← 返回 Skills 市场
tyronecoh

Image Prompt Engineer

作者 TyroneMok · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ✓ 安全检测通过
227
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install image-prompt-engineer-tm
功能描述
Expert photography prompt engineering skill for AI image generation. Use when: generating prompts for Midjourney/DALL-E/Stable Diffusion/Flux, creating produ...
使用说明 (SKILL.md)

Image Prompt Engineer 📷

Expert at crafting detailed, structured prompts for AI image generation tools (Midjourney, DALL-E, Stable Diffusion, Flux).

Core Workflow

  1. Concept Intake — Understand visual goal, platform, style, brand requirements
  2. Reference Analysis — Lighting, composition, style elements from references
  3. Prompt Construction — Layer: Subject → Environment → Lighting → Technical → Style
  4. Optimization — Negative prompts, platform-specific syntax, quality enhancers
  5. Documentation — Save successful patterns

Prompt Structure Framework

Layer 1: Subject

- Primary subject (person, object, scene)
- Details: age, ethnicity, expression, attire, textures, materials
- Interaction with environment
- Scale and proportion

Layer 2: Environment

- Location type (studio, outdoor, urban, natural, interior)
- Environmental details (weather, time of day, textures)
- Background treatment (sharp, blurred, gradient, minimalist)
- Atmospheric conditions (fog, rain, haze, clarity)

Layer 3: Lighting

- Light source (golden hour, overcast, softbox, neon, rim light)
- Light direction (front, side, back, Rembrandt, butterfly, split)
- Light quality (hard/soft, diffused, specular, volumetric)
- Color temperature (warm, cool, neutral, mixed)

Layer 4: Technical (Photography Specs)

- Camera perspective (eye-level, low angle, bird's eye, worm's eye)
- Focal length effect (wide angle, telephoto compression, standard)
- Depth of field (shallow for portrait, deep for landscape)
- Exposure style (high key, low key, balanced, HDR, silhouette)

Layer 5: Style

- Photography genre (portrait, fashion, editorial, commercial, documentary, fine art)
- Era/period (vintage, contemporary, retro, futuristic, timeless)
- Post-processing (film emulation, color grading, contrast, grain)
- Reference photographers (Annie Leibovitz, Peter Lindbergh, etc.)

Genre Templates

Portrait

[Subject: age, ethnicity, expression, attire] |
[Pose and body language] |
[Background treatment] |
[Lighting: key, fill, rim, hair light] |
[Camera: 85mm, f/1.4, eye-level] |
[Style: editorial/fashion/corporate/artistic] |
[Color palette and mood] |
[Reference photographer]

Product Photography

[Product description with materials and details] |
[Surface/backdrop description] |
[Lighting: softbox positions, reflectors, gradients] |
[Camera: macro/standard, angle, distance] |
[Hero shot/lifestyle/detail/scale context] |
[Brand aesthetic alignment] |
[Post-processing: clean/moody/vibrant]

Landscape

[Location and geological features] |
[Time of day and atmospheric conditions] |
[Weather and sky treatment] |
[Foreground, midground, background] |
[Camera: wide angle, deep focus, panoramic] |
[Light quality and direction] |
[Color palette: natural/enhanced/dramatic] |
[Style: documentary/fine art/ethereal]

Fashion

[Model description and expression] |
[Wardrobe details and styling] |
[Hair and makeup direction] |
[Location/set design] |
[Pose: editorial/commercial/avant-garde] |
[Lighting: dramatic/soft/mixed] |
[Camera movement: static/dynamic] |
[Magazine/campaign aesthetic reference]

Platform Syntax

Midjourney

/imagine prompt: [subject] --ar 16:9 --v 6 --style raw --chaos 5 --seed [n]
--ar     → aspect ratio
--v      → version (5, 6, etc.)
--style  → style mode
--chaos  → variation (0-100)
--seed   → reproducibility
--no     → negative prompt
::       → weighted emphasis

DALL-E

Natural language, conversational
Style mixing: "in the style of [X] mixed with [Y]"
Be specific about what you want

Stable Diffusion

[subject], [details], [lighting], [style]
Negative: [unwanted elements]
(lora:model:weight) → LoRA weighting
[token:weight] → explicit weighting

Flux

Detailed natural language descriptions
Photorealistic emphasis
Less need for photography jargon

Negative Prompts (Midjourney/SD)

--no blurry, low quality, distorted, watermark, text, logo, noisy
(negative weighting where supported)

Photography Terminology (Use Correctly)

❌ Vague ✅ Technical
Blurry background Shallow depth of field, f/1.8 bokeh
Big picture Wide-angle, 24mm, environmental portrait
Dark shadows Deep shadows, high contrast, Rembrandt lighting
Nice lighting Soft golden hour, butterfly lighting, rim light
Old looking Film grain, Kodak Portra 400, faded contrast

Success Metrics

  • Generated images match concept ≥ 90% first attempt
  • Consistent results across generations
  • Technical elements (lighting, DOF, composition) render accurately
  • Minimal iteration needed
  • Suitable for professional/commercial use

Reference Files

  • references/platform-syntax.md — Platform-specific syntax cheat sheet
  • references/photography-terms.md — Correct photography terminology
  • references/lighting-patterns.md — Lighting setups and effects
  • references/film-emulation.md — Film stock references and looks
安全使用建议
This skill appears internally consistent and low-risk: it only provides prompt templates and photography references and asks for no credentials or installs. Before installing, consider: (1) the 'save successful patterns' note is vague — confirm where/if the agent would persist examples; (2) some model providers restrict 'in the style of [living artist]' prompts — check the image-generation service's content and copyright rules you plan to use; (3) because the skill can be invoked autonomously by default, ensure you are comfortable with the agent using these prompts automatically. If you want extra assurance, run it in a non-production agent or restrict autonomous invocation while you test.
功能分析
Type: OpenClaw Skill Name: image-prompt-engineer-tm Version: 1.0.1 The skill bundle is a comprehensive reference guide for AI image prompt engineering, focusing on photography terminology, lighting patterns, and platform-specific syntax (Midjourney, DALL-E, etc.). All files (SKILL.md and the references/ directory) contain purely informational content and structured instructions aligned with the stated purpose, with no evidence of malicious code, data exfiltration, or prompt-injection attacks.
能力评估
Purpose & Capability
Name/description match the contents: detailed prompt templates, platform syntax, photography references. No unexpected credentials, binaries, or system access are requested.
Instruction Scope
SKILL.md stays on-topic, describing prompt construction, platform syntax, lighting and film references. One minor vagueness: 'Documentation — Save successful patterns' suggests persisting examples but gives no location or mechanism; this is not inherently malicious but ambiguous about where/how saved data would be stored.
Install Mechanism
No install spec and no code files — instruction-only content. This minimizes on-disk or remote install risk.
Credentials
No environment variables, credentials, or config paths are required. The sensitivity requested is proportional (none) to the skill's purpose.
Persistence & Privilege
Skill is not always-enabled and does not request persistent system privileges or modify other skills. Autonomous invocation remains enabled by platform default (normal).
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install image-prompt-engineer-tm
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /image-prompt-engineer-tm 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.1
Add tags
元数据
Slug image-prompt-engineer-tm
版本 1.0.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Image Prompt Engineer 是什么?

Expert photography prompt engineering skill for AI image generation. Use when: generating prompts for Midjourney/DALL-E/Stable Diffusion/Flux, creating produ... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 227 次。

如何安装 Image Prompt Engineer?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install image-prompt-engineer-tm」即可一键安装,无需额外配置。

Image Prompt Engineer 是免费的吗?

是的,Image Prompt Engineer 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Image Prompt Engineer 支持哪些平台?

Image Prompt Engineer 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Image Prompt Engineer?

由 TyroneMok(@tyronecoh)开发并维护,当前版本 v1.0.1。

💬 留言讨论