GPT-4o
⚡ Productivity
Advanced
AI Agent Error Recovery Planner
Design the error-handling and recovery logic for an AI agent so failures degrade gracefully instead of cascading.
The Prompt
# AI Agent Error Recovery Planner You are a reliability engineer for autonomous agents. Design error recovery logic for the agent workflow below. ## Workflow Context - Workflow being protected: [WORKFLOW DESCRIPTION] - Points of failure: [LIST LIKELY FAILURE POINTS, e.g., API timeout, malformed data, rate limits] - Acceptable downtime/delay: [SLA OR TOLERANCE] ## What To Produce 1. A failure mode table: each likely failure, its detection method, and the recovery action 2. A retry strategy (exponential backoff, max attempts, circuit breaker conditions) 3. A graceful degradation plan — what reduced-functionality mode looks like if full recovery isn't possible 4. A human notification policy: what failures get escalated immediately vs. batched into a daily report 5. A post-incident review template the agent (or a human) fills out after a significant failure ## Output Format Failure mode table, retry strategy, degradation plan, notification policy, and review template as separate sections.
📝 Fill in the blanks
Replace these placeholders with your own content:
[WORKFLOW DESCRIPTION]
[LIST LIKELY FAILURE POINTS, e.g., API timeout, malformed data, rate limits]
[SLA OR TOLERANCE]
How to use this prompt
1
Copy the prompt
Click "Copy Prompt" above to copy the full prompt text to your clipboard.
2
Replace the placeholders
Swap out anything in [BRACKETS] with your specific details.
3
Paste into GPT-4o
Open your preferred AI assistant and paste the prompt to get started.