checkout cleanup approval under e071b34e-c4d4-49d0-824d-2c58bb0ae496." priority: high state_hub_task_id: "46d8318b-d1fb-52e3-b632-52230ed1438f" decision_id: "e071b34e-c4d4-49d0-824d-2c58bb0ae496" @@ -514,3 +515,12 @@ the historical repo:flex-auth node in the railiance-fabric discovery snapshot. T destructive cleanup of old aliases/redirects needs separate human approval) remains. 2026-10-04 CUST-WP-0075 owner routing: `FLEX-WP-0020-T11` remains a human wait under open decision `e071b34e-c4d4-49d0-824d-2c58bb0ae496`. Exact scoped human act and native receipts remain required; this metadata records no approval or execution. Completed predecessor edges removed: FLEX-WP-0020-T10; completed source evidence remains in the history and does not satisfy the human gate. + +2026-10-04 blocked-task review: owner source still records SBOM-IN-0001 open, +and secrets-engine/docs/exec-owner-binding.md:89 still names ../flex-auth. +Neither is closed by the rename's completed journal. The discovery snapshot +residual needs railiance-fabric's receiving record and disposition. No soak +window or success is invented; no checkout or alias is removed. Exact owner +receipts and operator acceptance requirements are in +`docs/blocked-task-dependencies-2026-10-04.md`. T11 and this workplan remain +blocked until those returns and the still-open cleanup decision permit closure. diff --git a/workplans/FLEX-WP-0027-t03-human-review.md b/workplans/FLEX-WP-0027-t03-human-review.md index d02b646..32ee87e 100644 --- a/workplans/FLEX-WP-0027-t03-human-review.md +++ b/workplans/FLEX-WP-0027-t03-human-review.md @@ -4,7 +4,7 @@ type: workplan title: "Admit scoped human review for the three T03 actions" domain: infotech repo: flex-auth -status: blocked +status: finished flavor: implementation owner: codex topic_slug: netkingdom @@ -44,9 +44,8 @@ checks pass with synthetic identities; evidence is retained in docs/evidence. ```task id: FLEX-WP-0027-T03 -status: wait -needs_human: true -blocking_reason: "Needs the operator's exact signed-in account and actual human acknowledgements; synthetic policy checks are not acceptance." +status: done +needs_human: false priority: high state_hub_task_id: "9417d64a-308c-566f-af29-217c5c45d294" decision_id: "e1868a8d-cad7-48c6-84d1-78a5a566d4a4" @@ -62,3 +61,16 @@ account exercising the review), so the workplan is marked `blocked` rather than `active` until that happens. 2026-10-04 CUST-WP-0075 owner routing: `FLEX-WP-0027-T03` remains a human wait under open decision `e1868a8d-cad7-48c6-84d1-78a5a566d4a4`. Exact scoped human act and native receipts remain required; this metadata records no approval or execution. + +2026-10-04 owner-evidence review: the historical wait above is superseded. +`SECRETS-WP-0010-T03` is done; its 2026-09-16 completion record explicitly +records real human acceptance of all three renewed version-2 memos. The native +receipt retains actor `worsch`, approved human-controlled CAS consumption and +decision/request bindings for apply, verify and exec; all three approvals are +consumed. Reviewed receipt hash and action references are retained in +`docs/evidence/2026-10-04-t03-owner-completion-review.json`. This closes the +original scoped review task and finishes this workplan without new execution. +The version-3 records now in `examples/informed-decision-t03/records.json` +serve the next `IR-WP-0005-T02` cycle; they do not reopen this original task or +permit reuse of the consumed approvals. The old decision link is retained as +history, not a claim that the later human decision has been answered. diff --git a/workplans/FLEX-WP-0031-decision-record-emission.md b/workplans/FLEX-WP-0031-decision-record-emission.md index 7c66357..dcf7ca0 100644 --- a/workplans/FLEX-WP-0031-decision-record-emission.md +++ b/workplans/FLEX-WP-0031-decision-record-emission.md @@ -14,7 +14,7 @@ related_workplans: - FLEX-WP-0030 - FLEX-WP-0019 created: "2026-09-21" -updated: "2026-09-27" +updated: "2026-10-04" state_hub_workstream_id: "84f5d9fe-b4c9-584a-b964-efe3e48af095" --- @@ -63,7 +63,8 @@ and the node disk is at 84 %. id: FLEX-WP-0031-T02 status: wait blocked_on: message-from:audit-core -blocking_reason: "audit-core intake: sender registration, evidence_kind and heartbeat classes as cadence.yaml publishes them; gate is a registered sender." +depends_on: [GH-WP-0004-T12] +blocking_reason: "AUDIT-IN-0006 sender rows and governed token delivery remain pending; gate-house GH-WP-0004-T12 must rule on the load-bearing failure-path exception." priority: high state_hub_task_id: "89661908-ca9a-5e4a-a0a5-62d1a9e02568" ``` @@ -135,33 +136,36 @@ Done 2026-09-23: - `make test` (race) passes, and a local serve smoke test committed and counted one deny and one allow. Nothing is deployed and no chart changed. -## 5. Heartbeat and drain to audit-core +## 5. Implement heartbeat and drain to audit-core ```task id: FLEX-WP-0031-T05 -status: wait -depends_on: [FLEX-WP-0031-T02] -blocking_reason: "Per-class heartbeat sender and drain need the registered sender from T02." +status: done priority: high state_hub_task_id: "14bd6648-11da-53d7-a512-027005bd4772" ``` -2026-09-27: `cadence.yaml`'s heartbeat declaration corrected — it now names one -class per rare load-bearing event (`deny`, `redact`, `not_applicable`, -`audit_only`) instead of a single combined `flex-auth.decision.heartbeat` -class. `go build`/`go test ./...` pass unchanged. The remaining scope — the -actual per-class heartbeat sender and the drain to `POST /v1/events` — is -correctly blocked on T02's sender registration with `audit-core`; the task -stays `wait` rather than in progress here. +2026-10-04: split implementation from native delivery acceptance (new T08). +`internal/emission/delivery.go` implements daily per-class `audit-core.heartbeat` +events with independent correlation IDs and offset timestamps, persisted in +our durable outbox. Restart rebuilds each class's last heartbeat from the log. +Heartbeats expose committed and released-uncommitted counts; a known failure +changes the assertion to `released-uncommitted-finding`. -- Emit a daily `audit-core.heartbeat` event per rare class, with `data.class` - set to the class, following audit-core's `stream_findings` shape. -- Correct `cadence.yaml`, which still names a single - `flex-auth.decision.heartbeat` class. -- Drain committed events to `POST /v1/events` with `Idempotency-Key` set to the - event id, keep a drain cursor, and expose outbox depth and age against the - `lag_bound`. -- The drain stays disabled until T02 admits the senders. +The asynchronous drain sends the eight-field envelope with event-ID +idempotency, reads a mounted token file per event for rotation, refuses +redirects, and advances a fsynced/renamed cursor only on 200 or 202. Every other +response, transport failure or cursor-write failure retains the event. A +bounded drain runs each minute, without holding the decision lock during HTTP. +`GET /v1/emission` exposes pending depth, oldest pending timestamp and lag-bound +exceedance (100 events / 1h). No response body or token is logged. + +`serve --audit-url --audit-token-file ` +enables the worker only with an outbox. Both default empty, and no pin enables +it. Local tests cover per-class cadence, restart, rejection, stable replay, +ambiguous acceptance, token rotation and cursor durability. The current outbox +retains acknowledged log entries; retention/compaction is not implemented, so +capacity planning must include cumulative log growth. ## 6. Reconciliation, profile check, and storage @@ -169,7 +173,8 @@ stays `wait` rather than in progress here. id: FLEX-WP-0031-T06 status: wait depends_on: [FLEX-WP-0031-T02] -blocking_reason: "Reconciliation compare needs audit-core GET /v1/reconciliation, which exists only once T02's sender is registered." +blocked_on: message-from:audit-core +blocking_reason: "Receiver counts use accepted_at, while local counts use commit time; audit-core must supply compatible window semantics or durable acceptance timestamps before a valid comparator can be implemented." priority: high state_hub_task_id: "bf92a951-3b69-59dd-8907-f6748c91a264" ``` @@ -193,13 +198,22 @@ undone. Stays `wait`. (84 % on 2026-09-23). Rolling it out to any pin is a production change and needs the founder's go-ahead. +2026-10-04: storage/profile preparation is now complete as T07. The remaining +comparator is independently blocked by receiver time semantics, not just a +missing registration: audit-core counts accepted_at in half-open windows, +while our log timestamps local commits. Ingest returns status/reference only. +A receiver-approved matching basis or stable acceptance-time receipts (also +on duplicates) is required before implementing a comparison. Do not widen +may_read to fetch archive payloads as a workaround. T02 and T08 still gate +native custody/rollout evidence; see the dependency record for exact returns. + ## 4. Close G2 ```task id: FLEX-WP-0031-T04 status: wait -depends_on: [FLEX-WP-0031-T05, FLEX-WP-0031-T06] -blocking_reason: "Closing G2 needs the silence-finding gate, which needs T05's sender and T06's reconciliation." +depends_on: [FLEX-WP-0031-T08, FLEX-WP-0031-T06] +blocking_reason: "G2 needs T08 native heartbeat/drain and silence-finding evidence plus T06 reconciliation; local tests cannot close the delivery gap." priority: medium state_hub_task_id: "423b3090-1b72-58fd-9353-5c907c7683bb" ``` @@ -212,3 +226,57 @@ Change `cadence.yaml` `state` to emitting, move G2 out of the gap table with the evidence, and tell `gate-house`, `audit-core` and `kings-guard`. Gate: a silence finding is observed on a deliberately withheld heartbeat in a non-production run. + +## 7. Prepare persistent storage and validate the cadence contract + +```task +id: FLEX-WP-0031-T07 +status: done +priority: high +state_hub_task_id: "151fe609-ba68-506b-9c2c-9fdb0b149eef" +``` + +Split from T06 on 2026-10-04; its historical schema/storage blockers above are +superseded by this implementation. `charts/flex-auth` has optional outbox PVC, +mount, source and flags, defaults disabled, a one-replica constraint and +Recreate strategy to avoid two writers during rollout. It supports an existing +claim. Rendering is tested with caller auth enabled and disabled. No values pin +or live deployment is changed; 1Gi is a configurable preparation value, not an +assertion of available node space. + +Published `docs/contracts/decision-emission-cadence.json` against the actual +InfoTechCanon wire schema and NetKingdom profile. All five decision classes +are explicitly supplied; contract/profile assessment passes. The generic +expected-rate schema forbids reconciliation on that entry and the profile +forbids duplicate classes, so allow's additional reconciliation guarantee is +in the `extensions.flex-auth` namespace. Receiver acceptance of that mapping +remains part of T02; a profile pass does not establish observer delivery. + +## 8. Admit and exercise native heartbeat/drain delivery + +```task +id: FLEX-WP-0031-T08 +status: wait +priority: high +depends_on: [FLEX-WP-0031-T02] +needs_human: true +blocking_reason: "Needs admitted sender/token mounts, founder approval of exact image/PVC/network rollout, and audit-core observer evidence; no production delivery is enabled." +state_hub_task_id: "b3b52b30-c73d-5a90-9935-d98afe5a9eaf" +``` + +Remaining acceptance split from T05: audit-core's owner must return admitted +exact per-pin sources, scope, heartbeat classes and cadence mapping under +AUDIT-IN-0006. Platform must provide governed attended token custody and +mount references, never values. The founder must approve exact immutable image, +PVC capacity and rollout; network owners must admit the decision pins to the +receiver. Record native 202 and idempotent 200 delivery, restart cursor/lag +behaviour, and an audit-core silence finding from a separately authorized +non-production withheld-heartbeat exercise. No test engagement/window is +created by this workplan. T06 still owns the compatible-time reconciliation +implementation and its native acceptance. See +`docs/blocked-task-dependencies-2026-10-04.md` for the concrete dependency gates. + +2026-10-04 review: workplan stays blocked. Local heartbeat/drain and storage +preparation no longer wait on another repository, but admission, statutory +ruling, valid reconciliation time basis and approved native acceptance do. +`cadence.yaml` remains `declared-not-yet-emitting` and G2 remains open.