What is illo-skill?

illo is a skill (a reusable plugin) for AI coding agents that generates editorial-style illustrations from ideas or article text. Instead of generic stock images, it produces original scenes featuring a recurring mascot character, with each image built around a single visual metaphor drawn from the input. The default mascot is Blot, described as a deadpan ink-drop figure, and the project also includes a built-in character builder so you can define your own. The skill ships with ten visual styles: riso (the default, resembling risograph print), blueprint, woodcut, pixel, clay, manila, chalk, phosphor, enamel, and gouache. Each style produces a distinct look, from draftsman linework to stop-motion plasticine to CRT phosphor traces. You can also use named color palettes, custom palettes, and reference images to keep character appearance consistent across a set of illustrations. Image generation runs through two backends. If you have the Codex CLI installed and are logged in as a subscriber, it uses gpt-image-2 at no extra API cost. Otherwise it falls back to OpenRouter, where you can select from several available image-generation models. The README names a few options including Grok Imagine and GPT-based image models. Installation works across a wide range of agent runtimes. A single command via the skills CLI (npx skills add tmchow/illo-skill) covers Claude Code, Cursor, Codex, Copilot, Gemini CLI, and others. Platform-native install paths are also available for each major runtime, using each platform's own plugin or extension system. The repository ships manifests for Claude Code, Codex, Cursor, and Gemini CLI so managed updates are handled through the platform's own tooling. Community-contributed character packs live in a companion repository at tmchow/illo-characters. The project is MIT licensed.