Skip to content

@remotion/skills: Add video layout guidance#7386

Open
samohovets wants to merge 4 commits into
mainfrom
igor/layout-typography-skill
Open

@remotion/skills: Add video layout guidance#7386
samohovets wants to merge 4 commits into
mainfrom
igor/layout-typography-skill

Conversation

@samohovets
Copy link
Copy Markdown
Member

Summary

  • Add a Remotion skill rule for video-first layout, readable text sizing, safe areas, and preventing overlaps by construction.
  • Link the rule from the main Remotion skill so agents load it when designing visual scenes, promos, motion graphics, or text-heavy videos.
  • Add a simple vertical promo scenario to skills evals for testing text-heavy motion graphics output.

Closes #7256

Test plan

  • Ran bunx oxfmt src --write for affected packages where applicable.
  • Ran bunx oxfmt scenarios.ts --write for the skills eval scenario file.
  • Ran diagnostics for the edited skill files.

@vercel
Copy link
Copy Markdown
Contributor

vercel Bot commented May 14, 2026

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
bugs Ready Ready Preview, Comment May 15, 2026 10:28am
remotion Ready Ready Preview, Comment May 15, 2026 10:28am

Request Review

Copy link
Copy Markdown
Contributor

@pullfrog pullfrog Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No new issues found.

TL;DR — Adds a video-layout.md skill rule that codifies video-first layout, safe areas, text sizing, and overlap-prevention guidance, links it from the main Remotion skill, and seeds a vertical-promo eval scenario to exercise it.

Key changes

  • New video-layout rule — Adds guidance on framing, default scene structure, layout containers, text minimums, and a pre-render checklist.
  • Wire-up from main SKILL.md — Inserts a one-line pointer under ## Designing a video so agents load the rule when composing visual scenes.
  • Vertical-promo eval scenario — Replaces the empty scenarios array with a vertical-promo-text-motion scenario targeting text-heavy motion graphics.

Summary | 3 files | 2 commits | base: mainigor/layout-typography-skill


Video-first layout guidance

Before: No prescriptive guidance on safe areas, default scene structure, or text sizing; agents tended to produce dashboard-like, dense frames with tiny text.
After: A dedicated rule with concrete pixel minimums for 1080px-wide compositions (84/44/32px), a "prevent overlap by construction" section steering away from manual absolute positioning, and a pre-render check list.

The rule is loaded on-demand via the pointer in SKILL.md, matching the existing pattern for sibling rules like text-animations.md and subtitles.md.

SKILL.md · rules/video-layout.md


First skills-eval scenario

Before: scenarios was an empty array — no eval coverage existed for the skill system.
After: A vertical-promo-text-motion scenario drives a 10-second FlowPilot promo with six required phrases, giving the new layout rule a concrete artifact to compare against per the plan in #7256.

scenarios.ts

Pullfrog  | View workflow run | Using Claude Opus𝕏

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Skills: How do we prevent the AI from making the font sizes / proportions too small?

1 participant