image-prompt-writer

تایید شده

Craft optimized image generation prompts for tacosdedatos illustrations. Use this skill when you need to create a prompt for Gemini image generation, select an appropriate illustration mode, or help someone write a prompt for the tacosdedatos visual style. Outputs ready-to-use prompts with negative prompts included.

@majiayu000
MIT۱۴۰۴/۱۲/۳
رد شده
(0)
۸۲
۱۴
۳۲

نصب مهارت

مهارت‌ها کدهای شخص ثالث از مخازن عمومی GitHub هستند. SkillHub الگوهای مخرب شناخته‌شده را اسکن می‌کند اما نمی‌تواند امنیت را تضمین کند. قبل از نصب، کد منبع را بررسی کنید.

نصب سراسری (سطح کاربر):

npx skillhub install majiayu000/claude-skill-registry/image-prompt-writer

نصب در پروژه فعلی:

npx skillhub install majiayu000/claude-skill-registry/image-prompt-writer --project

مسیر پیشنهادی: ~/.claude/skills/image-prompt-writer/

بررسی هوش مصنوعی

رد شده
این مهارت معیارهای کیفیت را ندارد

Rejected because the skill is exclusively designed for one blog's visual identity system and requires an internal reference file that's not included. Not a contributing bundle candidate (not a major OSS project).

محتوای SKILL.md

---
name: image-prompt-writer
description: Craft optimized image generation prompts for tacosdedatos illustrations. Use this skill when you need to create a prompt for Gemini image generation, select an appropriate illustration mode, or help someone write a prompt for the tacosdedatos visual style. Outputs ready-to-use prompts with negative prompts included.
---

# Image Prompt Writer

Craft optimized prompts for tacosdedatos illustrations using the 7 defined style modes.

**Required reference**: Read `agents/shared/illustration-style-guide.md` first for color palette, mode templates, and negative prompts.

---

## The Process

### Step 1: Understand the Content

Gather these inputs:
- **Post title**: What's the article called?
- **Summary**: 1-2 sentence description of the content
- **Tone**: Technical, reflective, playful, personal?
- **Key concept**: What's the ONE main idea to visualize?

### Step 2: Select the Mode

Match content type to illustration mode:

| Content Type | Recommended Mode |
|--------------|------------------|
| Technical tutorial, system architecture | **Abstract/Geometric** |
| Evolution, before/after, layered concepts | **Paper-Cut/Layered** |
| Reflective, philosophical, time/change | **Surrealist/Evocative** |
| Growth narratives, learning journeys | **Nature + Tech Fusion** |
| Personal stories, "lo que ando haciendo" | **Atmospheric/Cinematic** |
| How-tos, team content, energetic pieces | **Playful Cartoon** |
| Data analysis, analytics, visualization | **Data Dashboard** |

### Step 3: Identify the Visual Metaphor

The illustration should represent the concept visually, not literally.

**Good metaphors** (from published examples):
- AI processing → geometric shapes flowing into brain network
- Time/timestamps → melting clock with butterflies
- Building something → puzzle pieces forming a path
- Data cultivation → garden with floating data cards
- Orchestrating AI → conductor's baton with flowing lines

**Bad approaches**:
- Generic person at laptop
- Literal representation of the topic
- Stock illustration style

### Step 4: Build the Prompt

Use the template for your selected mode from `illustration-style-guide.md`.

**Template structure**:
```
[Style specification]
[Background color]
[Subject/metaphor description]
[Color specifications with hex codes]
[Composition notes]
[Texture/style notes]
[Negative constraints]
```

### Step 5: Add Negative Prompts

Always append these exclusions:

**Universal** (always include):
```
no text, no words, no labels, no watermarks, no signatures, no logos
```

**Style protection**:
```
no photorealism, no 3D render, no harsh gradients, no glossy surfaces
no stock illustration style, no corporate clip art
```

**Quality**:
```
no blur, no artifacts, no low resolution
```

---

## Output Format

When delivering a prompt, provide:

```markdown
## Generated Prompt

**Mode**: [Selected mode]
**Rationale**: [Why this mode fits the content]

### Prompt
[Complete prompt text]

### Negative Prompt
[All negative constraints]

### Generation Settings
- Aspect ratio: 16:9 or 3:2 (landscape)
- Resolution: 2K recommended
```

---

## Examples

### Example 1: Technical Post

**Input**: "Subagentes de Claude Code: Primeras impresiones del día cero"

**Output**:
- **Mode**: Abstract/Geometric
- **Metaphor**: Terminal window branching into multiple output streams

```
Flat digital illustration on very dark background (#121212).
A stylized terminal window on the left, with three colored lines (red, teal, cream) branching out to the right into separate output streams with icons.
Colors: Crema Nevada (#F1FAEE) for the terminal outline, Rojo Fuego (#E63946), Azul Niebla (#A8DADC), and Naranja Cempasúchil (#F4A261) for the branching lines.
Style: clean geometric, subtle grain texture, generous negative space, minimal elements.
No text, no words, no labels, no photorealism, no 3D render, no gradients.
```

### Example 2: Personal Reflection

**Input**: "Desde el precipicio" - reflective piece about a turning point

**Output**:
- **Mode**: Surrealist/Evocative
- **Metaphor**: Melting clock representing time at a crossroads

```
Surrealist digital illustration, dreamlike atmosphere.
A melting clock (Dalí-style) as the central element, with butterflies and small birds emerging from its dripping form.
Dark background (#121212) with Naranja Cempasúchil (#F4A261) and Azul Niebla (#A8DADC) for the butterflies.
Crema Nevada (#F1FAEE) for the clock face.
Style: flowing organic forms, emotional resonance, subtle grainy texture.
No text, no photorealism, no harsh edges.
```

---

## Troubleshooting Prompts

| Issue | Solution |
|-------|----------|
| Image too literal | Focus on metaphor, not direct representation |
| Wrong colors | Include hex codes explicitly in prompt |
| Too cluttered | Reduce elements, emphasize "generous negative space" |
| Generic style | Add "subtle grain texture" and specific style terms |
| Text appearing | Add "no text, no words, no labels" to negatives |