Back to Blog

design-creative

Remake Poster: Turn Any Flattened Shot into a Live Design Board

Learn how the Remake Poster skill converts competitor screenshots into editable tldraw design boards, and when to trigger it.

OpenAgentSkillPublished:

Where this fits

You're building an agent that helps designers or marketers iterate on ad creatives. A user drops a screenshot of a competitor's poster or a UI shot and says "make me something like this." Without a structured approach, your agent either returns a static image or tries to rebuild the layout in HTML. This skill fills that gap: it converts a flattened raster into a Paw Work Design (tldraw) board with real, independent text boxes and cropped photo regions—so the user can click, edit, and reuse the design as a starting point.

Why agents benefit

  • Precise triggering logic – The skill's description explicitly tells the agent when not to fire: skip new original posters (no flatten) and fused PNGs (compose-image). This prevents false positives and wasted tool calls.
  • Source-pixel fidelity – The agent inspects the screenshot with inspect view=item, captures the exact pixel dimensions, and creates text boxes in source coordinates. No guesswork about scale or placement.
  • Structured decomposition – Instead of a single full-bleed image, the skill guides the agent to run fromRaster with scan: "auto" and kind=poster, cutting the flatten into color planes and photo regions. The agent supplies the words; the host does the heavy lifting.
  • Reuses a proven compile pipeline – It loads the visual-compile skill and follows the poster mutation path, so your agent inherits battle-tested layout and theme handling.
  • Clear failure mode – The README warns that a "vertical stack of black text on white paper is a failed remake." This gives your agent a quality check to self-correct.

Practical scenarios

Rebuilding a competitor's ad campaign

A user shares a carousel ad from a rival brand. The agent inspects each frame, extracts headline and CTA copy, and produces a tldraw board where the user can swap the photo, tweak the offer, and export a new version—all without touching a pixel editor.

Updating a UI mockup from a screenshot

A product manager sends a screenshot of a new app screen. The agent converts it into a design board with editable text labels and cropped UI elements. The PM can then adjust wording or colors directly, rather than asking a designer to recreate it from scratch.

Turning a static poster into a reusable template

A marketing ops user uploads a flattened event poster. The agent breaks it into layers: title, date, venue, and a photo block. The user can now reuse the template for future events by editing the text boxes—no need to start from zero.

Add it to your agent workflow

Install the skill with:

npx skills add Player-YN/PawWork_ZhuaZhua --skill "Remake Poster"

Then, when a user provides an @截图 or @图片 of a poster, ad, or UI shot, your agent should:

  1. Run inspect view=item to read copy and dimensions.
  2. Load visual-compile and follow the poster mutate flow.
  3. Execute fromRaster with scan: "auto" and kind=poster, then place the text nodes.

The result is a click-editable Paw Work Design board, not a flat image.

Compare before adopting

Before committing, weigh this skill against alternatives:

  • Quality signals – Test it on a few flattened posters with dense text and complex backgrounds. Does it preserve copy accurately and keep photo regions clean?
  • Maintenance freshness – Check the GitHub repo (Player-YN/PawWork_ZhuaZhua) for recent commits and open issues. A skill that tracks tldraw API changes is more reliable.
  • Alternatives – Compare with generic compose-image or HTML-based approaches. If your users only need static outputs, those might be lighter.
  • Workflow fit – Does your agent already use Paw Work Design? If not, the output format may not integrate with your existing tooling.

Why it is worth tracking

With 1,270 stars and an active author, this skill shows real community traction. The careful trigger logic and explicit failure modes indicate a mature, production-ready design—rare among agent skills. It's worth evaluating when your agent needs to turn visual references into editable assets, and you can track its evolution on the OpenAgentSkill listing to stay current with updates and community feedback.

Featured Skill

Remake Poster

Skill by Player-YN