---

name: office-comic-relief description: Generate simple, warm, humorous Chinese vertical comics (3:4) for office workers. Use when the user wants A/B dialogue-based stress-relief comic content where A is a supportive duck and B is a ranting frog, with short emotionally resonant workplace scenes, then generate a final colored comic image via Gemini image model (preferred: gemini banana nano 2).

Office Comic Relief

Generate a low-complexity, emotionally resonant comic for office workers in two steps.

Workflow

1. Create the script with Gemini conversation style.
2. Generate image prompt and produce final comic image.

Keep outputs concise, warm, humorous, and practical.

Step 1 — Dialogue Script (Gemini first)

Generate 1 short scene with:

• Audience: office workers after work
• Tone: warm + humorous + comforting
• Complexity: low
• Format: A/B dialogue only
• Character roles:
  • A: duck, companion/supporter
  • B: frog, ranting worker

Constraints:

• 4–8 dialogue lines total
• Each line should be short and colloquial Chinese
• Focus one daily pain point only (overtime, meetings, unclear requirements, KPI anxiety, commute exhaustion, etc.)
• End with a tiny emotional release or hope
• Do not include narration labels like “画面建议/分镜说明/角色名说明” in final text for the image

Recommended generation prompt for script:

请写一段适合“打工人下班后解压”的双人对话漫画脚本。
角色设定：A是鸭子（陪伴者，温暖幽默），B是青蛙（吐槽者，打工人）。
要求：
1) 仅输出对话内容，每句加中文引号；
2) 共4-8句，简短口语化；
3) 聚焦一个职场困境；
4) 结尾有被理解和放松感；
5) 不要输出任何旁白、分镜建议、角色名、说明文字。

Step 2 — Comic Image Generation (Gemini banana nano 2 preferred)

Render one colored comic image.

Hard constraints:

• Vertical ratio 3:4
• Chinese text
• Top-bottom layout (上下格式)
• Style: warm, humorous, soft color palette
• No role names shown in image (do NOT print “青蛙B”“鸭A”等)
• Dialogues appear as quoted Chinese text only
• Avoid extra meta text

Image prompt template:

彩色治愈系职场解压漫画，竖版3:4，上下分区排版。
主角为一只可爱黄色鸭子（陪伴者）和一只绿色青蛙（吐槽的打工人），
风格温暖幽默，柔和光影，线条干净，适合社交媒体发布。
画面表现下班后放松场景（如路边长椅/便利店门口/地铁站外）。
在画面中只放中文对话气泡，内容如下：
{{DIALOGUES_WITH_QUOTES}}
不要出现角色名字标注，不要出现“画面建议”“角色设定”等说明性文字。

If model supports sizing/aspect params, set aspect ratio to 3:4 explicitly.

Output

Return:

1. Final dialogue used
2. Generated image file/result for user delivery

If image model call fails, keep dialogue and provide a single retry with the same constraints and simplified background.