What problem does it solve? Cola agents lack a persistent visual identity and a way to express emotions in chat. This Skill generates a consistent pixel-art character, a profile card, animated expression GIFs, and meme stickers, then wires them into the agent's behavior so it can react with stickers during conversations. ## Core Features & Use Cases - Character Generation: Builds a deterministic pixel-art persona from the agent's name, birthday (five-elements color palette), personality, and a hash-based rarity tier (common/rare/legendary species pools), then renders it via the listenhub image API. - Animated Sticker Pack: Produces 4 expression GIFs (happy, sad, angry, thinking) and 3 meme stickers (confused, annoyed, cracked) in 128px and 256px sizes, with background removal, watermarking, and a degraded static-PNG fallback when Python/Pillow is unavailable. - Behavior Injection: Writes sticker-usage rules into the agent's AGENT.md so the agent proactively sends expressions at emotional moments, and displays the avatar on wake-up. - Use Case: A user says "画一个自画像" — the Skill generates the character, shows a profile card for confirmation, then produces and persists the full sticker pack under ~/.cola/avatar/. ## Quick Start Ask the agent to draw a self-portrait or generate an emoji sticker pack, for example by saying "generate your avatar and expression stickers".
