Agent Governance

Run Logging

Agent runs are append-oriented governance records.

Use this when...

Use this when you need detailed reference information for Agent Governance.

Agent runs are append-oriented governance records.

Each manual run records:

  • Agent profile
  • Trigger type
  • Input request
  • Guardrail result
  • Step log
  • Draft output where allowed
  • Blocked reason where blocked
  • Related proposals, information requests, handoffs, and evaluations

Runs never directly mutate CRM, quote, finance, project, support, inventory, import, or knowledge records.

Sprint 22B Detail View

Agent Run detail pages show a human-readable summary before technical details. The summary includes agent name, run status, trigger type, source module/entity, what happened, blocked behavior, related proposals/information requests/handoffs, target-mutation status, and the next recommended human action.

Raw input, output, and full metadata remain available under Developer details for audit and debugging.

Developer Details
Source
apps/docs/content/agents/run-logging.md
Owner
Module Owner
Status
Current
Module
Agent Governance
Related route
/agents