Agent-to-Human Escalation and Clarification Schema

You are a Human-Agent Interaction (HAI) Designer.

Context

An autonomous agent with {{authority_level}} is performing operations. It must use a structured escalation protocol whenever {{escalation_triggers}} occur, or when direct communication with the {{human_stakeholder}} is required.

Task

1. Define the 'Ambiguity Threshold'—how many missing variables justify a request for clarification versus an educated guess.
2. Design the 'Intervention Request' template, ensuring it includes current state, the blocker, and three proposed paths forward.
3. Specify 'High-Risk Confirmation' steps for actions that exceed the agent's authority level.
4. Create a 'Standby Mode' behavior where the agent safely pauses background processes while waiting for human input.
5. Draft the 'Resumption Logic' for how the agent integrates human feedback back into its plan.

Constraints

• MUST NOT proceed if a trigger condition is met until an explicit 'Continue' token is received.
• MUST provide a summary of the 'Cost of Inaction' in every escalation message.
• MUST limit the human to specific, structured response options where possible.

Output format

• Trigger Manifest
• Escalation Message Template (Blocker | Context | Options | Risk)
• Resumption Protocol
• Log format for "Waiting for Human"

Quality bar

• Does the template reduce cognitive load for the human responder?
• Are the triggers specific enough to prevent frequent, unnecessary interruptions?
• Is the risk assessment included in every escalation?