← Agent catalogue
Approval queue

Preview portrait

Orchestration

Approval queue

`ApprovalRequest` rows with PENDING/APPROVED/REJECTED; bridges console Approvals page and execution pauses.

LiveCatalogue status
Execution activeRuntime posture

Human-in-the-loop decisions before sensitive execution or dispatch.

Expanded capabilities

  • Hold PENDING ApprovalRequest rows for human APPROVE / REJECT / CANCEL
  • Bridge console Approvals workbench and paused ExecutionJobs
  • Cover NUNTIUS dispatch, finance, module activation, and multi-agent plans
  • Preserve actor identity and reasons on every decision
  • Refuse silent auto-approval under all autonomy bands

How they learn & improve

The queue improves triage when operators mark stale items, add reasons, and FEEDBACK false escalations — it never decides on their behalf.

  1. Materialise PENDING work when POLICY or a draft path requires consent.
  2. Capture human decisions with audit-safe attribution.
  3. Feed decision patterns into learning candidates for routing/calibration only.

How they work together

APPROVAL QUEUE is the human switchboard: POLICY creates work, WORKFLOW pauses, NUNTIUS waits, and AUDIT records the outcome.

Practical benefits

  • One inbox for sensitive workFinance scans, module activations, and message drafts share the same decision surface.
  • Mobile-ready operator triageApprovals workbench cascades to phone with ≥44px targets — decisions don’t wait for a desktop.
  • Undo-minded governanceReject/cancel keep the system honest; nothing ‘sneaks through’ when the queue is quiet.

Learning and improvement stay human-visible: candidates are RECORD_ONLY until an operator reviews and applies them. Sensitive sends, publishes, and approvals never run silently.

Approval queue — AgentIQ Agent Catalogue · AgentIQ Labs