← All skills
🔁
Prompt Engineering
Self-Eval Prompt Wrapper
Forces the AI to grade its own output and rewrite if it scores below 8 of 10. Quality control built in.
Step 1
Get the skill
Pick one. Copy to paste straight away, or download self-eval-prompt.md to keep.
Step 2
Paste it into your AI
Wrap any high-stakes prompt with this. Saves a manual review pass on important content.
▸Preview the skill· what gets copied
Run this task with a self-check loop.
Task: {{TASK}}
Output rubric (1 to 10): {{RUBRIC · accuracy, clarity, completeness, brand voice}}
Process:
1. Generate the first draft.
2. Score the draft against each rubric criterion (1 to 10 with 1-line reason).
3. If any score is below 8, rewrite the draft.
4. Score again.
5. Return only the final draft that scores 8+ on every criterion.
Show your scores for transparency at the end.Works with: Claude · ChatGPT · Gemini · Any model
By Rishi Jain
More skills like this
📐
Output Format Forcer
Stops the AI from giving you walls of text. Forces JSON, tables, CSV, or numbered lists. Every single time.
⚡CRAFT Prompt Framework
Rishi Jain's 5-part prompt structure. Context, Role, Action, Format, Tone. Works across every AI model.
🧩Chain-of-Thought Prompt
Force the AI to think step-by-step before answering. Reduces hallucination, surfaces reasoning, catches mistakes.