feat: Enhance Authority Identity Provider Registry with Bootstrap Capability
- Added support for bootstrap providers in AuthorityIdentityProviderRegistry. - Introduced a new property for bootstrap providers and updated AggregateCapabilities. - Updated relevant methods to handle bootstrap capabilities during provider registration. feat: Introduce Sealed Mode Status in OpenIddict Handlers - Added SealedModeStatusProperty to AuthorityOpenIddictConstants. - Enhanced ValidateClientCredentialsHandler, ValidatePasswordGrantHandler, and ValidateRefreshTokenGrantHandler to validate sealed mode evidence. - Implemented logic to handle airgap seal confirmation requirements. feat: Update Program Configuration for Sealed Mode - Registered IAuthoritySealedModeEvidenceValidator in Program.cs. - Added logging for bootstrap capabilities in identity provider plugins. - Implemented checks for bootstrap support in API endpoints. chore: Update Tasks and Documentation - Marked AUTH-MTLS-11-002 as DONE in TASKS.md. - Updated documentation to reflect changes in sealed mode and bootstrap capabilities. fix: Improve CLI Command Handlers Output - Enhanced output formatting for command responses and prompts in CommandHandlers.cs. feat: Extend Advisory AI Models - Added Response property to AdvisoryPipelineOutputModel for better output handling. fix: Adjust Concelier Web Service Authentication - Improved JWT token handling in Concelier Web Service to ensure proper token extraction and logging. test: Enhance Web Service Endpoints Tests - Added detailed logging for authentication failures in WebServiceEndpointsTests. - Enabled PII logging for better debugging of authentication issues. feat: Introduce Air-Gap Configuration Options - Added AuthorityAirGapOptions and AuthoritySealedModeOptions to StellaOpsAuthorityOptions. - Implemented validation logic for air-gap configurations to ensure proper setup.
This commit is contained in:
master
@@ -373,6 +373,7 @@ Authority now understands two flavours of sender-constrained OAuth clients:
|
||||
- Configure under `security.senderConstraints.dpop`. `allowedAlgorithms`, `proofLifetime`, and `replayWindow` are enforced at validation time.
|
||||
- `security.senderConstraints.dpop.allowTemporaryBypass` toggles an emergency-only bypass for sealed drills. When set to `true`, Authority logs `authority.dpop.proof.bypass`, tags `authority.dpop_result=bypass`, and issues tokens without a DPoP `cnf` claim so downstream servers know sender constraints are disabled. **Reset to `false` immediately after the exercise.**
|
||||
- `security.senderConstraints.dpop.nonce.enabled` enables nonce challenges for high-value audiences (`requiredAudiences`, normalised to case-insensitive strings). When a nonce is required but missing or expired, `/token` replies with `WWW-Authenticate: DPoP error="use_dpop_nonce"` (and, when available, a fresh `DPoP-Nonce` header). Clients must retry with the issued nonce embedded in the proof.
|
||||
- Refresh-token requests honour the original sender constraint (DPoP or mTLS). `/token` revalidates the proof/certificate, enforces the recorded thumbprint/JKT, and reuses that metadata so the new access/refresh tokens remain bound to the same key.
|
||||
- `security.senderConstraints.dpop.nonce.store` selects `memory` (default) or `redis`. When `redis` is configured, set `security.senderConstraints.dpop.nonce.redisConnectionString` so replicas share nonce issuance and high-value clients avoid replay gaps during failover.
|
||||
- Telemetry: every nonce challenge increments `authority_dpop_nonce_miss_total{reason=...}` while mTLS mismatches increment `authority_mtls_mismatch_total{reason=...}`.
|
||||
- Example (enabling Redis-backed nonces; adjust audiences per deployment):
|
||||
|
||||
@@ -13,7 +13,7 @@ completely isolated network:
|
||||
| Component | Contents |
|
||||
|-----------|----------|
|
||||
| **Merged vulnerability feeds** | OSV, GHSA plus optional NVD 2.0, CNNVD, CNVD, ENISA, JVN and BDU |
|
||||
| **Container images** | `stella-ops`, *Zastava* sidecar (x86‑64 & arm64) |
|
||||
| **Container images** | `stella-ops`, *Zastava* sidecar, `advisory-ai-web`, and `advisory-ai-worker` (x86‑64 & arm64) |
|
||||
| **Provenance** | Cosign signature, SPDX 2.3 SBOM, in‑toto SLSA attestation |
|
||||
| **Attested manifest** | `offline-manifest.json` + detached JWS covering bundle metadata, signed during export. |
|
||||
| **Delta patches** | Daily diff bundles keep size \< 350 MB |
|
||||
@@ -24,6 +24,8 @@ completely isolated network:
|
||||
**RU BDU note:** ship the official Russian Trusted Root/Sub CA bundle (`certificates/russian_trusted_bundle.pem`) inside the kit so `concelier:httpClients:source.bdu:trustedRootPaths` can resolve it when the service runs in an air‑gapped network. Drop the most recent `vulxml.zip` alongside the kit if operators need a cold-start cache.
|
||||
|
||||
**Language analyzers:** the kit now carries the restart-only Node.js, Go, .NET, Python, and Rust plug-ins (`plugins/scanner/analyzers/lang/StellaOps.Scanner.Analyzers.Lang.Node/`, `...Lang.Go/`, `...Lang.DotNet/`, `...Lang.Python/`, `...Lang.Rust/`). Drop the directories alongside Worker binaries so the unified plug-in catalog can load them without outbound fetches.
|
||||
|
||||
**Advisory AI volume primer:** ship a tarball containing empty `queue/`, `plans/`, and `outputs/` directories plus their ownership metadata. During import, extract it onto the RWX volume used by `advisory-ai-web` and `advisory-ai-worker` so pods start with the expected directory tree even on air-gapped nodes.
|
||||
|
||||
*Scanner core:* C# 12 on **.NET {{ dotnet }}**.
|
||||
*Imports are idempotent and atomic — no service downtime.*
|
||||
|
||||
@@ -35,7 +35,7 @@ _Source:_ `docs/assets/authority/authority-plugin-component.mmd`
|
||||
Capability flags let the host reason about what your plug-in supports:
|
||||
|
||||
- Declare capabilities in your descriptor using the string constants from `AuthorityPluginCapabilities` (`password`, `mfa`, `clientProvisioning`, `bootstrap`). The configuration loader now validates these tokens and rejects unknown values at startup.
|
||||
- `AuthorityIdentityProviderCapabilities.FromCapabilities` projects those strings into strongly typed booleans (`SupportsPassword`, etc.). Authority Core will use these flags when wiring flows such as the password grant. Built-in plugins (e.g., Standard) will fail fast or force-enable required capabilities if the descriptor is misconfigured, so keep manifests accurate.
|
||||
- `AuthorityIdentityProviderCapabilities.FromCapabilities` projects those strings into strongly typed booleans (`SupportsPassword`, `SupportsMfa`, `SupportsClientProvisioning`, `SupportsBootstrap`). Authority Core uses these flags when wiring flows such as the password grant, bootstrap APIs, and client provisioning. Built-in plugins (e.g., Standard) will fail fast or force-enable required capabilities if the descriptor is misconfigured, so keep manifests accurate.
|
||||
- Typical configuration (`etc/authority.plugins/standard.yaml`):
|
||||
```yaml
|
||||
plugins:
|
||||
|
||||
@@ -20,14 +20,16 @@ Focus: Identity & Signing focus on Authority (phase II).
|
||||
| # | Task ID & handle | State | Key dependency / next step | Owners |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| 1 | AUTH-DPOP-11-001 | DONE (2025-11-08) | DPoP validation now runs for every `/token` grant, interactive tokens inherit `cnf.jkt`/sender claims, and docs/tests document the expanded coverage. (Deps: AUTH-AOC-19-002.) | Authority Core & Security Guild (src/Authority/StellaOps.Authority/TASKS.md) |
|
||||
| 2 | AUTH-MTLS-11-002 | DOING (2025-11-07) | Deliver mTLS-bound token issuance/validation (cert thumbprint storage, JWKS rotation hooks) required for high-assurance tenants and plugin mitigations. (Deps: AUTH-DPOP-11-001.) | Authority Core & Security Guild (src/Authority/StellaOps.Authority/TASKS.md) |
|
||||
| 3 | PLG4-6.CAPABILITIES | BLOCKED (2025-10-12) | Finalise capability metadata exposure, config validation, and developer guide updates; remaining action is Docs polish/diagram export. | BE-Auth Plugin, Docs Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 2 | AUTH-MTLS-11-002 | DONE (2025-11-08) | Refresh grants now enforce the original client certificate, tokens persist `x5t#S256`/hex metadata via shared helper, and docs/JWKS guidance call out the mTLS binding expectations. (Deps: AUTH-DPOP-11-001.) | Authority Core & Security Guild (src/Authority/StellaOps.Authority/TASKS.md) |
|
||||
| 3 | PLG4-6.CAPABILITIES | DONE (2025-11-08) | Finalise capability metadata exposure, config validation, and developer guide updates; remaining action is Docs polish/diagram export. | BE-Auth Plugin, Docs Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 4 | PLG6.DIAGRAM | DONE (2025-11-03) | Component + sequence diagrams rendered (Mermaid + SVG) and offline assets published under `docs/assets/authority`; dev guide now references final exports. | Docs Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 5 | PLG7.RFC | DONE (2025-11-03) | LDAP plugin RFC reviewed; guild sign-off captured and follow-up implementation issues filed per review notes. | BE-Auth Plugin, Security Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 6 | SEC2.PLG | BLOCKED (2025-10-21) | Emit audit events from password verification outcomes and persist via `IAuthorityLoginAttemptStore`. ⛔ Waiting on AUTH-DPOP-11-001 / AUTH-MTLS-11-002 to stabilise Authority auth surfaces (PLUGIN-DI-08-001 closed 2025-10-21; re-run once sender constraints land). | Security Guild, Storage Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 7 | SEC3.PLG | BLOCKED (2025-10-21) | Ensure lockout responses and rate-limit metadata flow through plugin logs/events (include retry-after). ⛔ Pending AUTH-DPOP-11-001 / AUTH-MTLS-11-002; PLUGIN-DI-08-001 already merged, so limiter telemetry just awaits final Authority surface. | Security Guild, BE-Auth Plugin (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
| 8 | SEC5.PLG | BLOCKED (2025-10-21) | Address plugin-specific mitigations (bootstrap user handling, password policy docs) in threat model backlog. ⛔ Final documentation now hinges on AUTH-DPOP-11-001 / AUTH-MTLS-11-002 (PLUGIN-DI-08-001 landed 2025-10-21). | Security Guild (src/Authority/StellaOps.Authority/StellaOps.Authority.Plugin.Standard/TASKS.md) |
|
||||
|
||||
- 2025-11-08: PLG4-6.CAPABILITIES marked DONE – bootstrap capability surfaced in code/docs, registry logs updated, and bootstrap APIs now gate on providers that advertise it (`dotnet test` across plugins + Authority core).
|
||||
|
||||
## 100.D) __Libraries
|
||||
Dependency: None specified; follow module prerequisites.
|
||||
Focus: Identity & Signing focus on __Libraries.
|
||||
|
||||
@@ -9,6 +9,7 @@ Active items only. Completed/historic work now resides in docs/implplan/archived
|
||||
- 2025-11-03: AIAI-31-002 landed the configurable HTTP client + DI defaults; retriever now resolves data via `/v1/sbom/context`, retaining a null fallback until SBOM service ships.
|
||||
- 2025-11-03: Follow-up: SBOM guild to deliver base URL/API key and run an Advisory AI smoke retrieval once SBOM-AIAI-31-001 endpoints are live.
|
||||
- 2025-11-08: AIAI-31-009 marked DONE – injection harness + dual golden prompts + plan-cache determinism tests landed; perf memo added to Advisory AI architecture, `dotnet test src/AdvisoryAI/__Tests/StellaOps.AdvisoryAI.Tests/StellaOps.AdvisoryAI.Tests.csproj --no-build` green.
|
||||
- 2025-11-08: AIAI-31-008 moved to DOING – starting on-prem inference packaging, remote inference toggle, Helm/Compose manifests, scaling guidance, and Offline Kit doc refresh.
|
||||
- **Concelier** – CONCELIER-CORE-AOC-19-004 is the only in-flight Concelier item; air-gap, console, attestation, and Link-Not-Merge tasks remain TODO, and several connector upgrades still carry overdue October due dates.
|
||||
- **Excititor** – Excititor WebService, console, policy, and observability tracks are all TODO and hinge on Link-Not-Merge schema delivery plus trust-provenance connectors (SUSE/Ubuntu) progressing in section 110.C.
|
||||
- **Mirror** – Mirror Creator track (MIRROR-CRT-56-001 through MIRROR-CRT-58-002) has not started; DSSE signing, OCI bundle, and scheduling integrations depend on the deterministic bundle assembler landing first.
|
||||
|
||||
@@ -26,6 +26,11 @@ Advisory AI is the retrieval-augmented assistant that synthesizes advisory and V
|
||||
- Redaction policies validated against security/LLM guardrail tests.
|
||||
- Guardrail behaviour, blocked phrases, and operational alerts are detailed in `/docs/security/assistant-guardrails.md`.
|
||||
|
||||
## Deployment & configuration
|
||||
- **Containers:** `advisory-ai-web` fronts the API/cache while `advisory-ai-worker` drains the queue and executes prompts. Both containers mount a shared RWX volume providing `/var/lib/advisory-ai/{queue,plans,outputs}`.
|
||||
- **Remote inference toggle:** Set `ADVISORYAI__AdvisoryAI__Inference__Mode=Remote` to send sanitized prompts to an external inference tier. Provide `ADVISORYAI__AdvisoryAI__Inference__Remote__BaseAddress` (and optional `...ApiKey`) to complete the circuit; failures fall back to the sanitized prompt and surface `inference.fallback_*` metadata.
|
||||
- **Helm/Compose:** Bundled manifests wire the SBOM base address, queue/plan/output directories, and inference options via the `AdvisoryAI` configuration section. Helm expects a PVC named `stellaops-advisory-ai-data`. Compose creates named volumes so the worker and web instances share deterministic state.
|
||||
|
||||
## CLI usage
|
||||
- `stella advise run <summary|conflict|remediation> --advisory-key <id> [--artifact-id id] [--artifact-purl purl] [--policy-version v] [--profile profile] [--section name] [--force-refresh] [--timeout seconds]`
|
||||
- Requests an advisory plan from the web service, enqueues execution, then polls for the generated output (default wait 120 s, single check if `--timeout 0`).
|
||||
|
||||
@@ -145,3 +145,10 @@ All endpoints accept `profile` parameter (default `fips-local`) and return `outp
|
||||
- **Plan determinism:** `AdvisoryPipelineOrchestratorTests` shuffle structured/vector/SBOM inputs and assert cache keys + metadata remain stable, proving that seeded plan caches stay deterministic even when retrievers emit out-of-order results.
|
||||
- **Execution telemetry:** `AdvisoryPipelineExecutorTests` exercise partial citation coverage (target ≥0.5 when only half the structured chunks are cited) so `advisory_ai_citation_coverage_ratio` reflects real guardrail quality.
|
||||
- **Plan cache stability:** `AdvisoryPlanCacheTests` now seed the in-memory cache with a fake time provider to confirm TTL refresh when plans are replaced, guaranteeing reproducible eviction under air-gapped runs.
|
||||
|
||||
## 13) Deployment profiles, scaling, and remote inference
|
||||
|
||||
- **Local inference containers.** `advisory-ai-web` exposes the API/plan cache endpoints while `advisory-ai-worker` drains the queue and executes prompts. Both containers mount the same RWX volume that hosts three deterministic paths: `/var/lib/advisory-ai/queue`, `/var/lib/advisory-ai/plans`, `/var/lib/advisory-ai/outputs`. Compose bundles create named volumes (`advisory-ai-{queue,plans,outputs}`) and the Helm chart mounts the `stellaops-advisory-ai-data` PVC so web + worker remain in lockstep.
|
||||
- **Remote inference toggle.** Set `AdvisoryAI:Inference:Mode` (env: `ADVISORYAI__AdvisoryAI__Inference__Mode`) to `Remote` when you want prompts to be executed by an external inference tier. Provide `AdvisoryAI:Inference:Remote:BaseAddress` and, optionally, `...:ApiKey`. When remote calls fail the executor falls back to the sanitized prompt and sets `inference.fallback_*` metadata so CLI/Console surface a warning.
|
||||
- **Scalability.** Start with 1 web replica + 1 worker for up to ~10 requests/minute. For higher throughput, scale `advisory-ai-worker` horizontally; each worker is CPU-bound (2 vCPU / 4 GiB RAM recommended) while the web front end is I/O-bound (1 vCPU / 1 GiB). Because the queue/plan/output stores are content-addressed files, ensure the shared volume delivers ≥500 IOPS and <5 ms latency; otherwise queue depth will lag.
|
||||
- **Offline & air-gapped stance.** The Compose/Helm manifests avoid external network calls by default and the Offline Kit now publishes the `advisory-ai-web` and `advisory-ai-worker` images alongside their SBOMs/provenance. Operators can rehydrate the RWX volume from the kit to pre-prime cache directories before enabling the service.
|
||||
|
||||
Reference in New Issue
Block a user