---
name: critique-yourself
description: Use a structured expert self-evaluation cycle to substantially improve instructional content, slide decks, and accompanying AI image prompts before presenting a final version.
---

# Critique Yourself

Use this workflow when creating or revising instructional content, presentation slides, learning experiences, or text-to-image prompts. Work through the critique and iteration privately; the user should receive only the final 10x version unless they specifically ask to see the working.

## Phase 1: Expert Self-Critique

Adopt three perspectives at once:

- Senior instructional designer
- Expert AI prompt engineer
- Award-winning visual artist

Evaluate the current output against these questions:

1. **Instruction compliance** — Which user instructions were followed well? Which were missed, weakened, or only partially addressed?
2. **Inclusion and exclusion principles** — What principles guided what was included, omitted, prioritized, or simplified? Are those choices aligned with the audience, outcome, and constraints?
3. **Instructional-design weakness** — What is the weakest part according to instructional-design best practices? Identify the specific cause: unclear objective, poor scaffolding, weak engagement, cognitive overload, missing practice, weak assessment, or insufficient context.
4. **Visual-prompting weakness** — What is the weakest artistic or text-to-image prompt? Identify the cause: vague subject, incoherent style, weak composition, unhelpful lighting, unsuitable palette, low visual hierarchy, or lack of usable negative space.
5. **Image-prompt improvements** — What concrete changes would make every prompt more likely to generate professional, high-quality AI imagery?
6. **Professional rating** — Rate the response from 1 to 10 against an ideal senior-instructional-designer standard. State the exact changes that would move it closer to 10.

Be candid, specific, and evidence-based. Avoid generic praise or vague criticism.

## Phase 2: Iterative Improvement

Based on the critique:

1. Create **Version A** and **Version B** as genuinely different improved approaches.
2. For each version, assess its strengths and weaknesses against the brief, learner needs, instructional quality, and visual quality.
3. Synthesize the strongest insights into a **10x version**: a dramatic, not incremental, improvement over the original.
4. Track all material changes internally, with a specific explanation of what changed and why.

Choose the final structure that best supports clear objectives, scaffolded complexity, engagement, and retention. Reduce cognitive load, eliminate decorative or redundant material, and make every section earn its place.

## Phase 3: Final Output

Present **only the final 10x improved version**. Do not expose the private critique, Version A, Version B, ratings, or internal change log unless the user explicitly asks for them.

The final output must include:

- Complete polished content
- For every slide or applicable section, one individual, production-ready text-to-image prompt
- Layout direction: **Image LEFT, text RIGHT**
- Any necessary notes that make the content immediately usable

## Text-to-Image Prompt Standard

Each prompt must specify:

- **Subject:** a clear, concrete main subject and action
- **Style:** an intentional artistic style, such as photorealistic, editorial illustration, cinematic 3D render, or refined watercolor
- **Composition:** camera angle, framing, positioning, and visual hierarchy
- **Lighting:** source, mood, and atmosphere
- **Color palette:** dominant colors and intended emotional effect
- **Quality modifiers:** professional, high-resolution, detailed, studio quality, or equivalent meaningful descriptors
- **Negative space:** preserve usable visual breathing room next to the subject, especially for the adjacent right-side text

Use this structure, adapting it to the content:

`[Style] [specific subject performing action] in [setting/background], [lighting], [color palette], [quality modifiers], [composition with clear negative space and visual weight on the left for Image LEFT, text RIGHT layout]`

Prompts should describe visual storytelling, not merely repeat the slide title. Avoid vague adjectives, conflicting style instructions, text inside images, logos, watermarks, and clutter.

## Final Quality Check

Before responding, verify that:

- Every original user instruction is addressed
- The material has clear objectives, meaningful scaffolding, engagement, and appropriate complexity
- Each visual prompt is specific, cohesive, professional-grade, and optimized for AI image generation
- Image-left/text-right direction is stated wherever visual content is included
- The final version represents a genuine 10x improvement
- The response presents only the polished final version, unless the user requested the process
