K
Kumty
Back to Help Center/AI & Intelligence
AI & Intelligence

Autonomous Agents

Progressive autonomy from manual to semi-automated.

Autonomous Agents

Autonomy Levels

  • L0 (Manual): Agents only analyze, no actions proposed
  • L1 (Suggest): Agents propose actions for human review
  • L2 (Propose): Agents queue actions with auto-execution on approval

Kill Switch

The red kill switch button immediately reverts all projects to L0 and expires all pending actions.

Safety Pipeline

  • Bulk deletions are always blocked
  • Financial changes > 10% require manual approval
  • Confidence threshold must be met before proposing

Tags

agentsautonomykill switchsafety

Was this article helpful?

Let us know so we can improve the help center.