Add CRUD endpoints for script variables with validation and diff-compatible
models (ScriptEndpoints, ScriptApiModels, IScriptService, InMemoryScriptService).
Update PlatformScopes and PlatformPolicies for script.read/write permissions.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Implement remediation-aware health checks across all Doctor plugin modules
(Agent, Attestor, Auth, BinaryAnalysis, Compliance, Crypto, Environment,
EvidenceLocker, Notify, Observability, Operations, Policy, Postgres, Release,
Scanner, Storage, Vex) and their backing library counterparts (AI, Attestation,
Authority, Core, Cryptography, Database, Docker, Integration, Notify,
Observability, Security, ServiceGraph, Sources, Verification).
Each check now emits structured remediation metadata (severity, category,
runbook links, and fix suggestions) consumed by the Doctor dashboard
remediation panel.
Also adds:
- docs/doctor/articles/ knowledge base for check explanations
- Advisory AI search seed and allowlist updates for doctor content
- Sprint plan for doctor checks documentation
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Empty states:
- Audit Log "Recent Events" table: show "No audit events recorded yet." when empty
- Audit Log "All Events" table: show "No events match the current filters." when empty
Descriptions:
- Jobs & Queues: trim verbose 4-line description to single line
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Naming consistency:
- Dashboard heading "Mission Board" → "Dashboard" to match breadcrumb/title
- Vulnerabilities page: "Artifact workspace" → "Vulnerabilities" (heading, breadcrumb, route title)
- Doctor checks: "NOT RAN" → "NOT RUN" (grammar fix)
- Doctor pack categories: add label overrides for "servicegraph" → "Service Graph" and "binaryanalysis" → "Binary Analysis"
- Release digest placeholder: "digest-unavailable" → "Pending digest"
Bug fixes:
- Register locale data for all 8 supported locales (bg-BG, de-DE, ru-RU, es-ES, fr-FR, uk-UA, zh-TW, zh-CN) to fix NG02100 InvalidPipeArgument errors on Audit Log and other pages using Angular built-in pipes
- Null-safe json/number pipes in audit-log components (audit-log-table, audit-event-detail, audit-integrations, audit-export)
- Approval version fallback: use empty string instead of releaseName to prevent duplicate text in approval cards
- Approval card template: hide version span when it matches the release name
Layout/UX:
- stella-page-tabs: enable horizontal scroll on desktop (was mobile-only), prevents tab wrapping on Diagnostics (11 tabs), Audit Log (9 tabs)
- Triage date formatting: use DateFormatService for locale-aware dates instead of bare toLocaleString()
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add nginx proxy blocks for /api/v1/release-orchestrator/,
/api/v1/release-control/, /api/v2/releases/, /api/v1/releases/,
/api/v1/registries/ in Dockerfile.console
- All release UI calls now reach JobEngine (401 not 404)
- Registry search reaches Scanner service
- Pipeline page uses ReleaseManagementStore (real API, no mock data)
- Deployment wizard uses BundleOrganizerApi for create/seal
- Inline version/hotfix creation in deployment wizard wired to API
- Version detail shows "not found" error instead of blank screen
- Version wizard has promotion lane + duplicate component detection
- Sprint plan for 41 missing backend endpoints created
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: the | slice pipe was used in the template but SlicePipe
was not in the standalone component's imports array. This caused
Angular's resolveDirective to throw 'Cannot read factory' on every
change detection cycle, preventing mock version cards from rendering
and breaking the Continue button validation.
Also: removed unused RouterModule import, converted computed signals
to methods for PlatformContextStore-dependent values, added
platformCtx.initialize() in constructor.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Create unified releases pipeline page with decision capsules
(Deploy, Approve, Review Gates, View Evidence, Promote)
- Replace raw select filters with app-filter-bar on releases and activity pages
- Dashboard: single-column layout with Pending Actions card (pipeline + action
badges), 4-column status lane (Vuln Summary + Feed Status | SBOM Health |
Env Health | Environments at Risk), loading skeleton, reduced-motion support
- Sidebar: Dashboard at Release Control root, flat menu items
(Releases, Versions, Approvals, Activity), remove Promotions/Hotfixes
- Metric card labels: proper font size with ellipsis + title tooltip
- Badge cap changed from 99+ to 9+
- Action badges on sidebar: blocked gates, critical findings, failed runs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add UseIdentityEnvelopeAuthentication() to Authority service, fixing
/console/tenants 403. Authority was the last service missing it.
- Add ui.preferences.read and ui.preferences.write to OIDC scope request
in config.json and PlatformServiceOptions, fixing preferences/language 403.
The scopes existed in Authority but were never requested by the frontend.
All services now have identity envelope middleware:
Gateway, Platform, Scanner, JobEngine, Timeline, Integrations,
Concelier, Authority (8/8 complete)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Platform was the ONLY service missing UseIdentityEnvelopeAuthentication().
Sprint 008 added it to Scanner, JobEngine, Timeline, Integrations, Concelier
but missed Platform. This caused ALL context API calls to return 403:
- /api/v2/context/regions
- /api/v2/context/environments
- /api/v2/context/preferences
The gateway strips JWT and sends signed identity envelope headers instead.
Without the middleware, Platform couldn't read the scopes from the envelope,
so every authorization check failed.
This also fixes the "Events: DEGRADED" SSE status indicator since the
context store's error was caused by these 403 responses.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix 1 (CRITICAL): BackendProbeService now auto-recovers from "unreachable".
When probe fails, schedules retry every 10s (max 5 attempts). Guard
re-probes before redirecting if probe was previously reachable. This
fixes ALL 9 guarded route groups that redirected to /setup mid-session.
Fix 2: wasEverAuthenticated latch now persists in sessionStorage instead
of class field. Survives page reloads (F5), cleared on logout.
Fix 3: Console-admin routes now use requireAnyScopeGuard with ui.admin
bypass instead of plain requireAuthGuard. Admin user can access all
/console-admin/* pages (tenants, clients, tokens, branding).
Fix 4: Governance sub-tab routerLinks changed from relative to absolute
paths. Clicking Trust Weights, Staleness, etc. now stays on governance
instead of navigating to random pages.
Fix 5: Integration empty-state buttons show proper display names
("Runtime Host" not "runtimehost", "SCM Integration" not "scm").
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- integration-hub-ui.component.spec: fix integrationId → id property
- orphan-revival-regression: fix index signature access for getViewMode
- integration-detail-page.spec: fix mock Integration type
- Install @vitest/browser-playwright for test runner
- Sprint 025 FE-CLN-004: DONE — build verified, cleanup confirmed clean,
test runner Karma→Vitest migration is infrastructure not regression
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Backend (Scanner .NET):
- New ScanPolicyEndpoints.cs with GET/POST/PUT/DELETE /api/v1/scan-policies
- In-memory ConcurrentDictionary storage (no migration needed)
- Auth: scanner:read for list, orch:operate for mutations
- Registered in Scanner Program.cs
Frontend (Angular):
- New scan-policy.component.ts with table view, inline create/edit form,
enable/disable toggle, dynamic rules (type/severity/action)
- Route added at /security/scan-policies in security-risk.routes.ts
Gateway route already exists in router-gateway-local.json.
Sprint 002: all 7 tasks now DONE.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
flicker, pack creation, export tooltip, audit guidance
Sprint A: Harbor fixture now returns realistic search results (7 repos)
and artifact digests (3 versions with tags). Release creation wizard
Step 2 now shows actual images to select.
Sprint B: Scan polling caps at 60 polls (3 min). Shows timeout banner
with guidance link to Scheduled Jobs and "Keep Waiting" button.
Sprint C: /console/profile route now renders InsufficientPermissions
component instead of 404. Shows user/tenant, guidance, and nav links.
Catches all 24 guard redirect dead-ends.
Sprint D: Event stream chip no longer flickers DEGRADED during context
reloads. Loading state treated as connected (transient, not error).
Sprint E: Policy Packs empty state now has inline Create Pack form.
Calls existing PolicyApiService.createPack() backend endpoint.
Sprint F: Diagnostics Export button shows disabled tooltip "Run a
diagnostic check first" when no results available.
Sprint G: Audit Log shows guidance text when all module counts are 0.
Lists automatically captured event types. Confirms audit is active.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Harbor fixture returns plain text, not JSON. The RegistrySearchEndpoints
deserialization crashed with an unhandled JsonException causing 500 on
every /api/v1/registries/images/search request. This blocked the release
creation wizard at Step 2 (Components).
Fix: catch JsonException and return empty results gracefully. The release
wizard now shows "no results" instead of silently failing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: AuthSessionStore.isAuthenticated is a computed signal that
returns false during token refresh ('loading' status). Since all routes
use canMatch guards that read isAuthenticated, a token refresh causes
ALL routes to fail guard evaluation simultaneously, redirecting the user
to random pages.
Fix: Add wasEverAuthenticated latch that stays true once set. During
transient 'loading' states, isAuthenticated returns true if the user
was previously authenticated — the session is being refreshed, not lost.
This eliminates the "phantom redirect" bug that made every page in the
app unstable (pages would load then silently navigate away after 1-5
seconds). Verified stable on /setup/identity-access and /evidence/audit-log
with 12-second wait after navigation.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Policy Packs: show "No policy packs configured" with description and
link to overview when pack list is empty
- Shadow Mode: add prerequisite text below disabled Enable/View Results
buttons — "Shadow mode requires at least one active policy pack" with
link to Packs tab. Applied to both indicator and dashboard components.
- JobEngine: show guidance "No jobs have been submitted yet..." when all
counts are 0, auto-hides when jobs appear
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sprint 020 complete: scan ID mismatch, user ID display, feed status text,
route redirect race condition, exception scope mismatch, admin scope bypass.
All verified via Playwright on fresh install.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Critical fixes:
- Replace router.navigateByUrl() with Location.replaceState() in
PlatformContextUrlSyncService to prevent re-evaluating canMatch guards
during query param sync. This was causing random page redirects across
all routes when auth session signals hadn't settled yet.
- Fix exception scope mismatch: Authority issues 'exceptions:read' (plural)
but guards checked 'exception:read' (singular). Aligned to plural form.
- Fix admin scope bypass: guards checked 'admin' scope but token has
'ui.admin'. Now both are accepted as superuser bypass.
- Remove duplicate scope entries in description map.
UX polish (from fix agents):
- Integration detail: formatActor() truncates raw user ID hashes to
"User 9a2d0730..." instead of showing full 32-char hex string.
- Dashboard feed status: show "Not checked yet" instead of "0 healthy"
when no advisory source health checks have run.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Bug: canSubmit computed() used plain string imageRef which computed()
can't track. Submit button was permanently disabled.
Fix: Changed imageRef to signal(''), updated template to use
[ngModel]="imageRef()" (ngModelChange)="imageRef.set($event)",
updated all class references to imageRef() for reads and .set() for writes.
Bug: Scan submit called /scanner/api/v1/scans/ (console nginx prefix)
instead of /api/v1/scans/ (gateway route).
Fix: Removed /scanner/ prefix from all API URLs.
KNOWN ISSUE: Scanner service needs identity envelope middleware (same
as Concelier fix) — ReverseProxy strips JWT, scanner returns 400
"tenant required". The scan submission reaches the scanner but auth
fails. Fix requires adding the same pre-auth envelope middleware
pattern to the Scanner service.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The scan route was added to security.routes.ts which isn't loaded by the
app router. The app loads security-risk.routes.ts at /security/*. Added
the scan route to security-risk.routes.ts so /security/scan resolves.
Verified: Scan Image page loads at /security/scan with heading, image
input, and submit button. 111/111 canonical routes passing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
S5-T02: Dashboard 3-column layout with real APIs
- CSS Grid: security posture (1/3) | environments + actions (2/3)
- Left column: Vulnerability Summary (from VULNERABILITY_API getStats()),
SBOM Health (from computed sbomStats), Advisory Feed Status (from
SourceManagementApi getStatus())
- Right column: mission summary strip (4 cards), promotion pipeline
(env card grid), environments at risk table, quick links
- Footer: platform health bar with status dots
- Real API calls with independent loading states and catchError defaults
- Refresh button re-fetches all data
- Responsive: collapses to single column on mobile
- Welcome guide still spans full width when no environments
- Removed old: reachabilityStats signal, nightlyOpsSignals signal,
NightlyOpsSignal interface, TitleCasePipe/UpperCasePipe imports
S5-T03: Security reports as downloadable exports
- New shared utility: download-helpers.ts (downloadCsv, downloadJson)
- Risk Report tab: "Export CSV" + "Generate PDF" (print-friendly CSS)
- VEX Ledger tab: "Export VEX Decisions" as JSON download
- Evidence Export tab: explainer + "Open Export Center" link
- Tests updated: 6 new test cases for export functionality
All 6 sprints now complete. Angular build: 0 errors.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sprint 4 — Sidebar restructure (S4-T01+T02):
5 groups: Release Control, Security, Operations, Audit & Evidence, Setup & Admin
Groups 4+5 collapsed by default for new users
Operations extracted from Release Control into own group
Audit extracted from Security into own group
groupOrder and resolveMenuGroupLabel updated
Approvals badge moved to section-level
Sprint 2 — Demo data badges (S2-T04+T05):
Backend: isDemo=true on all compatibility/seed responses in
PackAdapterEndpoints, QuotaCompatibilityEndpoints, VulnerabilitiesController
Frontend: "(Demo)" badges on Usage & Limits page quotas
Frontend: "(Demo)" badges on triage artifact list when seed data
New PlatformItemResponse/PlatformListResponse with IsDemo field
Sprint 6 — Audit emission infrastructure (S6-T01+T02):
New shared library: src/__Libraries/StellaOps.Audit.Emission/
- AuditActionAttribute: [AuditAction("module", "action")] endpoint tag
- AuditActionFilter: IEndpointFilter that auto-emits UnifiedAuditEvent
- HttpAuditEventEmitter: POSTs to Timeline /api/v1/audit/ingest
- Single-line DI: services.AddAuditEmission(configuration)
Timeline service: POST /api/v1/audit/ingest ingestion endpoint
- IngestAuditEventStore: 10k-event ring buffer
- CompositeUnifiedAuditEventProvider: merges HTTP-polled + ingested
Documentation: docs/modules/audit/AUDIT_EMISSION_GUIDE.md
Angular build: 0 errors. .NET builds: 0 errors.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sprint 4:
S4-T04: Empty state guidance for sparse pages
- Unknowns: graceful "No unknown components" with "Scan an Image" link
(replaces API error message)
- Evidence Capsules: enhanced with "Create your first release" guidance
S4-T05: Canonical route cleanup
- All /administration/* routes now redirect to /setup/* canonical
- /admin/* redirects to /setup instead of /administration
- 8 admin-only routes migrated to /setup/* (identity-providers, offline,
configuration-pane, security-data, workflows, ai-preferences)
- Sidebar nav config updated to /setup/identity-providers
- Settings routes updated to /setup/* targets
Sprint 6:
S6-T03: Record Decision as modal overlay
- Both decision-drawer and decision-drawer-enhanced converted from
sliding drawer to centered fixed-position modal (z-index 1000)
- Semi-transparent backdrop, click-outside-to-close, scrollable body
- Max width 480/520px, max height 90vh
S6-T04: Topology wizard skip agent step
- "Skip — assign agent later" button when no agents available
- Deploy Agent guidance with CLI command
- Next button shows "Next (without agent)" when skipping
- Done summary shows "None (skipped)" for agent
- agentSkipped signal in wizard service
Angular build: 0 errors.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Walked through the full vulnerability triage workflow as a security
engineer. Found the artifact workspace — the product's killer feature
(evidence-grade findings with reachability, attestations, policy gating,
delta comparison, deterministic replay, VEX decisions). Recorded a VEX
decision for CVE-2023-38545.
Critical UX findings:
- UX-D1: No "Scan" entry point anywhere in the UI — scanner exists
(2 containers) but has no discoverable trigger from the console
- UX-D2: Triage workspace (best feature) hidden under "Triage" label —
security engineers look for "Vulnerabilities" or "Findings"
- UX-D3: Record Decision dialog unreachable on smaller viewports —
needs proper modal overlay instead of in-page drawer
- UX-D4: Security Posture shows 0 findings while Triage has 1 active
HIGH finding — different data sources
Assessment: The triage artifact workspace is 10/10 UX. The discoverability
is 2/10. Three changes would transform the security engineer experience.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
