Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions .github/workflows/ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -326,6 +326,12 @@ jobs:
# consumers must be emitted by a producer in src/ (zero baseline/waivers).
run: |
uv run python scripts/check_producer_consumer_keys.py --verbose
- name: Mechanism reference & TTM signing-hold gate (POAM-2026-222)
# Verifies every cage_* Prometheus series in alert policies/dashboards,
# every compliance-bridge field read by Lula specs, every metric/alert
# citation across OSCAL and DOC_SCOPE, and POAM-2026-222 TTM signing hold.
run: |
uv run python scripts/check_mechanism_references.py --verbose
- name: OSCAL evidence-claim census & register parity check
# Verifies 1-to-1 claim completeness and SHA-256/status freshness across
# all 7 OSCAL YAML files, path resolution in cites/read_back.check, and
Expand Down
2 changes: 2 additions & 0 deletions Makefile
Original file line number Diff line number Diff line change
Expand Up @@ -142,6 +142,8 @@ security:
docs-check:
@echo "==> Checking documentation references (broken links, dangling paths, stale symbols)..."
@uv run python scripts/check_doc_references.py
@echo "==> Checking monitoring, Lula, OSCAL, and TTM mechanism references..."
@uv run python scripts/check_mechanism_references.py --verbose
@echo "==> Checking vendor brand leakage into executable code..."
@uv run python scripts/check_vendor_brands.py --verbose
@echo "==> Checking domain literal leakage into kernel/integrations..."
Expand Down
2 changes: 1 addition & 1 deletion docs/architecture/EVIDENCE_CHAIN.md
Original file line number Diff line number Diff line change
Expand Up @@ -306,7 +306,7 @@ out-of-cluster GCP Cloud Audit Log corroboration (`AS-1` decisions `D-AS-1`..`D-
cluster teardown.

**Operational Metrics vs. Evidentiary Non-Repudiation (`R-9` Comparison-Owed Note)**:
GKE Managed Prometheus metrics (`cage_evidence_*`, `cage_custody_batches_total`,
GKE Managed Prometheus metrics (`cage_evidence_*`, `cage_evidence_custody_batches_total`,
`cage_kms_rpc_total`) provide operational pipeline health and alerting (NIST
SP 800-53 AU-12) and are intentionally non-evidentiary. Cryptographic
non-repudiation and out-of-cluster tamper detection rest exclusively on
Expand Down
2 changes: 1 addition & 1 deletion docs/architecture/GATEWAY_ARCHITECTURE.md
Original file line number Diff line number Diff line change
Expand Up @@ -487,7 +487,7 @@ While **Langfuse** serves as the regional operational LLM trace store, **Prometh
| **GCP KMS Provider** ([`integrations/gcp/kms_provider.py`](../../src/integrations/gcp/kms_provider.py)) | `cage_kms_rpc_total{op}` | Cloud KMS RPC call counter partitioned by operation (`sign`, `get_crypto_key_version`, `get_public_key`, `list_crypto_key_versions`) |
| **CBF Safety Engine & Reconciliation** ([`cbf_engine.py`](../../src/gateway/governance/safety/cbf_engine.py)) | `cage_reconciliation_replay_rejected_total{source}`, `cage_cbf_epoch_regression_detected_total`, `cage_cbf_current_fence_epoch`, `cage_cbf_wait_latency_seconds`, `cage_cbf_wait_timeout_total`, `cage_cbf_strict_replication_rollback_total` | R-04 replay rejections, R-05 fence-epoch double-spend defense, and Redis `WAIT` synchronous replication latency/timeout/rollback tracking |
| **Evidence Stream Producer** ([`evidence/stream.py`](../../src/gateway/governance/evidence/stream.py)) | `cage_evidence_commit_total{status}`, `cage_evidence_commit_duration_seconds`, `cage_evidence_append_conflicts_total`, `cage_evidence_blocking_disabled{env}`, `cage_evidence_stream_disabled{env}` | Redis Stream commit throughput/latency, Lua compare-and-append contention, and fail-closed posture gauges |
| **FTRA Boundary Gate** ([`governor/metrics.py`](../../src/gateway/governance/governor/metrics.py), [`ftra/node_factory.py`](../../src/gateway/governance/ftra/node_factory.py)) | `cage_ftra_boundary_checks_total{result}`, `ftra_llm_parse_failures_total` | Pre-pipeline reachability outcomes (`passed`, `conditional_clear`, `hitl_required`, `error`) and LLM structured-output parse failures |
| **FTRA Boundary Gate** ([`governor/metrics.py`](../../src/gateway/governance/governor/metrics.py), [`ftra/node_factory.py`](../../src/gateway/governance/ftra/node_factory.py)) | `cage_ftra_boundary_checks_total{result}`, `cage_ftra_parse_failures_total{failure_class}` | Pre-pipeline reachability outcomes (`passed`, `conditional_clear`, `hitl_required`, `error`) and structured-output parse failures |
| **Gateway Shared GCRA Rate Limiter** ([`rate_limit.py`](../../src/gateway/server/rate_limit.py), [`governance_middleware.py`](../../src/gateway/server/governance_middleware.py), [`mcp_tool_server.py`](../../src/gateway/server/mcp_tool_server.py)) | `cage_rate_limit_hits_total{route,scope}`, `cage_rate_limit_backend_errors_total{route}`, `mcp_rate_limit_active_buckets` | Shared Redis dual-bucket GCRA (`identity` + `principal`) rate-limit rejections, Redis fallback error counter, and fallback bucket gauge |

Downstream evidence custody and cold-store verification metrics are documented in [`EVIDENCE_CHAIN.md`](EVIDENCE_CHAIN.md) (§3–§4.2).
Expand Down
Loading
Loading