Multi-Agent Workflow Circuit Breaker Policy Directive
Author an operational planning email establishing guardrail thresholds, fallback loops, and triage protocol for agent chains.
Use this template when defining runtime failure thresholds, human-in-the-loop intervention policies, and recovery plans for complex workflow chains. It delivers a crisp executive directive email for operational stakeholders.
Role: Director of Autonomous Workflow Reliability with deep domain expertise in agentic control loops and fault-tolerant system architecture.
Context
- Workflow pipeline: {{workflow_pipeline_name}}
- Error threshold trigger: {{failure_threshold_rate}}
- Human review team: {{fallback_human_in_the_loop_team}}
- Tool call timeout window: {{tool_execution_timeout}}
- Retry and backoff rule: {{retry_policy_parameters}}
- Affected business units: {{impacted_business_units}}
Task
Compose a policy alignment email to operational managers and system owners defining circuit-breaker triggers, automatic pause conditions, and triage responsibilities for chained agent workflows in {{workflow_pipeline_name}}.
Method
- State the operational scope and risk surface across {{impacted_business_units}}.
- Detail the exact conditions under {{failure_threshold_rate}} that trigger an automated chain halt.
- Outline the execution boundaries established by {{tool_execution_timeout}} for individual node steps.
- Explain the stepped recovery behavior governed by {{retry_policy_parameters}} before tripping the circuit breaker.
- Define handoff procedures to route suspended agent state files to {{fallback_human_in_the_loop_team}}.
- Specify the required operational criteria to manually reset the chain and resume autonomous execution.
- Provide an audit checklist for post-incident RCA reviews within 24 hours of any trip event.
Constraints
- MUST define clear operational ownership for every triage phase.
- MUST NOT leave human handoff SLAs undefined.
- Tone MUST remain decisive, process-driven, and focused on business continuity.
- Limit the email length to a maximum of 400 words.
- Do not include pseudocode; use clear operational instructions.
Output format
Structure the response strictly as an email with:
- Subject: Operational Policy: Circuit Breaker Rules for {{workflow_pipeline_name}}
- Policy Overview (2 sentences)
- Automated Halt & Retry Thresholds (bulleted parameters)
- Human-in-the-Loop Escalation SOP (3 numbered action steps)
- Chain Resumption Protocol (2 criteria)
Self-review
- Are the timeout and retry constraints explicitly grounded in {{tool_execution_timeout}} and {{retry_policy_parameters}}?
- Is {{fallback_human_in_the_loop_team}} clearly designated as the action owner for paused chains?
- Does the email fulfill all stated length and section requirements?
Explicit role, a named task, and discrete steps the model can follow.
Background, inputs and variables the model needs before it starts.
Hard boundaries — what the model must and must not do.
A named, field-level shape for the response.
Ordered work items that force analysis before an answer.
Length and structure that travel across frontier models.
Signal density — instruction weight without padding.
Documented variables so the scaffold adapts to new inputs.
Quality bar, assumptions and behaviour when inputs are thin.
How much real usage the template has behind it.