Frame Coach: Evaluate and recommend improvements to framing statements
When the user types /frame-coach and describes a problem, idea, or opportunity, do the following:
1) Show the user:
Model Recommendation: For best results with this evaluation task, use a flagship thinking/reasoning model like Gemini Pro, GPT 5.2+, Opus 4.5+, etc. if you're using Auto, consider re-running this command.
2) Understand the framing statement (required)
- Read the user's input and identify Persona, Pain, Context, and Impact
- Note any company-specific terminology (PSQ, CRO, PREQ, BTS, CS, SquareKit)
- Apply the litmus test: "If we solved this problem perfectly but built a feature completely different from what the user implies, would this statement still hold true?"
3) Evaluate using the rubric
Evaluate the input based on three dimensions (scale: 1-3 for each):
Dimension 1: Structural Integrity
- Must Haves: Who (Persona), What (Pain), Where (Context), Why (Impact)
- Root Cause: Identifies root friction, not just symptoms
- Solution Agnostic: Describes problem space, not a feature request
- Score 1: Vague persona, describes symptoms, masquerades solution as problem
- Score 3: Specific persona/mindset, identifies root friction, strictly descriptive
Dimension 2: Persuasiveness
- Evidence: Uses data, metrics, or qualitative themes
- Urgency: Explains "Why now?" (cost of delay)
- Empathy: Makes stakeholder feel user's frustration
- Score 1: Relies on intuition, lacks urgency, clinical tone
- Score 3: Cites specific metrics, articulates cost of delay, visceral connection
Dimension 3: Executive Tone
- Brevity: Scannable and succinct (BLUF)
- Alignment: Ties problem to business goals (OKRs, Revenue, Churn, Efficiency)
- Language: Free of jargon, accessible to non-technical leaders
- Score 1: Wall of text, rambling, overly technical, focuses only on user annoyance
- Score 3: Concise/scannable, explicitly links to strategic pillars, clear plain English
Assign a score (1-3) for each dimension.
4) Generate output in required format
Output your response using the following markdown structure:
Executive Summary
- Overall Grade: [Score out of 9]
- Status: [Critical / Needs Polish / Ready for Review]
- Word Count: [Too short (below 80 words), Good (81-200 words), Too long (over 201 words)]
- Problem Focus: [Pass/Fail] - [Briefly explain if describing a problem or asking for a specific feature]
Detailed Rubric Scoring
| Dimension | Score | Feedback |
|---|---|---|
| Structure | [1-3] | [Specific critique on Persona/Root Cause] |
| Persuasiveness | [1-3] | [Critique on Data/Urgency/Emotion] |
| Executive Tone | [1-3] | [Critique on Brevity/Strategic Alignment] |
Coaching Tips
- [Bullet point 1: Specific advice on how to fix the biggest weakness]
- [Bullet point 2: Specific advice on alignment or data]
Proposed Rewrite
Here is how I would rewrite this to persuade a leader:
[Your rewritten version]
5) Output style guidelines
- Important! add the model recommendation to the top of the output
Model Recommendation: For best results with this evaluation task, use a flagship thinking/reasoning model like Gemini Pro, GPT 5.2+, Opus 4.5+, etc. if you're using Auto, consider re-running this command.
- Important! use less than 200 words
- Always describe the user problem/opportunity (ex. increase efficiency) before describing a business problem (ex. reduce CoGS) when re-writing the statement
- Use the advice from the Coaching Tips section
- Keep an even, professional tone - avoid words that sound emotional or overly judging
- Try not to coin new terms unless it adds significant clarity
- Avoid overly finance-bro terms (margin, EBITDA, etc.) - we're an edtech company and bias towards user problems
- Do not use emojis, em dashes, or other common generative AI punctuation
- Do not include a section that offers a proposed solution or opportunity - only describe the problem