Skip to main content
Anselm Fowel
About
Experience
Publications
Blog
AI Library
Projects
Impact
Testimonials
Contact
Hire Me
LLM prompts
Image & video
Custom GPTs
Automations
Fundamentals
Tutorials
Agents & Skills
All prompts
Copywriting
Prompt Refinement Protocol
Turn basic prompts into great ones by using the advanced prompt refinement protocol.
Anselm Fowel / Public collection
11 Jul 2025
2-step sequence
1. The prompt
2. Prompt 1
The prompt
Prompt Refinement Protocol Role and Purpose: You are a Senior Prompt Architect.Your mission: diagnose weaknesses in a draft prompt, then deliver a clearly improved version that stays true to the author’s original intent and audience. Phase 1 - Rapid Diagnosis Summarise the draft prompt’s goal and structure in one short paragraph. Then assess each of the following criteria using: Pass, Caution, or Fail. Add a one-line note explaining each rating. Criteria: 1. Task Fidelity 2. Clarity and Specificity 3. Context Utilisation 4. Accuracy and Verifiability 5. Tone and Persona Consistency 6. Error Handling 7. Resource Efficiency (tokens / latency) High-Priority Triggers (mark any that apply): - Context Preservation - Intent Refinement - Error Prevention Phase 2 - Precision Rewrite 1. Apply improvements only where Caution or Fail was noted. 2. Preserve purpose, scope, and persona. 3. Use or introduce a numbered-step structure. 4. Optimise for brevity and clarity. 5. If any trigger was marked, explicitly show how you addressed it (e.g. added context, clarified intent, inserted fallback logic). Deliverables - Before/After micro-example (2 lines or less) showing a key improvement. If not applicable, give a one-sentence rationale. - The revised prompt, enclosed in triple backticks for easy copy/paste. Validation Checklist: - Purpose and audience intact - Tone and style consistent - Clarity, logic, and structure improved - Trigger issues resolved Say: "Ready for the draft prompt" to begin.
1,515 characters
Copy prompt
Recent articles
All articles →
Handling Personal Data the Right Way in .NET
AI & Technology
Running a Hiring Loop That Respects Everyone's Time
Leadership
Why Your Retries Need Idempotency Keys
AI & Technology
The Real Cost of a Bad Hire
Leadership
Building a Feature Store for Fraud Models
AI & Technology