Repository navigation
v4.1.0 — Cybernetic Governance Engine
Release Summary: Corrective verification and compliance-custody release backed by the matched-SHA two-project Gate W certification archive at commit
fa7721221d44a53d01def9d810dfbd01c96d4293(gs://<PROJECT_ID>-evidence-worm-staging/gate-w-certification/fa772122/,archive_manifest_sha256: 70513815ba2cfe7f57a1bf5d5106528cad5f605387307717569bd254e6551be4,compliance/gate_w_certification.json). Ships everything accumulated in[Unreleased]sincev4.0.0(cage-audit/3.0stream invariants, live repository-ruleset verification,CAGE_INFERENCE_BACKEND, ValkeyAOFpersistence, model-free kernel/install isolation) together with the full Wave 1–4 andP0–P2GKE remediation (#675–#730): custom-VPC binding (CP-6), derived data-store/32egress (DE-1), non-loopback Linkerd mesh identity (LT-1), blocking refusal/deferral evidence commits (BR-1–BR-3),deployment/k8s/shadow-manifest retirement (KM-1–KM-2), in-cluster advisor proxy routing and GEAP/AGW retirement (AR-1,EA-1–EA-3), human OIDC + WebAuthn HITL quorum binding (HI-1), mesh-authenticatedINFRAcustody (IC-1–IC-4), server-bound Python STPA enforcement (ST-1), native KubernetesinitContainersOPA UDS sidecar (OP-1–OP-2), per-workload network/mesh segmentation (NS-1–NS-2), persistent 120-day CMEK evidence domain and bound cross-teardown genesis rotation (ED-1), KMS Cloud Audit Log corroboration, in-cluster Lulav0.16.0CronJob auditing (26/26active specs satisfied understartup_grace_hours == 0) with a2 × scheduledead-man's switch, and regionalEU_ECB/APAC_MASscope alignment (5adopter-operated institutional controls markednot-applicable,gdpr-art22split betweenHOLDSruntime enforcement andplannedEU-006contestation intake). Numbered4.1.0rather than5.0.0because this work makes4.0.0's existing governance, identity, and evidence-custody claims true on GKE rather than introducing a new major capability generation; seedocs/BREAKING_CHANGES.md(IB-1–OP-2) and thev4.0.0post-release errata below. Ships alongsideclient-v0.3.0, the first functional release ofpackages/cage-client.
Added
-
Matched-SHA two-project Gate W certification (
fa772122), persistent WORM/KMS custody corroboration, in-cluster Lulav0.16.0auditor + dead-man's switch, nativeinitContainersOPA sidecar, regional SSP scope alignment, andcage-clientv0.3.0(#675–#730) (gate_w_certification.json,staging_lifecycle.sh,evidence_domain,lula_cron,gateway,policy.py,packages/cage-client,POAM.md).- Gate W empirical certification (
fa772122,2026-11-12): Verified at identicalmainSHA across both a zero-state clean-room GKE project (fresh_project_clean_room, max cold pull+queue156.247svs600sPhase 2 timeout archived inkubelet-image-pull-durations-*.json) and a recovery validation project (fresh_project_with_phase0_kms_residue, exercising tombstonedDELETEDWIF pool/provider undelete,--import-staterecovery of the Binary Authorization Grafeas note with surviving image occurrences + KMS/WORM set, andread_worm_chain_anchor.pycold-start binding to a prior multi-batch WORM chain): Phase 1 two-stage zero-driftterraform plan, Phase 29/9workloadsReadywithrestartCount == 0, Phase 326/26active Lulav0.16.0specssatisfied(0not-satisfied) understartup_grace_hours == 0, Phase 4US_FED/EU_ECB/APAC_MAS/residency/Langfuse compliance isolation, Phase 5 live sweeps (live_external,integration,12/12test_gke_trade_flow.pypreview -> 2-human OIDC+WebAuthn quorum ->POST_HITLexecution,safety_rate == 1.0withblocked_traces == 0across all 5 controls, GMPcage_*timeSeries, bidirectional KMS Cloud Audit Log corroboration), and Phase 6 intra-run teardown -> rebuild (bound_rotations >= 1,declared_gaps == 0) -> final destroy (0unexpected survivors). Closes 34 POA&M items with finding-ID-bound regression tests while trackingPOAM-2026-201(lula_audit->A.5.3observer-subject feedback spiral) openly. - Durable post-W detection & quarterly expiry (
#705,#707):CronJob/lula-auditemits structuredevaluated_count/satisfied_count/failed_controlstelemetry enforced against a>= 26active-manifest corpus floor; failing runs evaluate to blocking (False) onA.5.3, firenotifier.send_critical_alert(), and pair with a2 × scheduledead-man's switch (sla_monitor.py+google_monitoring_alert_policybound to@lahlfors), whilescripts/check_gate_w_freshness.pyfails scheduled CI once the 90-day certification window expires. - Regional SSP & Lula scope alignment (
#730): Reclassifies 5 institutional adopter obligations (eu-ai-act-art9,dora-art10,mas-feat,mas-notice-655,mas-trm-s6) tonot-applicable(control-origination: adopter-operated) with concrete 3-step adopter integration runbooks, splitsgdpr-art22into animplemented(HOLDS) runtime component (REQUIRE_APPROVAL, distinct-principal quorum,POST_HITLdrift revalidation) and aplannedcontroller contestation component (EU-006), keepsra-5plannedunderPOAM-010, and refreshes the 209-row OSCAL evidence-claim census. cage-clientv0.3.0(client-v0.3.0,#675,#676): First functional SDK release. Discloses three facts together: (1)client-v0.1.0andclient-v0.2.0never completed a call against a live gateway (defer_tokenvs.deferred_id/ envelope schema mismatch); (2) out-of-bandkidGovernanceEnvelope.signatureverification (ES256/EdDSAover RFC 8785 JCS,expires_atfreshness, andsubject.action_hashbinding) is now enforced by default and cannot be disabled in enforcing postures; and (3) previewGovernanceEnvelopedropped the unpopulatedsubject.record_hashandgovernance_context.controls_satisfiedfields (EP-1).
- Gate W empirical certification (
-
CI compares the repository's live protection with the documented configuration, and POAM-2026-116, POAM-2026-133 and POAM-2026-170 close on it (
ci(ci),check_repo_rules.py,repo-rules.yml,test_check_repo_rules.py,branch-protection-rules.md,CONTRIBUTING.md,POAM.md).check_repo_rules.pyreads the rules onmainand the release-tag ruleset from the GitHub API and fails on any difference from what the document says, in either direction: a rule, check, pin or ruleset missing or disabled, and also one added. A fetch error fails it too.repo-rules.ymlruns it on every pull request, every push tomainand daily, because what it checks can change without a commit.- The tests run seventeen planted drifts against the API's recorded responses and require each to be reported. They also hold the document's required-check table equal to the checker's pins, so the two cannot drift apart. Eight mutants of the checker each fail them.
- Run live against a branch no ruleset covers, the checker reported all four of
main's rules missing and exited 1. Bypass actors are visible only to admins, so a bypass actor added to a ruleset goes unseen. CONTRIBUTING.mdhad its own protection table and settings list: CODEOWNERS review, protectedrc-v*andrelease/**, seven checks it called required, three of which are not and one of which (Marker Contract Check) does not exist, tag patterns no rule uses, and read-and-write workflow permissions. It now summarises what is enforced and points to.github/branch-protection-rules.md, asJURISDICTION_INVENTORY.mddoes, so one list of required checks remains.
-
Pure-stdlib
cage-audit/3.0stream-invariant oracle (INV-1–INV-5), live Compose--streamCI gate, andGOVERNANCE_THRESHOLDS_LOADEDattestation (POAM-2026-151) (feat(governance),stream_invariants.py,check_trace_conformance.py,thresholds.py,ci.yml,docker-compose.yml).proof/stream_invariants.py(zerosrc/imports, pure-stdlib RFC 8785 JCS + SHA-256) verifiescage-audit/3.0chain integrity,GOVERNANCE_THRESHOLDS_LOADEDdigest agreement (POAM-2026-151Option 2 producer, together withthresholds_digestonGET /governance/policy-version),INV-1actuation seal provenance,INV-2single-use nonce consumption,INV-3refusal receipt integrity (proof_hash+payload_hash),INV-4barrier non-violation, andINV-5settlement ledger conservation.POST /tools/executesurfacesproof_hashon tool-caught governance refusals,docker-compose.ymlprovisions an internalredis:7.2-alpine(noeviction) service and defaults local dev signing toCAGE_KMS_PROVIDER=ed25519(SoftwareEd25519ProviderEdDSAJWT seals soCAGE_SEAL_STRICT_MODE=truestays enforced out of the box),QUICKSTART.mdadds an explicit dev-only Cloud KMS callout and an end-to-end/tools/executeseal mint→verify→consume + first-DENY step, andquickstart-smokerunsscripts/check_trace_conformance.py --streamover an actual live Docker Compose evidence stream.
-
Unfiltered
pull_requestworkflow triggers for stacked PRs (POAM-2026-153) and formal-verification / audit-posture disclosures (POAM-2026-154,POAM-2026-115,POAM-2026-120) (ci(ci),ci.yml,security-scan.yml,license_guard.yml,test_workflow_pr_triggers.py,FORMAL_VERIFICATION.md,POAM.md). Dropson.pull_request.branchesacross all nine pull-request workflows so stacked pull requests targeting feature base branches run all seven required status checks on push (WF-M1–WF-M4), adds the explicitDistributedCBF.tlaatomic-read model-vs-implementation verification-gap disclosure (POAM-2026-154), and records the 93/30-day review-bypass velocity analysis (POAM-2026-115) and test-fixture enforcement-masking pattern (POAM-2026-120). -
CI runs
terraform testagainst the gcp-gke target (ci,ci.yml,w6.tftest.hcl). The newterraform-testjob plans the staging profile with mock providers and module overrides, so it needs no credentials and no state. Its 10 runs check what each inference backend deploys (no vLLM, advisor or NeMo server undernone; vLLM and the selected advisor undervllm) and that the plan refusesexternal, anyenable_vllm,vllmwithout a GPU pool, and the NeMo server withoutvllm. The job is notcontinue-on-error, so it can fail (POAM-2026-150). Two mutations ofmain.tfwere checked to fail it. -
CAGE_INFERENCE_BACKENDchooses the inference backend independently of jurisdiction:none,externalorvllm(feat(gateway),inference_backend.py). Undernonethe gateway starts with noVLLM_*settings and builds no NeMo rails, and every model-dependent path refuses with its own reason code instead of admitting or reaching for a default URL.- Content safety answers
BLOCKEDwithCONTENT_SAFETY_NO_INFERENCE_BACKEND(mcp_tool_server.py,semantic_rail_factory.py). - The inference proxy answers 503 with
INFERENCE_BACKEND_NONE(inference_proxy.py). - Consensus at or above its threshold is a DENY with
CONSENSUS_NO_INFERENCE_BACKEND, reached before any critic is consulted. Below the threshold it skips, as it does with a backend (engine.py). - The compliance bridge records the remediation advisory as
NOT_PERFORMEDwithINFERENCE_BACKEND_NONE(audit_workflow.py).
An unknown value fails startup. Unset keeps the behaviour from before the setting existed: a backend is required. No infrastructure profile changes here; choosing
noneper profile is a separate change. - Content safety answers
-
A pre-merge double-check of every required status check (
feat(compliance),check_pr_mergeable.sh). The script reads the required checks GitHub applies to the base branch, verifies each on the exact head SHA through the REST check-runs and commit-status APIs, fails closed on an empty rule set or any API error, and posts the result on the pull request. -
AOF persistence with
append_fsync = ALWAYSon the governance Valkey instance in every enforcing posture (POAM-2026-156item 3) (feat(infra),memorystore_valkey,main.tf).- The module always renders
persistence_config(DISABLED, orAOFwith an explicitappend_fsync) from two validated inputs, and refuses AOF without a policy and a policy without AOF. - The gcp-gke target names its enforcing predicate once,
local.enforcing_posture(prod or staging, or any compliance flag), which the retention lock now reads too. The governance instance plansAOF+ALWAYSunder it and an explicitDISABLEDotherwise; the app instance plans an explicitDISABLED. - Applying it updates an existing instance in place: nothing under
persistence_configisForceNewin the google provider (checked atv6.43.0andv6.50.0), and Google notes the instance might need maintenance.governance_persistence.tftest.hclplans every committed profile. Not yet applied to staging;POAM-2026-156closes only on the measured latency delta. - The
enable_apac_mas_compliancedescription now states what the flag does: exactly what the other two compliance flags do (POAM-2026-130).
- The module always renders
Changed
- Evidence documents that asserted controls not in force are recorded as a class, POAM-2026-169, and five of its seven instances are corrected (
docs(compliance),POAM.md,branch-protection-rules.md,TECH_STACK.md,SECURITY.md).- The class: a statement that a control is in force, written from a file, module or intent and never checked. Its detection rule: such a statement names the observation that showed it, with a date, or it is not evidence.
.github/branch-protection-rules.mdnow describes the protection the GitHub API returns, with the commands that read it. It had listed six settings onmainthat no rule applied, arelease/**rule for branches that don't exist, read-and-write workflow permissions, and a release workflow that doesn't exist.TECH_STACK.mdno longer credits VPC Service Controls, Cloud DNS and Cloud Armor toperimeter.tf, which enables the first two in no shipped profile and attaches the third to nothing.- POAM-2026-115's merge counts were less than half the real ones: 233 of the 240 pull requests merged from 2026-09-30 to 2026-10-31 had no approving review, not 105. Its skew policy now records that Compliance Matrix has failed on
mainsince 2026-10-30, and that before #602 the same failure was reported as a pass. - Open: the OSCAL SC-7 statements, which cite network-policy manifests no deployment path applies, and Compliance Matrix, whose workflow never installs the
uvits tests need.
POAM-2026-115is an accepted risk, not an open finding, andPOAM-2026-120gains a detection rule (docs(compliance),POAM.md).POAM-2026-115moves to a new Accepted Risks section. No work closes it: it closes when a second maintainer exists. Its compensating controls are stated as verified on 2026-11-02: ruleset 25966171 has no bypass actor, so--adminskips only the review rule, andcheck_pr_mergeable.sh --commentposts a pre-merge record on the pull request. Nothing enforces the record: 55 of the 75 merges from #552 to #632 carry one posted before the merge, and none of the 13 merged on 2026-11-01 (UTC) does. The entry had said all eight required checks were pinned to the GitHub Actions app; three were (all eight since 2026-11-02, see Security).POAM-2026-120records how to spot its next instance: a result that agrees with itself is suspect on its face.
- Image digests and the project are rendered per deployment, never committed;
deploy_all.shbuilds, then renders, then plans (POAM-2026-131, re-rated High) (refactor(infra)!,variables.tf,main.tf,render_image_digests.sh,build_images.sh,deploy_all.sh,image_digests.tftest.hcl,test_image_digests_rendered.py).- No profile was deployable from the repository. Staging's committed digests named
gcr.io/cage-reference/, which refuses every pull; every other profile inheritedsha256:e3b0c442…, the SHA-256 of empty input. The renderer's output was never read, and the plan was saved before the images were built. var.image_digestsandvar.project_idhave no default, and no profile commits either. The render command is in each description (interactive prompt) and in a trailing comment on the declaration line, which Terraform quotes in its-input=falseerror.- The
image_digestsvalidation requires exactly the images the profile deploys (local.required_image_digests) and refuses the empty-input digest, so staging underinference_backend = "none"needs no vLLM, advisor or NeMo image. - The renderer writes a gitignored
image_digests.<sha>.tfvars.jsonwith no fallback to another build's image.deploy_all.shrenders before it plans and gains--image-sha.build_images.shselects images withIMAGES. - BREAKING — infrastructure: see RD-1–RD-4 in
BREAKING_CHANGES.md.
- No profile was deployable from the repository. Staging's committed digests named
- The GKE GPU node pool uses pool-wide
total_min_node_count/total_max_node_count(location_policy = "ANY"), starts at 0 GPUs in every profile, and allows only verified regional L4 +g2-standard-8zones (POAM-2026-132) (refactor(infra)!,main.tf,variables.tf,variables.tf,gpu_pool.tftest.hcl).google_container_node_pool.gpu_nodesreplaces per-zonemin_node_count/max_node_countwith pool-widetotal_min_node_count/total_max_node_countandlocation_policy = "ANY", andgpu_node_pool_initial_countdefaults to0(0per zone is0total). Multi-zone sourcing no longer multiplies initial, minimum, or maximum GPU counts bylength(gpu_node_locations).- Every
.tfvarsprofile (dev,us-dev,eu-dev,apac-dev,staging,prod,eu-prod,apac-prod) andterraform.auto.tfvars.examplesetsgpu_node_pool_initial_count = 0andgpu_node_pool_min_count = 0(max = 2on dev/staging,max = 5on*prod.tfvars) and explicitly lists only verified zones offering bothnvidia-l4andg2-standard-8:us-central1-a/b/c(restoring 3-zone capacity sourcing onstaging.tfvars),europe-west1-b/c(removingeurope-west1-d,POAM-2026-132, closed), andasia-southeast1-a/b/c. - Plan-time preconditions on
gpu_nodesrefuse empty/out-of-regiongpu_node_locationsand any unverified L4/g2-*zone in those three regions, exercised against the realgcp_gke_clustermodule bygpu_pool.tftest.hcl. - BREAKING — infrastructure:
gpu_node_pool_min_countandgpu_node_pool_max_countnow bound the pool total rather than each zone,gpu_node_pool_initial_countdefaults to0,*prod.tfvarsshipmin = 0, andeurope-west1-dfailsterraform planwith[POAM-2026-132]. See AG-8 inBREAKING_CHANGES.md.
- On GKE,
inference_backenddecides what is deployed: staging runs the governance core with no model by default, andexternalis compose-only (POAM-2026-136) (refactor(infra)!,main.tf,variables.tf,staging.tfvars).inference_backendreplacesenable_vllmand setsCAGE_INFERENCE_BACKENDon the gateway and the compliance bridge. Under"none", the default and staging's setting, no vLLM, no advisor in either runtime and no NeMo server are deployed, so nothing is left pointing at a model service that does not exist. Under"vllm"the vLLM deployments and the advisor are deployed."external"is refused: the cluster denies external egress by default, and opening it for the agent would weaken that control.- Staging keeps its GPU pool at min 0 / max 2 / initial 0, so switching to
"vllm"is a variable change and an apply (W6 narrowed staging tous-central1-awhile counts were per-zone; AG-8 moves the pool tototal_min_node_count/total_max_node_countand restoresus-central1-a/b/c). enable_nemo_guardrailsdefaults tofalse. A plan guard refuses"vllm"without a GPU pool, and the NeMo server without"vllm".- The never-read
vllm_fast_url,vllm_reasoning_urlandvllm_gateway_urlvariables and theirdeploy_all.shandload_env.shexports are deleted (POAM-2026-136, closed). - Lula SI-2 requires only the deployments every backend has, and digest-checks the optional ones when present. CM-6 no longer requires
nemo-guardrails-config. CM-6 still cannot pass, because nothing creates the thresholds ConfigMap it asserts (POAM-2026-151, opened). - The SSP no longer lists LLM provider APIs as an interconnection or an outbound inference destination on GKE. The gcp-gke README documents the backends, including that the advisor calls its models directly by design.
- BREAKING — infrastructure: remove
enable_vllmfrom every profile and setinference_backend. A deployment that relied on staging's vLLM, advisor or NeMo server setsinference_backend = "vllm". See AG-7 inBREAKING_CHANGES.md.
CAGE_NEMO_ENABLED=falsenow holds at runtime, not only at startup; all three opt-in switches share one parser; and "rails off" has its own reason code (POAM-2026-144) (fix(governance)!,semantic_rail_factory.py,feature_flags.py,inference_proxy.py,hybrid_server.py,mcp_tool_server.py,nemo_node_factory.py,content_safety.py).- Every path that builds NeMo rails consults
semantic_rails_refusal(): gateway startup, the inference proxy,get_nemo_rails(),reload_nemo_rails(), thecage:nemo:reloadsubscriber and refinement approval. Before this change the inference proxy rebuilt rails on every request while the switch was off, and a reload broadcast re-enabled them on all sub-apps. Both routed around thenemo_rails_backendstartup check. - With the rails off, the inference proxy runs deterministic-only: the Stage 1 keyword scan, then
scrub_pii()output masking, with ISO 42001A.6.1.2stampedSKIPPED. The reload subscriber is not started, and applying an approved refinement returns409with the proposal leftstaged. verify_content_safetyreturnsCONTENT_SAFETY_RAILS_DISABLEDwhen the rails are off by configuration.CONTENT_SAFETY_RAILS_UNAVAILABLEnow means only "enabled but not initialised".- BREAKING — configuration:
CAGE_NEMO_ENABLED,CAGE_CONSENSUS_CRITICS_ENABLEDandCAGE_CAUSAL_TIER_ENABLEDaccept onlytrue/false/1/0(empty or unset is off).yes/noused to raise for NeMo but enable critics and the causal tier; it now raisesConfigurationErrorfor all three.consensus_critics_enabledandCAGE_CONSENSUS_CRITICS_ENABLED_ENVmove fromsrc.gateway.governance.consensustosrc.gateway.governance.feature_flags. See AG-6 inBREAKING_CHANGES.md.
- Every path that builds NeMo rails consults
- NeMo Colang semantic rails are opt-in (
CAGE_NEMO_ENABLED=falseby default),validate_output_semantics()callsrails.llm.ainvoke()directly with strictSAFEparsing (POAM-2026-125), and deadCustomSelfCheckOutputActionis removed (POAM-2026-140) (refactor(nemo)!,nemo_node_factory.py,mcp_tool_server.py,hybrid_server.py,posture.py,manager.py,actions.py,action_registry.py,definitions.co,nemo-rails-configmap.yaml).nemo_rails_enabled()innemo_node_factory.pynow defaultsCAGE_NEMO_ENABLEDto"false";mcp_tool_server.pyandhybrid_server.pyskipinitialize_rails()whenCAGE_NEMO_ENABLED=false; andassert_production_posture()(nemo_rails_backendinposture.py) refuses startup whenCAGE_NEMO_ENABLED=trueandCAGE_INFERENCE_BACKEND=none.- BREAKING — behaviour: set
CAGE_NEMO_ENABLED=true(backed by a>= 7Binstruction-tuned judge model; seesrc/integrations/nemo/README.md) to enable Colang 2.x input flows and LLM output auditing;CustomSelfCheckOutputAction(custom_self_check_output) is removed. See AG-5 inBREAKING_CHANGES.md.
- The STPA compiler refuses uncompilable declared enforcement targets,
UCA-3/4/7/10/12compile into structured checks, and deterministic PII redaction and Aho-Corasick screening are unbundled fromget_nemo_rails()(POAM-2026-141,POAM-2026-142) (fix(governance)!,stpa_compiler.py,core_system.yaml,stpa_control_structure.yaml,trade_hazards.yaml,uca_rules.py,nemo_node_factory.py,NIST_SP800_53.json).UCAModel._validate_enforcement_deps()instpa_compiler.pynow raisesValueErrorat parse/compile time whenever a UCA declaresopaorpythonwithout a structured condition (ConditionModel.is_structured():composite, orparam+operatorwiththreshold/threshold_refwhere required) or declaresnemowithoutcondition.semantic_pattern.- In
core_system.yamlandstpa_control_structure.yaml,UCA-7drops its uncompilednemotarget (enforcement: [opa]). Intrade_hazards.yaml,UCA-3(output_contains_pii: is_true,enforcement: [python]),UCA-4(partial_transaction_detected: is_true,enforcement: [langgraph, python]),UCA-10(stale_balance_data_detected: is_true,enforcement: [python]), andUCA-12(bounding_contract_violation_detected: is_true,enforcement: [python, opa]) now carry structured conditions and compile intosrc/cage_finance/stpa/uca_rules.py.NIST_SP800_53.jsonreplaces citations to the deleted2d4aea66^:src/gateway/governance/generated_stpa_validator.pywithsrc/cage_finance/stpa/uca_rules.pyandsrc/gateway/governance/stpa_validator.py, and the three OSCAL artifacts undercompliance/oscal/are regenerated. nemo_node_factory.pyaddsnemo_rails_enabled()(CAGE_NEMO_ENABLED, raisingConfigurationErroron invalid values) and runs deterministic Presidio_redact_pii()and Stage 1ac_keyword_scan()on input, and Presidio_redact_pii()plus regexscrub_pii()on output, unconditionally before consultingget_nemo_rails().- BREAKING — STPA schema & validator: STPA YAML files declaring
opa/pythonwith onlysemantic_patternornemowithoutsemantic_patternnow raiseValueErrorat parse time, andGeneratedSTPAValidatorinsrc/cage_finance/stpa/uca_rules.pynow enforces_check_uca_3,_check_uca_4,_check_uca_10, and_check_uca_12. See AG-4 inBREAKING_CHANGES.md.
- Consensus separates deterministic notional threshold escalation from optional multi-model critic voting (
CAGE_CONSENSUS_CRITICS_ENABLED=falseby default), andproof/model.pymodels the 7-tier deterministic core (refactor(governance)!,engine.py,consensus_tier.py,opa_node_factory.py,posture.py,model.py). WhenCAGE_CONSENSUS_CRITICS_ENABLEDis unset orfalse,ConsensusGatedoes not instantiateConsensusModelRegistryorGatewayClient, and any action at or aboveconsensus_threshold(or inhigh_stakes_actions) deterministically returnsESCALATEwithreason_code: NOTIONAL_THRESHOLD_ESCALATION(ViolationKind.HITL→REQUIRE_APPROVAL) with zero LLM calls, including underCAGE_INFERENCE_BACKEND=none. SettingCAGE_CONSENSUS_CRITICS_ENABLED=trueactivates the 3-critic LLM quorum and requires an active inference backend (assert_production_posture()refusesCAGE_CONSENSUS_CRITICS_ENABLED=truewithCAGE_INFERENCE_BACKEND=none).proof/model.pyremoves"causal"fromTIERS(8 → 7deterministic tiers:ftra,stpa,opa,confidence,consensus,cbf,fiscal;34gated states,17ungated states,31skipped-tier states oncbf,38EU_ECBstates).- BREAKING — behaviour (staged correction to #566
493727f5): by default (CAGE_CONSENSUS_CRITICS_ENABLED=false), actions at or aboveconsensus_thresholdescalate toREQUIRE_APPROVAL(NOTIONAL_THRESHOLD_ESCALATION) instead of invoking LLM critics or failing closed withCONSENSUS_NO_INFERENCE_BACKENDunderCAGE_INFERENCE_BACKEND=none; setCAGE_CONSENSUS_CRITICS_ENABLED=trueto run multi-model critic voting.proof.model.TIERSnow has 7 entries instead of 8. See AG-2 and AG-3 inBREAKING_CHANGES.md.
- BREAKING — behaviour (staged correction to #566
- The finance causal tier is opt-in (
CAGE_CAUSAL_TIER_ENABLED=falseby default) and plugins declarerequired_tiersenforced at startup (refactor(governance)!,plugin.py,__init__.py,contracts.py,assembly.py,posture.py).FinanceCagePluginandcreate_finance_tiers()now assemble the five deterministic finance tiers (trade_confidence,bounding,cbf,fiscal,consensus) by default and recordDisabledTier("causal", "CAGE_CAUSAL_TIER_ENABLED=false")whenCAGE_CAUSAL_TIER_ENABLEDis unset orfalse. SettingCAGE_CAUSAL_TIER_ENABLED=true(or passinginclude_causal=True) registersCausalTierPluginandcausal_world_model_refresher; an unknown value raisesConfigurationError.PluginContribution.required_tiersdeclares the domain's mandatory tier names (FINANCE_REQUIRED_TIERSin finance, plus"causal"when enabled), andassert_production_posture()verifiesrequired_tiersbefore any other posture check, refusing startup in enforcing postures if any required tier is absent or disabled.- BREAKING — behaviour: the default finance governor assembles five domain tiers instead of six; set
CAGE_CAUSAL_TIER_ENABLED=true(or passinclude_causal=True) to include the DoWhy/EconML causal tier. See AG-1 inBREAKING_CHANGES.md.
- BREAKING — behaviour: the default finance governor assembles five domain tiers instead of six; set
- The inference proxy has no
localhostfallback (fix(gateway)!,inference_proxy.py). With no endpoint configured it forwarded prompts tohttp://localhost:8000/v1. It now answers 503 withINFERENCE_BACKEND_URL_MISSINGand rolls back the quota step.- BREAKING — behaviour: set
VLLM_BASE_URL(orVLLM_FAST_API_BASE/VLLM_REASONING_API_BASE) explicitly. The default was undocumented, and the MCP server already refused to start withoutVLLM_BASE_URL. See IB-1 inBREAKING_CHANGES.md.
- BREAKING — behaviour: set
- Consensus admission is decided once, in the kernel (
engine.py).consensus_admits()admits onlyAPPROVE,APPROVEDorSKIPPEDinstatus, and never a result that carries areason_code. The finance, healthcare and physical-AI tiers use it, and record the kernel's reason code over their own.- BREAKING — behaviour: every domain tier refuses a consensus result without
status, or one that carries areason_code; finance used to fall back todecisionwhenstatuswas absent. See IB-2 inBREAKING_CHANGES.md.
- BREAKING — behaviour: every domain tier refuses a consensus result without
- The compliance bridge records why there is no remediation advisory (
audit_workflow.py,main.py).advisor_statusisNOT_REQUIRED,NOT_PERFORMED,ERROR,PERFORMED,FAILEDorPENDING, withadvisor_reason_codewhen no advisory was attempted. A required backend with noVLLM_BASE_URLis now anERRORwithINFERENCE_BACKEND_URL_MISSING, not a log line. - Gate G9 now covers
docs/POAM.md, checks Terraform paths, and verifies removed paths pinned to a commit (check_doc_references.py). A document that has to name a path removed from HEAD pins it in git's revision syntax,<sha>^:<path>, and the gate checks the pin against history: the commit must be an ancestor of HEAD and must contain the path. It fails closed in a shallow clone, so thelintjob now checks out full history. The scope lives only in the script'sDOC_SCOPE; CI andmake docs-checkno longer carry their own lists. The label of a link with a local target is no longer checked as a second path, since the target is the reference. Three entries that onlydocs/POAM.mdrelied on left theKNOWN_DOCUMENTED_GAPSwaiver.AGENTS.mdrecords that whoever rewritesmain's history owns re-pinning.
Fixed
-
Every gcp-gke profile sets
enable_nist_complianceitself, so a directterraformrun withstaging.tfvarsplans the compliance posture (POAM-2026-174item 1) (fix(infra)!,staging.tfvars,dev.tfvars,staging_compliance.tftest.hcl,posture_guards.tftest.hcl,outputs.tf,outputs.tf,test_control_plane_access.py,BREAKING_CHANGES.mdCP-2 and CP-3,POAM.md).- The staging flag was commented out and only
deploy_all.shandstaging_lifecycle.shpassed it. A directterraform planorapply -var-file=staging.tfvarstherefore planned0.0.0.0/0on the control plane, skippedcontrol_plane_access_guardand thePOAM-019key guards, and turned off Cloud SQL PITR. - Breaking for direct runs (CP-2):
planandapplynow needauthorized_networks(at least one CIDR, no/0) and distinctlangfuse_compliance_*keys, asdeploy_all.shruns already did. Teardown does not:terraform destroy -var-file=staging.tfvarsreports these resource preconditions only as warnings, verified on Terraform 1.15.8 and 1.17.1. The same check written as a variable validation would block teardown, and a comment inmain.tfkeeps the guards as preconditions. dev.tfvarssetsfalse, which it always got, andtest_every_profile_sets_the_nist_flag_itselfrequires every committed profile to set the flag. The classifier intest_control_plane_access.pyno longer modelsdeploy_all.sh's default. Theus-dev.tfvarscomments that credited the flag with Redis replication and PDBs, and placed the key guard inapp_secrets/main.tf, are corrected.staging_compliance.tftest.hclplans the profile with no flag override and asserts that compliance is active, PITR is on, only the supplied CIDRs reach the control plane, and an empty or/0list stops the plan. A control run with the flag forced off shows the0.0.0.0/0fallback. Re-commenting the flag fails the file. Of the five existing test files the change touched, three supply the compliance key fixtures, so every run in them plans the new default; two force the flag off in four runs, each of which keeps the tier and flag it planned before.- Item 2 of
POAM-2026-174stays open:master_authorized_networks_configanddatabase_encryptionare in the cluster'signore_changes, so a cluster created open by an earlier direct apply is not corrected by this change. POAM-2026-179's remediation is amended to match. Its owner-format and TTL checks become resource preconditions, because as variable validations they would block teardown for anyone still holding a pre-convention list. Its plan-time HMAC options are marked as putting the helper on the teardown path.POAM-2026-180is opened: the fiveterraform_dataguards inmain.tfare standalone, so a plan targeted at what one protects skips it. Two of those gaps have no preventive backstop:-target=module.gkewith an empty list, and-target=module.app_secretswith shared compliance keys.POAM-2026-126,POAM-019and the docs that describe these guards now say they refuse untargeted plans only.- A comment below
control_plane_access_guardinmain.tfholds the placement matrix (targeted plan and destroy plan, for each of four places a check can go, all run on both versions), andAGENTS.mdadds the rule to its Terraform invariants: a check goes where a plan targeted at the resource it protects still evaluates it, as a precondition on that resource or a guard it depends on. The comment also records that a root-levelterraform testcannot expect a failure inside a child module, so a precondition there is tested from the module's own tests.test_guard_placement.pyfails on any new standalone guard ininfra/. Two more false comments are corrected:main.tfsaidmodule.app_secretsskips the compliance secret when the keys are empty, andus-dev.tfvarscredited the compliance bridge with detecting shared keys. - Breaking (CP-3): the deployment guards follow the tier, not the jurisdiction flags (
POAM-2026-181). The control-plane guard, bothPOAM-019key guards, the cluster allowlist, PITR and the NeMo securityContext followed the three jurisdiction flags, which all do the same thing. So a staging or prod plan that selected none planned an open control plane with no guard: the base platform with no overlay, orstaging.tfvarspointed at EU_ECB with the NIST flag off as its description instructed (verified on Terraform 1.15.8 and 1.17.1). All of them now bind onlocal.enforcing_posture, which was already true for every staging and prod plan, and the flags decide only whether a dev plan enforces.posture_guards.tftest.hclplans staging and prod with no flag in all three jurisdictions.test_control_plane_access.pyfails ifany_compliance_activegates anything else again. No committed profile is affected, and teardown is unchanged. The flag descriptions, the cluster module README,deploy_all.sh's DEP-01 comment and messages, the DEP-01 decision record and the staging row inDEPLOYMENT_RULES.mdnow describe the code: staging is US_FED, becausedeploy_all.shignores aCAGE_DEPLOYMENT_REGIONthat disagrees with the profile. OSCAL output is unaffected, because the exporter followsCAGE_DEPLOYMENT_REGION.POAM-2026-130records that nothing checks a flag against its jurisdiction, while the jurisdiction itself is checked against the region.
- The staging flag was commented out and only
-
tlc-model-checkruns on every pull request and every push tomain, and its runner fails on any TLC verdict the pins do not cover (POAM-2026-090,POAM-2026-091,POAM-2026-150) (ci(ci),tlc-model-check.yml,verify_tla.py,POAM.md,FORMAL_VERIFICATION.md,proof/README.md).- The workflow had only a
workflow_dispatchtrigger from 2026-10-02 and first ran on 2026-11-01 (run41362761903). That run reproduced every pinned count, and every negative control failed as pinned. Both closures had cited the workflow as if it ran. Each POAM row now records local verification on its original closure date and the first CI run; the closure dates are unchanged. verify_tla.pyparsed only invariant violations, so it never read TLC's verdict onSP4_FenceEpochMonotonic, an action property declared underPROPERTIES. With SP-4 planted false, TLC v1.7.4 printed 768Action propertyerrors and the runner printed OK on all 8 cfgs. Under-continueTLC exits 0 and printsNo error has been foundeven after violations. The runner now fails on any TLCError:line other than a pinned invariant violation or its trace header, and on a run that never reports completion. Against real TLC it reports BAD on the planted tree's threeDistributedCBFcfgs and OK on all 8 real cfgs.test_gate_tlc_model_check_rejects_unpinned_tlc_verdictsplants six defects without Java, and the runner from before the fix fails it.tests/test_workflow_pr_triggers.pyfails if either trigger is removed. POAM-2026-150 records the workflow as a sixth gate that could not fail.- The pins measured the state space, not what was checked over it: a cfg that dropped an
INVARIANTSorPROPERTIESline kept every count and passed.DECLARED_CHECKSintla_pins.pynow pins each of the 8 cfgs' invariants and properties. The runner andtest_tla_specs_proof.pyboth enforce it; the test also proves that dropping any one declared line is caught, and it checks the Step 15 table inFORMAL_VERIFICATION.md, including "all 11 / all 10 invariants", against the pins. proof/README.mddocuments the out-of-repo toolchain (Java and the sha256-pinned jar). Without it,make verify-tlaand pytest both stay green while TLC never runs, and CI's pytest-logic always runs that way. The gate test and the workflow note that its verdict strings are TLC v1.7.4's and must be re-derived fromMP.classwhen the jar is bumped.FORMAL_VERIFICATION.mdStep 12 cited12,184states for the N=2 suite. That is the pre-#351 model's N=4 pin. It now cites that run's per-cfg counts, 1,811 + 3,972 + 2,388, which sum to 8,171 over three state spaces. The status row no longer saystests/test_no_direct_bind_proof.pypins 38 gated states (it pins 34), and the Step 7 output block is marked as predating #574.
- The workflow had only a
-
No LLM client sits in the kernel, the base install or the gateway image (
POAM-2026-168) (fix(governance)!,consensus/engine.py,seams/critic.py,critic_backend_factory.py,inference_client/,check_model_free_install.py).- Before this fix,
assemble_governor()loaded the openai SDK for any plugin contributing consensus (finance does). It did so even underCAGE_INFERENCE_BACKEND=none, where consensus can only refuse. The base install resolved five model packages and the gateway image fourteen. - The consensus gate keeps its deterministic half in the kernel: trigger, context selection, prompt rendering, vote parsing and aggregation. The model call crosses the new
CriticBackendseam. - The kernel's
GatewayClientandConsensusModelRegistrymoved to Layer 3 asInferenceClientandRoutedCriticBackend. The backend is bound only when critics are enabled and a model exists. Without theinferenceextra, assembly fails closed withCONSENSUS_CRITIC_BACKEND_UNAVAILABLE. litellm,openaiand the coverage plugin left the base install, andnemoguardrailsleft thegatewayextra. Two new extras hold them:nemoandinference.- The gateway image no longer installs the
advisorextra or copies the advisor's source. Its comment claimed the gateway imported it; nothing does. - G3 now forbids LLM clients and model runtimes in
src/gateway/. It listed none before. - The new model-free install gate runs in the Lint job. It fails if the base install or the gateway image resolves a model package, or if G3 stops listing an LLM client.
- The hermetic job's "no ML" check covers LLM clients too.
- A subprocess test assembles a finance governor and finds no model client in
sys.modules. Its positive control seesopenaionce critics and a model are configured. - BREAKING — kernel, packaging & images: see MF-1–MF-5 in
BREAKING_CHANGES.md.
- Before this fix,
-
Consequence single-use markers are replica-acknowledged on a pinned socket,
provider_01keysrecper determination rather than per mandate, andcbf_engine.py's two vacuous-WAITdefects are closed (POAM-2026-156items 1 & 4,POAM-2026-157,POAM-2026-070,POAM-2026-090) (fix(governance)!,redis_replication.py,consequence_authority_store.py,consequence_gateway.py,cbf_engine.py,redis_single_use.py,provider.py,replicated_valkey.py,test_consequence_consumption_durability.py,test_provider_01_payload_contract.py,test_flowsignal_step2_qualification.py).ConsequenceAuthorityStore.consume_once()pins a single connection viaredis_replication.py(pinned_connection,write_socket,acknowledged_replicas,ReplicationPolicy.from_env()), re-asserts its own marker via_REASSERT_IF_OURS_LUAif a lost-reply retry inset_nx_outcome()took theGETbranch (NxOutcome.APPLIED_UNCERTAIN), and sendsWAITdirectly on the underlyingwrite_socketrather than throughredis-py's retryingexecute_command. Under strict replication (CAGE_STRICT_REPLICATION), an unacknowledged write raisesConsumptionNotReplicatedErrorwithout deleting the marker on the primary, andConsequenceGateway.evaluate()returnsConsequenceDecision.BLOCKwithreason_code="CONSUMPTION_NOT_REPLICATED".provider_01removes the test-onlyprefer_receipt_id_for_recpayload flag, unconditionally setsConsequenceToken.rectoauthority_receipt.id(orauthority_receipt_id), and fails closed (cage.authority_denial) when anALLOWresponse carries no receipt id.TC-01–TC-06intest_flowsignal_step2_qualification.pynow build their payload throughFriaTier._payload(), andtest_provider_01_payload_contract.pyguards against top-level payload reads outsideFriaTier._payload().- Extracting
redis_replication.pyalso remediates two distinct defects incbf_engine.pythat left its priorWAITvacuous againstDistributedCBF.tla'sSyncReplication: (1)rollback_state()passed the unpinned pool client (self.redis_client) into_sync_to_replicas()soLUA_ROLLBACK_CBFandWAITcould run on different pooled connections on every call with no disconnect involved, and (2)atomic_verify_and_commit()/_sync_to_replicas()calledtarget_client.execute_command("WAIT", ...)through the retrying client so a socket drop during primary failover silently reconnected on a fresh socket with zero writes and returned>= 1vacuously. Nine mutations (M1–M9) are verified inPOAM.md. - BREAKING — governance & provider_01: see CD-1–CD-3 in
BREAKING_CHANGES.md.
-
Declared consequence magnitudes (
ConsequenceMagnitudes) are enforced atFriaTierpre-flight,ConsequenceTokenminting, andConsequenceGatewayconsumption (POAM-2026-155) (fix(governance)!,contracts.py,consequence_magnitude.py,fria_tier.py,consequence_token_service.py,consequence_gateway.py,assembly.py,provider.py,adapter.py,src/cage_finance/,src/cage_healthcare/).- Domain plugins declare per-action
ConsequenceMagnitudetuples overMagnitudePath(keys, minor_unit_digits)(paths=()when explicitly magnitude-free).ConsequenceMagnitudes.violations()refuses undeclared actions (MAGNITUDE_UNDECLARED), absent paths underparams(MAGNITUDE_ABSENT), and values that arebool, non-numeric, non-finite, sub-minor-unit dust,<= 0, or> 2^53 - 1minor units (MAGNITUDE_INVALID). - Enforced before
provider.validate_fria()inFriaTier.evaluate()(FRIA_MAGNITUDE_INVALID), before signing inmint_consequence_token_finding()(CONSEQUENCE_TOKEN_MINT_FAILED), and at Step 5 ofConsequenceGateway.evaluate()beforestore.consume_once()(leaving the single-use authority record unconsumed on refusal).provider_01drops its name-based"amount"/"magnitude"float checks. cage_financedeclares all 6 registered actions (execute_trade,execute_trade_boundedatminor_unit_digits=2; 4 read-only actions withpaths=());cage_healthcaredeclares its 5 dosing actions atminor_unit_digits=6(whole nanograms) and leavesadjust_dosageandoverride_contraindicationundeclared (MAGNITUDE_UNDECLAREDunderEU_ECB);cage_physical_airemains unrunnable (domain_config = None,POAM-2026-077) withconsequence_magnitudes=().- BREAKING — governance & integrations:
magnitudesis a required keyword-only argument onmint_consequence_token_finding()andConsequenceGateway(),minor_unit_digitsis required onMagnitudePath, and undeclared actions fail closed withMAGNITUDE_UNDECLARED. See CM-1–CM-4 inBREAKING_CHANGES.md.
- Domain plugins declare per-action
-
The GKE load test restores the gateway's replica count when its stdout closes or its terminal hangs up (
fix(infra),run_gke_load_test.sh,test_gke_load_test_script.py).- The script pins the gateway to one replica and restores the original count from its EXIT trap. It trapped INT and TERM but not SIGPIPE. With stdout closed (a
| teereader gone, a dropped session), its next write raised SIGPIPE, which killed bash before the EXIT trap could run, so the gateway stayed at one replica. Against the test stubs this happened every time stdout closed with no signal (returncode −13, no restore call). It also happened when SIGTERM arrived after the script's next write, which is howtest_closed_stdout_pipe_still_restores_the_original_countfailed once under parallel load. - SIGPIPE and SIGHUP are now trapped like INT and TERM, and all four exit through the restore. Bash already ran the EXIT trap for an untrapped SIGHUP, but only through its own signal handling.
- New tests close stdout with no signal (exit 141) and send SIGHUP (exit 129). The close-then-SIGTERM test accepts either signal handling first (141 or 143); the restore is asserted in every case.
- The script pins the gateway to one replica and restores the original count from its EXIT trap. It trapped INT and TERM but not SIGPIPE. With stdout closed (a
-
The Langfuse event bucket shares the WORM bucket's retention posture, so a staging walk locks it for one day instead of seven years (
POAM-2026-159) (fix(infra)!,main.tf,outputs.tf,retention_posture.tftest.hcl,test_clickhouse_operator.py,DEPLOYMENT_RULES.md).local.retention_period_seconds(86400 s forstaging, 220752000 s for every other tier, thedevdefault included) andlocal.retention_locked(always forstagingandprod, fordevonly under a compliance flag) feed bothgoogle_storage_bucket.langfuse_event_uploadandmodule.worm_bucket. The Langfuse bucket locked 2555 days whenever a compliance flag was set, whichdeploy_all.shdoes for US_FED staging, so every staging walk left a bucket, and a lien on its project, that nobody could remove for seven years.- GCS refuses to shorten a locked policy, so the bucket moves to
google_storage_bucket.langfuse_event_upload(langfuse-event-upload-<project>-<suffix>), and aremovedblock (destroy = false) makes the first apply forgetgoogle_storage_bucket.langfuse_eventswithout deleting it. On the live staging project the old bucket and its lien stay until at least 2033-09-27.DEPLOYMENT_RULES.mdgives the migration, why it must be applied before any destroy, and what still stops a full staging teardown: signing keys withprevent_destroy, buckets withoutforce_destroy, and the project lien. - Twelve plan-only runs on mock providers read every committed profile from its file: staging takes one day on both buckets, every prod profile keeps 2555 days, and an unknown tier is refused. Two static checks pin that both buckets read the shared locals and that an unset tier defaults to
dev. Fourteen mutations of the gate (M1–M14 in the pull request) each fail a named test. - The IAM comment no longer calls the bucket a transient queue.
- BREAKING — infrastructure: see LR-1 and LR-2 in
BREAKING_CHANGES.md.
-
The POAM-019 plan guard is split into one resource per condition, and each error headlines the condition it detected (
fix(infra),main.tf,poam_019.tftest.hcl,DEPLOYMENT_RULES.md,QUICK_START.md).- With a compliance posture active,
terraform_data.poam_019_empty_key_guardrefuses empty compliance telemetry keys ([POAM-019] Compliance telemetry keys are empty under compliance hardening.) andterraform_data.poam_019_shared_key_guardrefuses keys equal to the application project's ([POAM-019] Compliance telemetry keys equal the application project keys.). The single guard they replace headlined both as an AU-9 isolation breach, so unset variables ("" == "") read as a compliance defect when.envhad not been loaded. An empty key is no longer also reported as shared, the likely cause moves into the remediation text, and both messages citedocs/compliance/us_fed/POAM_US_FED.mdinstead of the nonexistentdocs/POAM_US_FED.md#POAM-019. - The POAM-2026-086 Memorystore CA precondition that shared the old resource moves unchanged to
terraform_data.memorystore_ca_pinning_guard. Existing state plans one destroy (the old resource) and three creates; none of them manages infrastructure. - Seven plan-only runs on mock providers test each guard on its own, since
expect_failuresnames a resource, not a precondition. - Troubleshooting entries carried over from the closed PR #591: the Linkerd network-validator crash loop (root cause unestablished), the unverified
sa-cage-deployerpath, the POAM-2026-126 authorized-networks lockout, the POAM-019 diagnostics, and diagnostic hygiene.
- With a compliance posture active,
-
The partner over-the-wire harness can fail, and runs every partner with only its own secrets (
POAM-2026-150) (ci(ci),ci-partner-integration.yml,partner_harness.py,require_live.py,selection_flags.py,test_partner_harness.py). It had never run, and as written no run could have tested a partner.--run-partner-integrationnow enables partner tests. Every partner live test also carrieslive_external, whichconftest.pyskipped unless--run-live-externalwas set, so the harness,make test-partnerand the runbook's commands all skipped every partner test.- One job per partner (
provider_01,02,03,05,07,08,actuator_01;07and08were missing), at the paths #218 moved the tests to. Each secret reaches only its own partner's test step; before, the secrets sat on a step that ran no tests. - The
require_liveplugin fails a job on any skip, and the verdict job fails unless every partner job passed. Before, an all-skipped run and a skipped job both reported success. - actuator_01 receives the file paths its adapter reads: its PEM secrets are written to files at run time. The workflow used to pass differently named variables that the adapter never read.
- Runs weekly as well as on dispatch.
test_partner_harness.pychecks the workflow against the live tests on every pull request.
-
Live
quickstart-smokeexercises Theorem 5.3 routing-seal nonce replay (INV-2) and HTTPtransaction_idreplay against a running Redis container, andPOAM-2026-115/ branch-protection docs match live ruleset25966171(eight required status checks) (ci(ci),ci.yml,test_quickstart_bootstrap_contract.py,test_workflow_pr_triggers.py,branch-protection-rules.md,POAM.md).quickstart-smokenow replays (1) the already-executedAAPLtransaction_idoverPOST /tools/execute(ExecutionLedgerRedisSET NXemitsGOVERNANCE_REFUSAL_RECEIPTand refuses a second actuation) and (2) a liveEdDSAJWT routing seal throughverify_and_consume_seal()twice inside the runninggatewaycontainer againstredis:7.2-alpine(SymbolicGovernorViolation: replay attack detected — seal already consumed), assertingACTUATION_RECEIPT == 1,GOVERNANCE_SEAL_CONSUMED == 2, and--expected-refusal-count 2viacheck_trace_conformance.py --stream.POAM.md(POAM-2026-115,POAM-2026-153),.github/branch-protection-rules.md, andtest_workflow_pr_triggers.pynow document all eight required status checks in ruleset25966171(includingQuickstart Bootstrap & First-DENY Smoke Gate,integration_id: 15368).
-
Concurrent CBF commits no longer false-regress the fence epoch (
POAM-2026-154), andmake lintrunsmypy src/(POAM-2026-153) (cbf_engine.py,test_cbf_settlement_ledger.py,Makefile).DistributedCBF.tlamodelsReadas an atomic observation ofavailable_balance,fence_epoch, andagent_epochs[a], whereascbf_engine.pysplit it across two multi-await windows:_get_fence_epoch()+_check_fence_epoch()and a second post-read HWM check inatomic_verify_and_commit(). When a concurrent commit landed inside either window, the losing coroutine loggedCRITICALCBF_EPOCH_REGRESSION_DETECTEDorFENCE_EPOCH_REGRESSIONand aborted onattempt=1instead of retrying — both flaking thePOAM-2026-070double-spend guard (test_concurrent_trades_over_the_limit_admit_one[one-cbf]) and rejecting concurrent admissible trades._read_fence_epoch_and_hwm()now reads(safety:fence_epoch, safety:fence_epoch_hwm)in oneMGET(and in one transaction pipeline in self-reported mode),_read_cbf_state_atomic()brackets the read withseen_before/epoch_beforeto re-read on forward concurrent commits while failing closed immediately on genuine stale-replica regressions, andatomic_verify_and_commit()reuses the pre-read verified epoch and atomically captured HWM. Deterministic skew tests (skew_ticks = 0..13) and mutationsCBF-M1–CBF-M4guard every path.Makefile'slinttarget now runsuv run mypy src/, closing the local-gate omission exposed when stacked pull requests #578–#582 skippedmain-filtered CI workflows (POAM-2026-153, opened for the workflow branch filters).
-
The quickstart and README describe the core as merged (
QUICKSTART.md,README.md,proof/README.md).- The quickstart expected
/healthto report"nemo":"active". NeMo rails have been opt-in since W5, so under the compose defaults it reports"off".test_quickstart_bootstrap_contract.pynow compares the documented response with the gateway's, so the two cannot drift apart again unnoticed. - The README's pipeline ended
→ consensus → causal. It now shows the deterministic core in the formal model's order, ending→ consensus → CBF → fiscal, and says that consensus escalates to a human, critics are optional and the causal tier is an optional extension. proof/README.mdrecords that the downscoping plan predicted 34 / 19 / 32 states with FULL / POST_HITL / DRY_RUN labels, and that the measured 34 / 17 / 31 / 38 are the values the tests pin.
- The quickstart expected
-
The chaos-agent playground no longer presents emulations as real components, and its trades can be allowed (
chaos_agent_playground.py,examples/README.md).- Only Tier 0, the finance plugin's
GeneratedSTPAValidator, is real. Yet Tier 2 printed a Python re-implementation of the STPA Rego rules under "OPA Rego Policy", and the README said every layer "executes its actual logic". Every tier now says whether it is real, emulated or simulated, and the README and QUICKSTART say the playground never calls the governor, OPA, consensus, CBF or a model. - Every trade payload lacked
order_sizeanddaily_vol, so UCA-6 refused every trade whatever the other inputs. The payloads now carry them. A new Scenario F, a compliant baseline, shows the real validator allowing the action it denies in Scenario A, and the summary checks each scenario against its expected verdict.
- Only Tier 0, the finance plugin's
-
deploy_all.shtakes the posture from the profile, not from the--envname (deploy_all.sh; ADR-005 inDEPLOYMENT_DECISION_RECORD.md). The banner andenable_nist_compliancefollowed the--envname andCAGE_DEPLOYMENT_REGION, while terraform applied the profile'senvironmentandcage_deployment_region.--env stagingor--env prodwithoutCAGE_DEPLOYMENT_REGIONskipped the NIST hardening their US_FED profiles declare, and--env eu-prodprinted "DEV MODE". ACAGE_DEPLOYMENT_REGIONthat disagrees with the profile is now ignored with a warning. Two error messages no longer tell operators to put Langfuse compliance keys in a committed tfvars file. -
Twelve references in
docs/POAM.mdand one in the PIA pointed at nothing, or at the wrong file (POAM.md,PRIVACY_IMPACT_ASSESSMENT.md). Seven name files whose deletion was the remediation or that were removed later; they are pinned to the commit before the deletion rather than repointed. One of them is POAM-2026-032's scaffold, which a documentation sweep on 2026-09-16 (a2a338b8) had repointed to a live production module that the entry does not describe. Three cited files never existed: POAM-2026-069's and POAM-2026-070's test module and POAM-2026-086's Terraform module. One test had moved (#218), and one verification step is reworded. The PIA cited an IAM file that has not existed at any commit; it now citesinfra/targets/gcp-gke/iam.tf. -
The compliance artifacts no longer credit NeMo with controls the deterministic path provides (
POAM-2026-145) (fix(compliance),oscal_ssp_exporter.py,NIST_SP800_53.json,aarm_mapper.py,compliance/oscal/). NeMo rails have been opt-in since AG-5 (#576), but the artifacts still described them as the enforcement.- The SSP listed the NeMo Guardrails Server as
operationalfor ISO 42001 A.9.2 and NIST PT-2/PT-3. Those claims now belong to a new Deterministic PII Redaction component: in-process Presidio (12 identity entity types, not 16) and regexscrub_pii. Wherever its statements claim fail-closed behaviour they carry the input-side fail-open recorded below as POAM-2026-147, and they stay that way while that item is open. The NeMo component isother, an optional tier. The system description, data flow, boundary and PII information type say the same. - The STPA narratives cited
config/rails/generated_stpa_rails.co, which defines no rails, as enforcing nine SP 800-53 controls (eleven mentions in the SSP with the artifact list and theartifact-nemoprop). The exporter now names only the generators a control's UCAs target, so the rails file, the prop and "NeMo Colang" in the compiler component appear only while some UCA targetsnemo. The narratives also stop citing OPA for controls whose UCAs target only Python. - The FTRA SI-10 statements credited "NeMo/AGP rails generated from UCA-7". UCA-7 compiles only to OPA; the always-on injection screen is the Stage 1 keyword scan. The SP 800-53 component that claimed SI-10 and SI-3 through NeMo, at a path that does not exist, is now the deterministic screening component, and SI-3 is
partialbecause nothing deterministic screens model output for harmful content. AARM-V2, V5 and V10 cite the deterministic code, and V10 no longer claims 20 entity types. - POAM-2026-094 and POAM-2026-114 are re-scoped to defects of the optional causal tier (
CAGE_CAUSAL_TIER_ENABLED, off since #573). Three findings from the review are recorded as open: A.9.2's Lula evidence has no producer (POAM-2026-146), input redaction fails open when Presidio cannot initialise (POAM-2026-147), and the component definition describes a TypeScript gateway that never existed in this repository (POAM-2026-148).
- The SSP listed the NeMo Guardrails Server as
-
Test sessions no longer read
.env, and a hermetic session refuses to start while one can reach it (POAM-2026-120) (fix(tests),dotenv_guard.py,conftest.py,test_dotenv_guard.py,ci.yml).tests/conftest.pyloaded the repository's.envin three places, one of them a bareload_dotenv()that also searched every parent directory.git cleankeeps the file, so one left in a reused worktree changed every later run: on 2026-11-01, one written by a quickstart run turned 87 tests red. The same file could as easily have turned failing tests green.- The loaders are gone, but litellm, which
srcimports, still callsload_dotenv()at import, and the virtualenv's place inside the repository puts the repository's.envon its search path. So the session checks instead of trusting: if a.envsits in the test directory, the working directory, site-packages or any of their parents, a hermetic session stops with a usage error that names the file. - Live runs (
--run-integration,--run-live-external,--run-partner-integration,--run-e2e) are exempt. They take their values from the shell, or explicitly fromuv run --env-file .env. CI's quickstart job runs its pytest step before it creates.envfor Docker Compose. - Contributors: a hermetic run beside the quickstart's
.envnow stops, where QUICKSTART §6 previously only asked you to move the file aside.
Security
-
maincan no longer be force-pushed or deleted, all eight required checks are pinned to GitHub Actions, and GitHub Actions can no longer approve pull requests (POAM-2026-170, POAM-2026-115) (repository settings;POAM.md,branch-protection-rules.md,SECURITY.md).- No rule on
mainblocked a force push or deletion: classic protection was off, and its rulesets held only the review rule, which admins bypass, and the required checks. The activity log records two force pushes tomain, both on 2026-07-09. Ruleset 26377312 now appliesnon_fast_forwardanddeletiontomainwith no bypass actor. It was seen refusing a force push and a deletion on a disposable branch before it was retargeted tomain. - Five of the eight required checks in ruleset 25966171 accepted a result of the same name from any source. All eight are now pinned to the GitHub Actions app, the only source any of them had used.
- A workflow's token could approve a pull request and so meet the review rule without anyone reviewing. That setting is off; no workflow used it.
- No rule on
-
Mesh identity is accepted only from the pod's own Linkerd proxy, and governance pods without the proxy are refused at admission (POAM-2026-167, High) (
fix(security)!,workload_identity.py,auth.py,proxy-required-policy.yaml,test_workload_identity_middleware.py,test_mesh_proxy_required_policy.py).- The gateway and the compliance bridge took the caller's identity from
l5d-client-idwithout checking where it came from. Through the proxy the header is sound: on staging, the proxy replaced a forged value with the caller's certificate identity and removed an unmeshed caller's. A pod without its proxy took whatever a caller sent. That included the approver identities on the gateway's approval routes and the operator identity on the bridge's dual-control path. - Pods without the proxy do get created. The injector's webhook fails open, and on 2026-10-29 a gateway replica was admitted without the proxy while the injector was stuck.
- The header now counts only on a connection from inside the pod (
is_pod_local_connection): loopback, or the pod's own IP, which is where the proxy connects from. From any other peer, the gateway refuses the request (INGRESS_IDENTITY_NOT_FROM_PROXY) and the bridge returns 401. Both, and the standalone MCP server, run uvicorn with proxy headers off, soX-Forwarded-Forcannot set the peer. - The
service_meshmodule'sproxy_required_namespacesinstalls thecage-mesh-proxy-requiredValidatingAdmissionPolicy. A pod created in a listed namespace without thelinkerd-proxycontainer is refused, with no opt-out, and its controller retries until the injector is back. The gcp-gke target lists its governance namespace. Tested live in a staging probe namespace. - BREAKING: a request that sets
l5d-client-iditself, from outside the pod, is refused. Onlykubectl port-forward, which connects from inside the pod, still works. Docker Compose has no proxy, socompose_gateway_call.shsends identified requests from inside the gateway container; the quickstart and its CI gate use it. On gcp-gke, a governance-stack pod can no longer opt out of the proxy.
- The gateway and the compliance bridge took the caller's identity from
-
The pip-audit gate's four suppressions are removed, since none of them matched a locked package version, and an
--ignore-vulnwithout a comment naming its ID now fails the tests (POAM-2026-161) (ci(security),security-scan.yml,test_workflow_vuln_suppressions.py).PYSEC-2026-139andGHSA-rrmf-rvhw-rf47covered torch, which leftuv.lockin #224; #90 had deleted the second one's comment and kept the flag.PYSEC-2026-3721is fixed in pip 26.2, the version the lock pins and the audit venv installs.GHSA-rg7c-g689-fr3xwas commented as the advisory for diskcache, which is not in the lock either, but the ID is google-adk's CVE-2026-4810, and the locked google-adk 2.3.0 is not affected. POAM-017 inPOAM_US_FED.md, which still said CVE-2026-4810 was ignored in CI through that ID, now records the removal.- The "Scan compliance extras" step audited the same venv as the gate, which already holds every extra including
compliance, and ended in|| true, so it could not fail. It is deleted. - Under the runner's
bash -e, the gate step exited before printing its JSON whenever pip-audit failed. It now prints the results, then fails. - The test parses every workflow. A flag passes only if a shell comment in its own
run:block names the exact ID; flags outside arun:block, non-literal IDs and the abbreviations argparse accepts in a pip-audit command, such as--ig=, are reported.
-
The gateway's egress allowlist no longer names model providers (POAM-2026-152) (
variables.tf,egress-lockdown.yaml).api.openai.com,api.anthropic.comandgenerativelanguage.googleapis.comare removed from the Terraform default and from the static manifest the SSP cites as SC-7 evidence. No GKE inference path called them: inference is in-cluster vLLM or none, and theexternalbackend is compose-only. A test keeps the two lists equal and free of providers.- Still open: the
*.googleapis.comwildcard the gateway needs for GCP APIs also matches Google's model API. Narrowing it needs an inventory of the GCP APIs the gateway calls. - The compliance documents that still described third-party providers as part of the system are corrected: the SP 800-53 profile's PT-2 guidance (which also repeated "16 entity types" and credited NeMo; it is 12, deterministic Presidio), the catalog's SA-9 and the SSP outline.
- POAM-2026-149 records the staging WORM check: none of the 30 archived actuation receipts holds a seal, because the evidence sink's sanitizer redacted each one as
[REDACTED_JWS]. That prevented the leak (no irremovable credential) and permanently broke decision linkage for those 30 receipts; restoring linkage must not weaken the sanitizer's coverage. - Healthcare no longer declares a causal graph that nothing read (POAM-2026-143, closed).
-
langgraph patched for GHSA-3fx2-cqw9-xr3c (Dependabot #54–#56, moderate) (
uv.lock,packages/cage-client/uv.lock). langgraph 1.2.7 → 1.2.14 in the root lock; langgraph 1.2.11 → 1.2.14 and langgraph-sdk 0.4.4 → 0.4.6 in the cage-client lock. Both are runtime dependencies. Nothing insrc/orpackages/cage-client/importslanggraph_sdkorRemoteGraph, so the vulnerable streaming client was not reached; patched because a fix exists. -
CI jobs that cannot fail are recorded as a pattern (POAM-2026-150) (
POAM.md).ci.yml'sintegration-smokejob runspytest -m integrationwithout--run-integration, so it skips all 153 tests it selects and passes. It has never run an integration test since it was added on 2026-08-11. It is the fourth job found this week whose test outcomes cannot fail the build, afterregional-posture.yml, the test step ofcompliance-matrix.yml(both POAM-2026-121) andlula-validate.yml. In its run on 2026-10-29,lula-validate.ymlsatisfied none of the 33 manifests and succeeded. Every Lula result committed to the repository is a schema error. The entry records the pattern and changes no workflow. While it is open, none of these jobs' status is cited as evidence that something was tested. -
Every writer of a decision digest validates it, and each digest field has one meaning (POAM-2026-149, reopened) (
fix(security)!,digests.py,quarantine.py,verdicts.py,quarantine_actuator.py,ocsf_ingestor.py,stream_invariants.py). The 2026-10-29 fix below covered one of the three writers ofgovernance_decision_digest.QuarantineDirectivehad no format check and two meanings: the kernel's quarantine set it to a refusal receipt'sproof_hash, and actuator_03 to a clearance's seal digest. actuator_02's OCSF model accepted any string of 8 to 128 characters. Both wrote the field into the WORM-archived evidence stream, and the oracle checked it only on actuation receipts.- One validator,
require_sha256_hex()in a new seam module that imports only the standard library, accepts only lowercase SHA-256 hex matched in full, and its error never echoes the value.ExecutionClearance,QuarantineDirectiveandOcsfEventModeluse it; the OCSF model also checks digest keys nested inmetadataand hides inputs in its error strings. governance_decision_digestnow means onlyseal_digest(seal). A refusal-triggered quarantine names its refusal in a new field,refusal_proof_hash. A directive carries exactly one of the two,to_dict()emits only that one, andWORKLOAD_QUARANTINE_RECEIPTrecords it under its own name. Should a refusal hash ever be malformed,handle_deny()refuses the quarantine and still raises the DENY.- The stream oracle gains
digest_format: both fields, at any depth of every record of every event type, must be lowercase 64-hex, and a finding never echoes the value. A refusal-triggered quarantine receipt must follow its refusal (refusal_evidence), and_is_hex64now matches in full. The packaged copy is byte-identical, andcage-conformance0.3.0 lists the new rule. - The fixtures built around the defect (
"digest-abc-123","digest-1") now carry real digests.test_decision_digest_format.pycovers every writer, the oracle and the packaged copy; in a recorded mutation check, each ofDD-M1–DD-M26fails a named test. POAM-2026-149 satisfies both technical closure criteria (DD-M1–DD-M26andcage-conformance0.3.0) and awaits only the merge SHA; over-the-wire partner sandbox verification of actuator_03's changed signed directive (partner notice) is tracked separately inPOAM-2026-166. - BREAKING — contract and wire:
QuarantineDirectiveis keyword-only and takes exactly one ofgovernance_decision_digestandrefusal_proof_hash; a kernel quarantine's receipt and actuator_03's signed body carryrefusal_proof_hash; actuator_02 refuses a non-digest; stream verification reportsdigest_format. See DV-1–DV-4 inBREAKING_CHANGES.md.
- One validator,
-
Actuation evidence records the routing seal's digest, never the seal (POAM-2026-149) (
fix(security)!,actuation.py,routing_seal.py,broker_actuator.py, finance and healthcaretool_provider.py). Finance from #190 and healthcare from #534 passed the raw routing seal asExecutionClearance.governance_decision_digest, and from #327dispatch_actuation()wrote that field into every actuation receipt on the evidence stream. The stream's durable copy is WORM-retained, so a seal written there can never be redacted. Each seal had been consumed before its receipt was written and expired after the seal TTL (30 s by default), so the current verifier refuses a replay; the problem is that the archived copy is permanent.- Both domains now record
seal_digest(seal), the SHA-256 the decision trace already uses asseal_ref, so a seal's issuance and its actuation join on one key.ExecutionClearancerefuses any digest that is not lowercase SHA-256 hex, so a seal in that field fails at construction in any domain. The finance broker takes the seal it re-checks from the out-of-bandrouting_seal, not the digest. test_seal_digest_evidence.pymints real HMAC and JWT seals and follows both domains throughdispatch_actuation()to the evidence record; in a recorded mutation check, each ofSD-M1–SD-M13fails a named test.- BREAKING — contract: an
ExecutionClearancewhosegovernance_decision_digestis not a lowercase SHA-256 hex digest raisesValueError, and the broker's refusal for a missing seal isMISSING_ROUTING_SEAL(wasMISSING_GOVERNANCE_DIGEST). See RS-1 inBREAKING_CHANGES.md.
- Both domains now record
-
A consensus critic with no endpoint no longer falls back to
api.openai.com(POAM-2026-127) (fix(gateway)!,llm.py). With onlyVLLM_BASE_URLandVLLM_API_KEYset, the documented minimum forCAGE_INFERENCE_BACKEND=external, a critic without a dedicated URL built an OpenAI client with no base URL. That client targets OpenAI's public API with the operator's key and the trade prompt. Critics now use the routed endpoint, elseVLLM_BASE_URL, else refuse, and a refusal is an ERROR vote.- BREAKING — behaviour: with no endpoint configured, critics vote ERROR and consensus escalates instead of calling OpenAI. See IB-4 in
BREAKING_CHANGES.md.
- BREAKING — behaviour: with no endpoint configured, critics vote ERROR and consensus escalates instead of calling OpenAI. See IB-4 in
-
Staging's GKE control plane no longer accepts every address (POAM-2026-126) (
fix(infra)!,main.tf,staging.tfvars).staging.tfvarsauthorized0.0.0.0/0under a header claiming VPN scope, and nothing read the list. Whenever a compliance posture is active, the plan now refuses an empty allowlist or a /0. The livecage-stagingallowlist was narrowed the same day.- BREAKING — deploy contract: supply
authorized_networks(TF_VAR_authorized_networksor a--var-fileafter the profile) when deploying staging or any profile with a compliance flag on. See CP-1 inBREAKING_CHANGES.md.
- BREAKING — deploy contract: supply
-
A content-safety check that reaches no verdict is now a refusal (POAM-2026-124) (
fix(gateway),manager.py,mcp_tool_server.py,api.py). With no NeMo rails, the gateway'sverify_content_safetyanswered an error string that the advisor labelledSAFE. When NeMo raised "No main flow found", three rail functions returned safe. Each such path now returnsBLOCKEDwithCONTENT_SAFETY_RAILS_UNAVAILABLE,CONTENT_SAFETY_NO_MAIN_FLOWorCONTENT_SAFETY_ERROR(content_safety.py), keeping the degraded span marking, and the advisor records only an exactSAFEas safe.test_content_safety_no_verdict.pypins each path; in a recorded mutation check, each test fails when its path is reverted.- BREAKING — output: absent rails answer
BLOCKED: CONTENT_SAFETY_RAILS_UNAVAILABLE: …instead ofERROR: NeMo rails not initialised., a check that raises refuses instead of failing the tool call, and refusal reasons lead with a reason code. See IB-3 inBREAKING_CHANGES.md.
- BREAKING — output: absent rails answer
-
Output semantic validation calls
rails.llm.ainvoke()directly and accepts only an exactSAFEverdict, and deadCustomSelfCheckOutputActionis removed (POAM-2026-125,POAM-2026-140) (refactor(nemo)!,manager.py,actions.py,test_validate_output_semantics.py).validate_output_semantics()previously calledrails.generate_async()with a system message, routing through the Colang input flow without ever sending the auditor prompt (POAM-2026-125). It now rejects emptyoutput_text, callsrails.llm.ainvoke()directly, strips closed<think>...</think>blocks, accepts only exact ASCII"SAFE", extracts the reason on"UNSAFE", and fails closed withCONTENT_SAFETY_ERRORon all other replies. DeadCustomSelfCheckOutputAction(custom_self_check_outputinconfig/rails/actions.py), which no Colang flow invoked and whose Stage 2safe_indicatorsshort-circuited on 22 common financial words, is removed (POAM-2026-140). Guarded bytest_validate_output_semantics.py(OUT-M1–OUT-M6mutation table verified inPOAM.md). -
POAM-2026-070 was closed while its control did not operate; a regression test now guards it, and POAM-2026-119 and POAM-2026-120 are opened (
POAM.md,test_cbf_settlement_ledger.py). POAM-2026-070, an externally reported concurrent double-spend within the reconciliation window, was closed on 2026-08-28 citing a function and a test module that have never existed. The debit tracking #102 added was not atomic. On Redis 7.4.12, two trades each admissible alone but not together both committed when run concurrently, at21ad845aand at46447c2f, the parent of #256; run one after the other, the second was refused. From #256 (2026-09-26) the fence-epoch compare-and-set refused the second concurrent commit; since #506 (2026-10-22) it retries, and the in-script debit total refuses it. The entry now records both dates: the original closure (struck, incorrect), the control operating from 2026-09-26, and verification on 2026-10-27.test_concurrent_trades_over_the_limit_admit_oneruns the two trades concurrently on one CBF instance and on two, with every Redis call yielding. In a sensitivity check it fails when the debit write is moved out of the Lua script, as #102 had it, even with today's fence-epoch check in place, and passes when only the fence-epoch check is removed.- POAM-2026-023 now records that its remediation removed the in-script balance read that had serialized concurrent commits. POAM-2026-069's test pointer and refusal message are corrected. The POA&M/Lula drift checker had counted all three as covered by Lula files that match their control IDs and never mention the CBF; it now counts them by their tests.
- POAM-2026-119: of 13 identifiers cited as commits, six are pre-squash branch commits, each mapped to the squash commit that carries it, and seven resolve nowhere, including the only commit POAM-2026-011's closure cites. None is an advisory-fork commit kept in the maintainer's private refs.
- POAM-2026-120: closed controls are not guarded by tests that fail when the control stops operating. 34 of the 80 closed entries cite no test, and the drift checker's 52 "covered" entries rest on an unchecked file-name match. The remediation is a forward closure requirement and a backfill of High entries only.
-
Required status checks on
maincan no longer be bypassed, and release tags can no longer be moved or deleted (repository rulesets; POAM-2026-115 and POAM-2026-116 inPOAM.md). The checks moved out of the review ruleset into ruleset 25966171, which has no bypass actor. Before that, the maintainer's admin merges bypassed both rules, and 163 of the 215 pushes tomainfrom 2026-09-27 to 2026-10-27 went in with required checks not passing. The one exception is an advisory fix landed from a private fork, where Actions does not run: the ruleset is disabled for that merge and re-enabled, asSECURITY.mddescribes. Tag ruleset 25965917 protects everyv*andclient-v*tag, includingv4.0.0, with no bypass actor; no tag was protected before. POAM-2026-115 now covers only AC-5, since one maintainer cannot approve their own pull requests..github/branch-protection-rules.mdno longer claims that code-owner review is enabled, and its tag section describes the new ruleset. -
The causal risk boundary can fail open as volatility rises; POAM-2026-117 records it and POAM-2026-094 may not close first (
POAM.md; documentation only). The kernel predicts risk as0.5 + β × amount / normalization_scale. That prediction has no volatility term and uses a 0.5 that is neither configured nor estimated. In the frame POAM-2026-117 states, the boundary permits $25,504.81 at volatility 0.9, while a volatility-aware reading of the same fit trips at $9,619.17. The defect is latent while POAM-2026-094 leaves every causal verdict untrusted in enforcing postures. POAM-2026-094 is now amended to say it must not close before POAM-2026-117, and its scope now covers the kernel'sMockTelemetryProvider, which generates the same calibration formula as the finance fixture. POAM-2026-118 records thatstagingandlocalcan select the mock telemetry provider: onlyprodandproductionare refused, and only the mock's missing timestamp column stops it today. Closed POAM-2026-057, which records the formula as implemented, now cross-references POAM-2026-117. -
§3 of the Regional Posture Validation Report makes a false composition claim, a plugin overlay can redefine a kernel control, and content-safety checks that did not run are reported as passed; POAM-2026-121 to POAM-2026-124 record them (
POAM.md,REGIONAL_POSTURE_VALIDATION_REPORT.md; documentation only).- POAM-2026-121 (High): §3's independence claim was false when written, because
cage_finance's US_FED overlay redefines the kernel'sCTRL_FTRA_001. Its 12 hashes reproduce exactly on3b69207c, the parent of the commit that published them, and match no tree that contains the report. §3 and the report's head now carry notes saying not to rely on it. The claim is withdrawn rather than proven: POAM-2026-123 shows independence violated in a path that changes a tier's verdict. - POAM-2026-122 (Low): region-pinned tests number 57 for
eu_ecb, 23 forus_fedand 13 forapac_mas, although US_FED carries the NIST SP 800-53 surface. - POAM-2026-123 (High): the control registry applies plugin overlays over kernel controls and only logs the collision. A test overlay that defines
CTRL_FRIA_006moved the EU_ECB FRIA tier past its Art. 27(2) staleness gate. With the kernel baseline alone, that gate denies without consulting the provider. - POAM-2026-124 (High): with no rails, or when NeMo has no main flow,
verify_content_safetyand the advisor report a check that did not run as passed. The control itself operates with the shipped Colang 2.x rails. The fix lands with the inference-backend kernel change, beforeCAGE_INFERENCE_BACKEND=noneexists. - POAM-2026-120 gains an instance in the consensus path. Finance's
HIGH_VALUE_TRADEhigh-stakes override is asserted present by a test but has never fired, so consensus is triggered by amount alone.
- POAM-2026-121 (High): §3's independence claim was false when written, because
-
filelock4.2.4,langgraph-api0.16.9 andlanggraph-cli0.4.40 clear four dependency advisories (pyproject.toml,uv.lock).filelock3.20.3 had two: CVE-2026-112909 (GHSA-9vhv-jxww-xfm5) and CVE-2026-112910 (GHSA-567q-5qvw-pf92), fixed in 4.2.1;[tool.uv].override-dependenciesforcesfilelock>=4.2.1pastlitellm's<4.0pin.langgraph-api0.13.5 had CVE-2026-118668 (fixed in 0.15.0) andlanggraph-cli0.4.31 had CVE-2026-115885 (fixed in 0.4.32), pulled in by thedevgroup (langgraph-cli[inmem]).
-
authlib1.9.1 andmsgpack1.2.4 clear eight dependency advisories (uv.lockonly). They were published aftermain's last clean dependency scan, and from then on the scan failed on every pull request.authlib1.7.2 had seven: GHSA-6v38-4h7w-rj9h, GHSA-fch4-3c6x-j4gc, GHSA-7vr2-2fgp-9w7p, GHSA-c5mf-xmp2-59rc, GHSA-x4rv-c3wx-4cr8, GHSA-jj7h-c3hg-m23x and GHSA-8m4g-j4wp-wgcg. Onlygoogle-adkpulls it in, under theadvisorextra.msgpack1.2.1 had one: GHSA-j586-36cw-2gc2. Onlypip-auditpulls it in, for development.
Nothing in
src/imports either package, and no other locked package changed.
Full Changelog: v4.0.0...v4.1.0