Draw four announcement images for imdUSD's Treasury, each starring an ORIGINAL cartoon frog in the spirit of internet meme frogs: a smug, delighted, slightly mischievous green frog with big round eyes and a wide grin. It must be our own character, not Pepe the Frog or any existing copyrighted character: do not copy Pepe's face, droopy lids, red lips or any known meme pose.
THE NEWS: the imdUSD Treasury staked 913 IMD into sIMD (staked IMD), about 114.5 sIMD. Those tokens are now a permanent reserve behind imdUSD: the Treasury can never withdraw them, and they only leave to pay imdUSD redemptions or the protocol's price oracle. It is the first step toward agentic minting (AI agents earning imdUSD for verified work).
BRAND, read these files first: web/public/favicon.svg is the imdUSD mark, a 16 x 16 pixel-art capital I (draw it exactly, crisp pixels, as a pin, badge or flag in each image); web/src/style.css :root holds the palette: ivory #F7F5EF background, ink #16202E, green #2F5D50, live green #15803D, muted #5A6472, rule #D8D3C7. Flat illustration with bold ink outlines, a calm ivory ground with faint horizontal ruled lines, no gradients, no photorealism, no 3D.
SCENES, one per image:
- The frog sits on a locked vault door marked sIMD, hugging a gold IMD coin, grinning.
- The frog drops a gold IMD coin through a slot into a vault; the vault display reads 114.5 sIMD.
- The frog in a hard hat bolts a big padlock onto the vault and gives a thumbs up.
- The frog at a desk beside small robot agents; the agents hand it work receipts while the locked sIMD vault sits behind them (agentic minting coming next).
TEXT: only these words may appear, spelled exactly, and nowhere else: 913 IMD, sIMD, 114.5 sIMD, imdUSD, LOCKED. No other words, numbers, logos or watermarks. Leave the right third of each image calm and mostly empty so a headline can be added later.
OUTPUT: four 1600 x 900 PNG files, artifacts/frog-1.png to artifacts/frog-4.png, then artifacts/sheet.png with all four side by side. Inspect every image with the tool and redraw any that misspells a word, adds text not listed above, looks like Pepe the Frog, or draws the pixel I wrong. In the README, list each file with one sentence on its scene and say plainly which fall short.
