illo

Agent Skills-format package for turning articles, posts, and abstract ideas into original editorial illustrations with recurring mascot characters.

What It Does

illo is a portable agent skill with a local engine, references, assets, and quality checks. It reads an idea, article, or post and renders one of four forms: editorial scene, mini-comic, explainer diagram, or transparent character cutout. The consistent unit is a recurring mascot and character pack, so the tenth image still belongs to the same visual system. Source: illo llms.txt, 2026-06-30

The launch thread matters because it names the cross-harness target directly: Claude Code, Codex, Hermes, and OpenClaw. It also uses OpenRouter image models by default in the author's setup, with Grok Image called out as the best result at launch. Source: X/@trevin, 2026-06-12

The visual artifact says "Give your ideas a face" and frames the skill as a way to turn articles and abstract ideas into original editorial illustrations without stock art. Source: local artifact wiki/assets/x-bookmarks/2065232610335641667/image-01.jpg, reviewed 2026-06-30

Local Skill Import

The full skill package is now imported at skills/engineering/illo/, including:

• SKILL.md with trigger rules, backend selection, workflow, and quality bar;
• scripts/illo.py, a stdlib Python engine;
• scripts/repair-hermes-assets.sh for Hermes binary asset repair;
• bundled default character assets;
• references for backends, models, character packs, composition, cutouts, palettes, prompt recipes, quality checks, and pack sharing.

Version snapshot: upstream tmchow/illo-skill main was imported at commit 2cb52e78b9a7401aee1933b1bcaddcb2d3c47dbc; the skill frontmatter reports version: 0.28.1; the latest observed Git tag is v0.28.1. Source: git ls-remote, 2026-06-30

Stack Fit

Use illo when Kevin explicitly asks for /illo, a recurring mascot system, editorial art for a post, a character cutout, or an explainer in the skill's house style. Do not route generic image requests into it automatically. Generic illustration, product imagery, and UI assets should still use the relevant design/image generation route unless the user wants the illo style.

For the brain, this is a reusable style system skill: it gives agents a repeatable visual voice, which is more valuable than one-off generated images. It belongs beside design/motion skills because it turns taste into a portable workflow with references, assets, checks, and installable backends.

---

Timeline

• 2026-06-30 | Page created and full upstream skill imported into skills/engineering/illo/. Source: tmchow/illo-skill commit 2cb52e78b9a7401aee1933b1bcaddcb2d3c47dbc
• 2026-06-12 | Launch bookmark introduced illo as a recurring-character illustration skill usable from Claude Code, Codex, Hermes, and OpenClaw. Source: X/@trevin, 2026-06-12