/apply-comments

/apply-comments closes the visual feedback loop. You drop comments through the inspector; the skill resolves them into code edits.

What it does

1. Scans the workspace for @slide-comment markers.
2. For each marker, reads the surrounding code and the comment text.
3. Edits the file to satisfy the comment.
4. Removes the marker.

› /apply-comments
✓ 3 markers · slides/q2-launch/index.tsx
✓ 1 marker  · slides/intro/index.tsx

Marker semantics

A @slide-comment marker is a normal JSX/JS comment that the inspector inserts next to the element you commented on. It's invisible at runtime — the rendered output never changes. The marker carries:

• The user's note (the message you typed).
• A pointer to the element (so an agent reading the file knows what this meant).

When to use it

After a batch of inspector comments — typically right before saving the deck for the day or right before a rehearsal. Refreshing the dev server after /apply-comments shows the new state with the markers cleared.

If a comment is ambiguous, the agent applies the smallest reasonable interpretation and notes the assumption in its summary. Comments it can't resolve at all are left in place and reported as skipped — refine the wording and re-run.