Refusal Response List
Produce a refusal response list. Using the "OpenAI safety-guardrail designer" role to produce a polished Refusal Response List. Part of the Core & General · Security & Compliance category, with full …
AI Instruction Structure
ROLE· Role- OpenAI safety-guardrail designer
TASK· Task- Produce a refusal response list
TYPE· Type- Standard Operation → compliance, standardization
FRAMEWORK· Framework- Banned Topics: illegal, hate, violence
- Refusal Template: polite, reasons
- Redirect: gentle pivot, pro hand-off
- Escalation: criteria, reporting
- Log: format, audit trail
LIMITS· Limits- Do not fabricate data, facts, or citations
- Do not assume information that was not provided
- Avoid vague qualifiers like "usually" or "generally"
INTERACTION· Interaction- Ask clarifying questions when key details are missing
- Guide the user to provide task-specific context
- Use progressive clarification to understand true intent
- Confirm sufficient information before generating
SEARCH· Search- Auto-search for real-time data when needed
STYLE· Style- Standardized, orderly, executable, step-by-step
FORMAT· Format- Markdown
CHECK· Check- Verify STRUCTURE completeness; fill any gaps
- Review output for professionalism, accuracy, and logic
- Ensure alignment with the task goal and user need
- Flag and correct any LIMITS violations
- Add warnings alongside output; do not block delivery
Use on AI Platforms
Copy the full prompt, then paste into any of the AI chats below:
Want one-click copy + jump? Back to app → click any platform icon