DocsNew Run

Feature matrix

Implemented in this project

AreaFeatureStatusNotes
DashboardKPI cardsRuns, signals, cost, success rate
DashboardChartsVolume, latency, error, signal trends (ISO timestamp aware)
DashboardService health summaryRuntime health surfaced
ScenariosPack + scenario catalogSearch and pack filters
ScenariosScenario detailTemplates, versions, launch schema
ScenariosPolicy hintsPolicy type, threshold, veto config on scenario cards and detail
Run creationSchema-driven launch formUses launch schema defaults
Run creationRaw JSON editorSwitchable input mode
Run creationCompile launch requestExamples Service integration
Run creationValidate execution requestControl Plane integration
Run creationSubmit runRedirects to live workbench
Run creationExample end-to-end bootstrapOptional quick path
Run creationSaved launch presetsSave/load/delete from localStorage
Live runsActive runs listQuick watch cards
Live runsSSE auto-reconnectExponential backoff, heartbeat timeout, bounded buffer
Run detailGraph viewReact Flow execution graph
Run detailNode inspectorPayloads, logs, signals, traces
Run detailSignal railSide-channel update surface
Run detailLive event railCanonical events timeline
Run detailFinal decision panelAction, confidence, rationale
Run detailArtifacts/messages panelTrace/report bundle visibility
Run detailReplay descriptor requestControl Plane replay hookup
Run detailClone with overridesClone run with custom tags (context overrides rejected under observer-only CP)
Run detailPolicy governanceRFC-MACP-0012 policy panel, commitment evaluations, policy badges
Run detailTimeline scrubberInteractive visual replay with frame markers
HistoryRun history pageSearch + status/environment filtering (server-side)
HistoryCSV/JSON exportExport filtered runs as CSV or JSON
HistoryBatch operationsCancel, archive, delete, export selected runs
CompareRun comparison pageSummary + raw diff payload
CompareDecision comparisonSide-by-side confidence bars and reasons
ComparePayload diff viewerStructural diff with color-coded changes
CompareSignal timeline overlayDual-lane chronological signal view
AgentsAgent catalogSearch + framework filter, enriched with CP metrics
AgentsAgent detailScenario coverage + related runs, direct /agents/:agentRef endpoint
LogsCanonical event explorerRun selection + event filters
TracesTrace summary and artifact explorerInline span support when available
ObservabilityRuntime health dashboardManifest, modes, roots, metrics text
ObservabilityRaw metrics surfacePrometheus text dump, parsed into sortable table
ObservabilityPercentile latency KPIsp50 / p95 / p99 computed client-side from histogram buckets
ObservabilityCircuit breaker timelineState transitions from /admin/circuit-breaker/history
ObservabilityReadiness probePer-subsystem health checks (/readyz)
ObservabilityJaeger trace deep-linksSpan waterfall via /api/jaeger/* proxy when configured
RuntimeRuntime mode browser/modes surface
RuntimeRuntime policy registry/policies surface, RFC-MACP-0012 governance rules
SettingsPreferences storePersisted in local storage
SettingsWebhook managementCreate/toggle/delete, delivery stats display
SettingsCircuit breaker resetAdmin action surface
SettingsAudit log viewRecent actions
PlatformDocs viewer/docs landing + UI Console + Examples Service doc collections with Mermaid rendering
PlatformDemo modeRich mock dataset
PlatformCommand paletteRoute jumping
PlatformTheme toggleDark/light
PlatformReal-mode integrationResponse normalization layer for CP/ES compatibility
PlatformServer-side proxy BFFSecret-safe browser integration
PlatformError boundariesGlobal + per-component crash recovery
PlatformCI/CDGitHub Actions: lint, typecheck, test, build
PlatformTestingVitest + React Testing Library, unit + integration suites
PlatformPre-commit hooksHusky + lint-staged, Prettier + ESLint

Partially implemented / foundation laid

AreaFeatureStatusNotes
CollaborationComments / notes🟡Not yet built
RBACRole-aware access control🟡UI structure can support it later
AlertsThreshold-based notifications🟡Metrics surfaces are in place
  • annotation / incident-notes layer
  • prompt and policy version diffing
  • RBAC with route guards
  • simulation / dry-run mode
  • threshold-based alerting on metric surfaces