mofa-infographic

name	mofa-infographic
version	0.5.0
description	AI-generated infographics and visual posters. Triggers: infographic, poster, 信息图, 海报, data poster, visual summary, mofa infographic. Generates multi-section infographic via Gemini with optional Qwen-Edit refinement, stitched into a single tall image.
requires_bins	mofa
requires_env	GEMINI_API_KEY

mofa-infographic

CLI: mofa infographic — generate multi-section infographic posters via Gemini, stitched vertically into a single tall PNG. Optional Qwen-Edit refinement (requires DASHSCOPE_API_KEY).

For LLM-facing discovery and routing, the manifest discovery card declares the hints below; this SKILL.md is the long-form reference loaded on demand.

Output paths (LOAD-BEARING)

ALWAYS use relative paths under skill-output/ with a unique per-request subdirectory:

skill-output/mofa-infographic-<YYYYMMDD-HHMMSS>/poster.png
skill-output/mofa-infographic-<YYYYMMDD-HHMMSS>/sections/

NEVER use absolute paths like /tmp/poster.png. The out argument MUST be a relative path — the tool description repeats this. Absolute paths break the workspace contract and prevent the output from being delivered.

Built-in styles (4)

--style	Best For
cyberpunk-neon (default)	Tech, AI, data
editorial	Reports, articles, longform
clean-light	Business, consulting
multi-panel	Comparisons, multi-topic summaries

Full style table, custom inline styles, and variants (header / normal / footer): see docs/styles.md.

Section JSON (top-level: array of section objects)

Field	Required	Description
prompt	yes	Section content description (data, titles, visual elements)
variant	no	"header" / "normal" / "footer" (auto-detected by position)
refine_prompt	no	Qwen-Edit refinement instruction (needs --refine + DASHSCOPE_API_KEY)

Full schema, prompt writing tips, and 4 worked examples (tech / business / editorial / batch): see docs/examples.md.

Quick reference

• CLI flags, resolution & quality, output structure, mofa/config.json: docs/cli-flags.md.
• API modes (rt vs batch), per-section timing, 10-minute timeout strategy, cached-section recovery: docs/api-modes-and-timing.md.
• Telegram preferences flow + inline-keyboard example for picking style / sections / resolution: docs/interaction.md.

Minimal example

[
  {"prompt": "Header: 'Q3 Review' bold title, gradient background.", "variant": "header"},
  {"prompt": "3 KPI cards: Revenue $12.4M, NPS 72, Churn 2.1%."},
  {"prompt": "Footer: sources and credits.", "variant": "footer"}
]

mofa infographic --style clean-light --out skill-output/review.png -i sections.json

Run notes

• Each section ~15-30s. Keep sections ≤8 for a single call to stay inside the 600s tool timeout.
• --api batch is 50% cheaper but 5-30 min — avoid in octos tool calls.
• Cached sections (section-XX.png >10KB) are reused on rerun; delete to regenerate.