Operate
Dashboard
Live overview of the connected openMiura instance. Visual by default; click raw on any card to fold out the JSON returned by the underlying endpoint.
Connect to load dashboard data
Open the auth dropdown in the topbar, configure the broker base URL and provide a token or login.
Active scope
from /auth/me- Tenant
- Workspace
- Environment
- Principal
- Role
No identity payload available.
Openclaw runtimes
Sessions
Canvas documents
System
Recent admin events
Loading…
No events.
| When | Source | Actor | Action |
|---|---|---|---|
Operate
Runtimes
Registered OpenClaw runtime adapters. Select a row to inspect its concurrency, alerts, notification targets and alert routing. The legacy HTTP path /admin/openclaw/* is preserved on purpose; see docs/_archive/refactor/ for context.
Connect to load runtimes
Use the auth dropdown in the topbar.
Registered runtimes
Loading…
No runtimes registered for this scope. No runtimes match the filter.
| Runtime | Environment | Tenant / Workspace | Updated | |
|---|---|---|---|---|
| / |
Selected runtime
Concurrency
/
—
Alerts
Notification targets
Alert routing
Selected runtime payload
Govern
Policies
Registered policy packs and the current policy snapshot, plus the three explorer endpoints: explain a decision, simulate a candidate YAML, and diff a candidate against a baseline. All three are read-modeled — they do not mutate the active policy.
Connect to load policies
Use the auth dropdown in the topbar.
Registered policy packs
Loading…
No policy packs match.
Policy snapshot
Loading…
Status:
Inspect the verbatim JSON for the active policy:
Explain a decision
POST /admin/policies/explainAsk the policy engine why a request was — or would be — allowed or denied. Fields with default values can be left blank.
Simulate a candidate policy
POST /admin/policy-explorer/simulateReplace the active policy with the YAML below and resolve the same request. The active policy is not modified.
Diff candidate vs baseline
POST /admin/policy-explorer/diffBoth YAMLs are optional — empty means "use the active policy". Samples is a JSON array of decision requests that get evaluated under both policies; the response lists which samples change verdict.
Last pack application
POST /admin/openclaw/runtimes/<id>/policy-packapplied
Govern
Secrets
Secret-governance surface for compliance and forensics. The catalog lists declared refs in the active scope; the values themselves never leave the backend. The explain form is read-modeled — it replays a decision without consuming the secret.
Connect to load secret governance
Use the auth dropdown in the topbar.
Filters
applied to catalog / timeline / usageSummary
GET /admin/secrets/summaryLoading…
Total events
Allowed
Denied
Deny rate
Catalog
Loading…
No declared secret refs in this scope.
| Ref | Source | Last seen | Tools |
|---|---|---|---|
Timeline
Loading…
No timeline events match.
Usage
Loading…
No usage groups match.
Explain access
POST /admin/secrets/explainReplay a secret-access decision for the given ref + tool and surface the policy reason. The secret value is never read or returned.
Govern
Identities & RBAC
Identity-link table (channel → global user key) and the read-modeled explain endpoints for sandbox capabilities and security access. Linking an identity is audit-only — it records the binding, it does not grant a role. Roles are derived from the principal scope (tenant / workspace / environment / role).
Connect to load identities
Use the auth dropdown in the topbar.
Identity links
Loading…
No identity links in this scope.
| Channel user key | Global user key | Linked by | Linked at |
|---|---|---|---|
Active sessions
Loading…
No active sessions.
| Session ID | Principal | Tenant / Workspace | Channel | Started |
|---|---|---|---|---|
| / |
Sandbox capabilities
POST /admin/sandbox/explainResolve the capability bundle for a (role × scope) tuple. Useful when an operator reports "tool X disappeared from my menu".
Allowed tools
Security access
POST /admin/security/explainWould this (role × scope) access the given resource? Like the policies B3 explain, but routed through the security layer.
Configure
Channels wizard
Inspect and reconfigure the channel adapters (Slack, Telegram, voice, ...). Two-step pipeline: validate is read-modeled and safe to call repeatedly; save rewrites the on-disk YAML and lives behind a confirmation modal.
Connect to load channels
Use the auth dropdown in the topbar.
Registered channels
GET /admin/config-center/channels-wizardLoading…
No channels registered.
Editor
Editing the raw YAML for . Validation parses the candidate without touching disk; save rewrites it.
The save succeeded but the channel needs a restart to pick the new config. Schedule it manually.
Last test message
POST /admin/channels/<name>/testdelivered
Operate
Evidence packs
Compliance summary + evidence-pack export for the active scope. The summary is read-only and surfaces a per-section count for the chosen time window. Export builds a downloadable pack and records the action with a report_label the auditor sees later.
Connect to load evidence
Use the auth dropdown in the topbar.
Scope & window
passed to summary + exportCompliance summary
GET /admin/compliance/summaryLoading…
Last export
POST /admin/compliance/exportexported Artifact ref:
The export was recorded on the audit trail; pull the artifact from the backend via its ref.
Configure
Secrets wizard
Inspect and reconfigure the secret-reference profiles (which env vars get mapped to which secret:// refs). Two-step pipeline: validate is read-modeled and safe to call repeatedly; save rewrites the on-disk dotenv reference file and lives behind a confirmation modal.
Connect to load secret profiles
Use the auth dropdown in the topbar.
Registered profiles
Loading…
No profiles configured.
Editor
Editing the raw YAML for . Validation parses the candidate without touching disk; save rewrites it.
The save succeeded but the profile needs a restart to pick the new mapping. Schedule it manually.
Operate
Dispatches
OpenClaw dispatch records across runtimes. Click a row to inspect its full timeline and act on it: cancel aborts a pending dispatch, retry re-enqueues with optional overrides, reconcile manually pins a canonical status (escape hatch when retry/cancel cannot untangle a half-applied state).
Connect to load dispatches
Use the auth dropdown in the topbar.
Filters
GET /admin/openclaw/dispatchesDispatches
Loading…
No dispatches match the filters.
| Dispatch id | Runtime | Action | Status | Updated | |
|---|---|---|---|---|---|
Selected dispatch
Loading…
Dispatch payload
Last action
applied
Govern
Approvals
Pending approvals across the active scope. Acting opens a confirmation modal — every decision lands on the audit trail with the reason text and the actor's principal id.
Connect to load approvals
Loading…
No approvals match the current filter.
Debug
Event log
Stream of recent admin events for the active scope. Read-only window onto what openMiura has recorded.
Connect to load events
Loading…
No events in the window.
Debug
Tool calls
Decision traces: every tool invocation the policy engine evaluated. Click a row to inspect the verbatim trace payload.
Connect to load traces
Filters
Traces
Loading…
No traces match the filters.
| Trace id | Tool / action | Status | Session | Timestamp |
|---|---|---|---|---|
Selected trace
Loading…
Configure
Workflows
Workflow records across the active scope (lists via the operator console, kind=workflow). Click a row to inspect its timeline; act on it (verb in the URL — the backend admits the appropriate actions).
Connect to load workflows
Loading…
No workflows in this window.
Selected workflow
Loading…
Configure
System & config
Full openMiura config-center: inspect every YAML section, edit it, validate it (read-modeled), save it (confirmed write to disk). The reload-assistant card below decides which sections can hot-reload vs. require a restart.
Connect to load system config
Configuration sections
GET /admin/config-centerLoading…
No sections registered.
Editor
Reload assistant
Loading…
applied
Operate
Portfolio evidence
Portfolio-aware evidence packs. Each portfolio carries its own bundle: export new packs, prune old ones, verify the integrity of a specific artifact, restore one from cold storage. All writes confirmed by modal.
Connect to load portfolios
Portfolios
GET /admin/openclaw/alert-governance/portfoliosLoading…
No portfolios in this scope.
Selected portfolio
Loading…
No evidence packages in this portfolio.
| Artifact ref | Report label | Status | Generated | |
|---|---|---|---|---|
Last action
applied
Pick a section
All 15 admin sidebar items are wired up. Pick one in the sidebar.