Photo-to-flat-vector-style illustration skill for Codex, Claude Code, and AI agents using semantic compression, fixed HSB colors, and no gradients.
