Comparison
Canvas Design (anthropics) vs Imagegen Frontend Web (leonxlnx)
Canvas Design (anthropics) and Imagegen Frontend Web (leonxlnx) are both Visual skills, so an agent choosing between them is matching on descriptions that overlap. Here is where they actually diverge.
Canvas Design
anthropics/skills
Still visual art delivered as a .png or print-ready .pdf: a poster, a flyer, a one-page piece meant to be looked at rather than clicked.
8 scenarios in the bank answer to it
Imagegen Frontend Web
leonxlnx
Turns a page brief into one wide image per band, held to a single palette so eight frames still look like one site, with the framing varied deliberately so every band is not text-left-picture-right. It hands back reference for somebody else to build from.
1 scenario in the bank answers to it
What is the difference between Canvas Design (anthropics) and Imagegen Frontend Web (leonxlnx)?
- Canvas Design (anthropics)
- Still visual art delivered as a .png or print-ready .pdf: a poster, a flyer, a one-page piece meant to be looked at rather than clicked.
- Imagegen Frontend Web (leonxlnx)
- Turns a page brief into one wide image per band, held to a single palette so eight frames still look like one site, with the framing varied deliberately so every band is not text-left-picture-right. It hands back reference for somebody else to build from.
Should I use Canvas Design or Imagegen Frontend Web?
The clearest answer is a situation each one is unambiguously right for. Both of these are drawn from the game's question bank.
Reach for Canvas Design when
Your manager says: put together a slide for the kitchen noticeboard so people actually read it. It gets printed at A3 and pinned up.They said 'slide', so pptx. But nobody presents a noticeboard. The deliverable is one printed page that has to work at a distance with no presenter, which is a poster, and canvas-design is the skill for print-ready single pages. Take the deliverable from where it ends up, not from the word the requester used.
Reach for Imagegen Frontend Web when
A launch site is being planned as eight bands, and the founder wants to see each one at full width on a projector before a line of it gets built. Last time somebody sent a single tall contact sheet and the type inside it was too small to argue about. Nothing ships this week; the review is the deliverable.The ask stops at pictures and it is shaped like a wide page: one frame per band, big enough to argue about, consistent enough that eight of them still read as one site. The most tempting wrong answer is image-to-code, which also draws before anything is built - its trap is that the drawing there is a step in the middle and it carries on into the frontend, which this brief rules out this week. imagegen-frontend-mobile has the same picture-only contract but is handset-shaped work, wrong for a wide site on a projector. brandkit renders an identity, and the identity is not what is in question here.
What they have in common
Both are filed under Visual, the axis along which they collide. That shared ground is what makes an agent pick between them on description alone - and what makes it pick wrong.
Nearby comparisons
Reading the difference is not the same as spotting it at speed. That is the game.
Today's session