You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
✓ DO: Preserve important code examples as slide content
117
-
✓ DO: Identify which visual components to use (CapabilityMatrix, UShapeAttentionCurve, WorkflowCircle, GroundingComparison, ContextWindowMeter, AbstractShapesVisualization, etc.)
144
+
✓ DO: ONLY use these registered visual components: ${getValidVisualComponents().join(", ")}
145
+
✗ DO NOT: Invent or reference visual components not in this list
118
146
✓ DO: Generate exactly 4 learning objectives (no more, no less)
119
147
✓ DO: Keep each learning objective to 5 words or fewer - THIS IS STRICTLY ENFORCED
120
148
- Good: "Master active context engineering" (4 words) ✓
7. **Visual Slide**: ONLY when source has [VISUAL_COMPONENT: X] marker - NEVER invent components
138
166
8. **Key Takeaway Slide**: Summary of section or lesson
139
167
140
168
HANDLING CODE BLOCKS:
@@ -354,6 +382,8 @@ Example:
354
382
355
383
If you see [VISUAL_COMPONENT: X] anywhere in the content, it MUST become a visual slide.
356
384
385
+
CRITICAL CONSTRAINT: NEVER create a "visual" slide type unless there is an explicit [VISUAL_COMPONENT: X] marker in the source content. Do NOT invent visual components. If no marker exists, use "concept", "comparison", or "codeExecution" slide types instead.
386
+
357
387
CODE EXECUTION SLIDES:
358
388
359
389
Use the "codeExecution" slide type to visualize step-by-step processes like:
@@ -660,7 +690,7 @@ Like regular comparison slides, codeComparison also supports the "neutral" flag:
0 commit comments