concept-comic-skill
v0.1.0
Published
Keyword-triggerable skill asset package for turning words and concepts into four-panel comic storyboards and image prompts.
Downloads
122
Maintainers
Readme
Concept Comic Skill
把黑话,画成笑话
Turn jargon into deadpan four-panel comics.
Transform AI terms, office jargon, and product concepts into a complete four-panel comic package — including angle options, Chinese storyboard, and English image prompt.
concept → angle → storyboard → prompt
Install
npx skills add SuperKatrina123/concept-comic-skillWhat it does
concept-comic-skill turns a word, phrase, term, or abstract concept into a comic-ready output package.
It is designed for people who want to turn things like:
- AI terms
- office jargon
- product language
- workflow concepts
- short meme premises
into clean, deadpan, social-media-friendly four-panel comics.
By default, the output is optimized for:
- Chinese
Idea BreakdownandStoryboard - English image prompts
- horizontal
4:3framing - clean
2x2four-panel layout - large readable title above the comic
- literal / satire / reflective mode control
- deadpan webcomic-style humor
What you get
Given one raw idea, the skill returns a structured package with:
- mode decision
- 3 angle candidates
- idea breakdown
- 4-panel Chinese storyboard
- English image prompt in a fixed production template
- anti-drift guardrail prompt
- recovery prompt for rerolling bad generations
- optional social copy
This makes the output useful both for human ideation and for downstream image-generation workflows.
Good for concepts like
- Human in the Loop
- Guardrails
- Context Window
- Tool Calling
- Chain of Thought
- Open Office
- Highest Priority
- Model Distillation
Quick example
Use $concept-comic-skill to turn "Human in the Loop" into a deadpan four-panel comic.用 $concept-comic-skill 把 “Human in the Loop” 做成四格小漫画。How It Thinks
The skill follows a fixed pipeline instead of dumping one loose answer:
- classify the idea and pick the best mode
- generate 3 candidate angles in that mode
- lock the strongest comic premise
- turn it into a 4-panel storyboard
- generate an English image prompt
- add guardrails and recovery prompts
It is especially strict about two things:
- not letting literal ideas drift into abstract reflection
- making sure Panel 4 is stronger than Panel 3
It also defaults to:
- no speaker-name prefixes in dialogue
- a meaningful Panel 2 that actually advances the joke
- a cleaner one-line payoff in Panel 4 when possible
Package Contents
This package includes:
SKILL.mdagents/openai.yamlreferences/schema.mdreferences/workflow.mdreferences/prompt-templates.mdreferences/guardrails.mdreferences/examples.md
The references include built-in cases such as:
Highest PriorityOpen OfficeHuman in the LoopModel Distillationguardrails
Runtime Notes
This is an asset package, not a CLI.
It does not auto-install itself into a host runtime. It is intended for runtimes that:
- load skill assets from installed npm packages
- load skills from explicit package paths
- support skill discovery by metadata and keyword matching
Exported asset paths:
concept-comic-skill/SKILL.mdconcept-comic-skill/agents/openai.yamlconcept-comic-skill/references/schema.mdconcept-comic-skill/references/workflow.mdconcept-comic-skill/references/prompt-templates.mdconcept-comic-skill/references/guardrails.mdconcept-comic-skill/references/examples.md
License
MIT
