← Agent catalogue
TEMPUS

Preview portrait

Execution stages

TEMPUS

Short remediation steps / plan status — not a full task system.

LiveCatalogue status
Execution activeRuntime posture

Action plan: governs what happens next from prior stages + approvals.

Expanded capabilities

  • Compose short, human-readable next-step plans from prior stages
  • Reflect approval and VIGILIA posture in plan status
  • Stay advisory — not a full task or calendar system
  • Surface plans on interaction detail when workflow evidence exists
  • Defer execution of recommended steps to human/approved paths

How they learn & improve

TEMPUS improves plan usefulness from which steps operators actually complete or dismiss — it never schedules silent automations.

  1. Read completed stage outputs and approval state.
  2. Propose a short remediation/action plan block for the operator.
  3. Tighten future step ranking after FEEDBACK and approved learning apply.

How they work together

TEMPUS turns MONETA/VIGILIA/NUNTIUS outcomes into ‘what next’ for humans, while WORKFLOW remains the durable executor.

  • VIGILIAIncorporates control posture into next steps.
  • MONETAUses finance stage outcomes when present.
  • Approval queueCalls out pending decisions as blocking steps.
  • RESPONSE (operator-facing outcome)Feeds structured plan text into operator read models.
  • DISPATCHShares governed context with DISPATCH so operators see one coherent path instead of conflicting module stories.

Practical benefits

  • Clear next moves after a scanOperators get a short checklist instead of re-reading every stage payload.
  • Blocked steps stay honestPending approvals appear as blockers — TEMPUS won’t pretend work is ready.
  • Plans that learn from dismissalsIgnored steps become FEEDBACK so future plans stay shorter and more relevant.

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.

TEMPUS — AgentIQ Agent Catalogue · AgentIQ Labs