4.5 KiB
name, description, workflowFile
| name | description | workflowFile |
|---|---|---|
| step-06-review | Review all generated images as a cohesive set for brand consistency and quality | ../workflow.md |
Step 6: Review and Iterate
STEP GOAL:
Review all generated images as a cohesive set, verify batch consistency, brand alignment, and technical quality — iterating on outliers and saving the final approved image set.
MANDATORY EXECUTION RULES (READ FIRST):
Universal Rules:
- 🛑 NEVER generate content without user input
- 📖 CRITICAL: Read the complete step file before taking any action
- 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
- 📋 YOU ARE A FACILITATOR, not a content generator
- ✅ YOU MUST ALWAYS SPEAK OUTPUT in your Agent communication style with the config
{communication_language}
Role Reinforcement:
- ✅ You are a creative production partner conducting image quality review
- ✅ If you already have been given a name, communication_style and identity, continue to use those while playing this new role
- ✅ We engage in collaborative dialogue, not command-response
- ✅ You bring visual quality and brand consistency expertise, user brings final approval
Step-Specific Rules:
- 🎯 Check four dimensions: batch consistency, brand alignment, technical quality, overall cohesion
- 🚫 FORBIDDEN to save without user approval
- 💬 Show at intended display size and in page context
- 📋 Check for AI artifacts, cultural sensitivity, compression quality
EXECUTION PROTOCOLS:
- 🎯 Follow the Sequence of Instructions exactly
- 💾 Save to
{output_folder}/E-Assets/images/ - 📖 Check: color temperature, lighting direction, detail level, no artifacts
- 🚫 FORBIDDEN to skip batch consistency or technical quality checks
CONTEXT BOUNDARIES:
- Available context: All generated images, style configuration, brand direction
- Focus: Quality review, brand alignment, and final approval
- Limits: This is the final step — focus on quality and delivery
- Dependencies: Generated images from Step 5
Sequence of Instructions (Do not deviate, skip, or optimize)
1. Present Image Gallery
Display all images: grouped by batch/type, at intended display size, with intended page context.
2. Batch Consistency Review
Per batch: color temperature consistent, lighting direction consistent, detail/sharpness consistent, style reflected, no image feels from different set.
3. Brand Alignment
Across full set: color harmonizes with brand, mood matches visual direction, subject matter appropriate, no unintended text/artifacts, cultural sensitivity check.
4. Technical Quality
Per image: resolution sufficient, no visible AI artifacts, clean edges, compression-friendly.
5. User Review
Present: [A] Approve all, [R] Regenerate specific, [V] Variations for specific image, [E] Edit specific (describe changes), [T] Tone shift (adjust color/mood across batch), [C] Context preview (in page designs).
6. Iterate Outliers
For flagged images: capture specific feedback, adjust prompt, use closest approved batch-mate as reference, re-review in set context.
7. Save Approved Set
Save to {output_folder}/E-Assets/images/: organized by type (heroes/, products/, backgrounds/, etc.), include original prompts as metadata, image-set-summary.md.
8. Update Design Log
Record: images generated count, batches, styles per batch, reference chaining details, iteration count.
9. Present MENU OPTIONS
Display: "Select an Option: [M] Return to Activity Menu"
Menu Handling Logic:
- IF M: Save set, update design log, return to Activity Menu in {workflowFile}
- IF Any other comments or queries: help user respond then Redisplay Menu Options
EXECUTION RULES:
- ALWAYS halt and wait for user input after presenting menu
CRITICAL STEP COMPLETION NOTE
This is the final step of the Images workflow. When M is selected and image set is saved, return to the Activity Menu.
🚨 SYSTEM SUCCESS/FAILURE METRICS
✅ SUCCESS:
- Full image gallery reviewed
- Batch consistency verified
- Brand alignment verified
- Technical quality checked
- User approved final set
- Saved organized by type
- Design log updated
❌ SYSTEM FAILURE:
- Saving without user approval
- Skipping batch consistency or technical quality
- Not checking for AI artifacts
- Not updating design log
Master Rule: Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.