MEL

Machine Expression Language: A guided journey to craft powerful, directive-based LLM prompts.

Step 1: Define the Persona

Student Assistant Collaborator Expert Specialist Mentor Visionary Devil's Advocate Oracle Simulator

Step 2: Define the Task

Execute Achieve Synthesize Precise-Exec Brainstorm Explain Critique Design Deconstruct Tutor

Step 3: Define the Audience

None A Child A Novice A Peer A Manager A Skeptic An Expert A Regulator Posterity The Self

Step 4: State the Purpose

None Practical Educational Durable Strategic Commercial Impactful Artistic Provocative Transformative

Step 5: Set the Innovation Level

Standard Conceptual Blend Unconventional Surprising Analogical Challenging First-Principles Heresy Paradigm Shift Exotic

Step 6: Choose the Style

Raw Text Direct Formatted Chain-of-Thought Analytical Narrative Socratic Literary Poetic Masterpiece

Step 7: Shape the Output

Raw Text Bullets Markdown Table Presentation Mind Map Plan Formal Paper Polished Artifact Dialogue

Step 8: Add a Constraint (Optional)

None Concise Word Limit Formal Tone No Jargon In-Character Step-by-Step Bullets Only Single Para Strict

Step 9: Set the Polish Level (Optional)

None Proofread Improve Flow Check Accuracy Add Summary Review Directives Simplify Elevate Tone Add Insight Make Inevitable
To use with OpenClaw: Paste the Message Prompt for a one-off task, or save the New Skill File as a .md file to create a reusable skill.

MEL v1.0 • github.com/polonski/mel

Experimental prompting syntax — January 2026