[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M790499R9CVJTSEM8THHHDZR":3,"skill-same:01M790499R9CVJTSEM8THHHDZR":62},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":31,"ownerHtmlUrl":32,"ownerBlogUrl":33,"ownerDescription":34,"updatedAt":35,"install":36,"resources":49},"01M790499R9CVJTSEM8THHHDZR","image-prompts",null,"Write AI image prompts covering subject, lighting and composition, countering the models' bias toward stereotyped depictions. Use when authoring Midjourney, Flux, or DALL-E AI image prompts.","poorvith-mp\u002Fskills-design","poorvith-mp","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F277478855?v=4","01M78Z5K0YH1Z3PX815PGQ96ND","Poorvith M P","01M79023NZMHA7EKM9K6KWHEET",0,"MIT","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fimage-prompts\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: image-prompts\nlast_reviewed: 2026-09-06\ngroup: Implementation\ndescription: >-\n  Write AI image prompts covering subject, lighting and composition, countering the models' bias\n  toward stereotyped depictions. Use when authoring Midjourney, Flux, or DALL-E AI image prompts.\n---\n\n# image-prompts\n\n## Core Philosophy\nAuthoring AI image generation prompts (Midjourney, Flux.1, DALL-E 3, Stable Diffusion) is not typing random adjectives like \"hyperrealistic, 8k, award-winning trending on artstation\" into a text box. Generative image models respond to precise photographic parameters, lighting physics, historical art references, spatial composition rules, and deliberate debiasing techniques. High-craft image prompts construct an exact visual scene while actively counteracting the models' algorithmic drift toward plastic, oversaturated stereotypes.\n\n---\n\n## 4-Step Generative Image Prompt Architecture\n\n### Step 1: The 5-Part Architectural Prompt Anatomy\n1. **The Structural Prompt Formula**:\n   - `[Subject & Action]` + `[Environment & Context]` + `[Composition & Framing]` + `[Lighting & Atmosphere]` + `[Camera, Lens & Film Stock \u002F Style Engine]`\n2. **Component Breakdown**:\n   - *Subject*: Specific, concrete details (clothing texture, age, posture, expression, action).\n   - *Environment*: Exact time of day, architectural materials (brutalist raw concrete, wet asphalt, mahogany paneling).\n   - *Composition*: Rule of thirds, low-angle hero shot, Dutch tilt, wide-angle cinematic perspective.\n   - *Lighting*: Volumetric god rays, directional Rembrandt lighting, diffuse cloudy softbox, neon back-rim lighting.\n   - *Camera \u002F Medium*: 35mm photograph, Leica M11, f\u002F1.8 aperture, shallow depth of field, Kodak Portra 400 film grain.\n\n### Step 2: Eliminating Generative Clichés & Plastic Skin\n1. **The AI Slop Kill-List for Prompts**:\n   - Ban empty fluff words: `photorealistic`, `hyperrealistic`, `ultra-detailed`, `4K`, `8K`, `masterpiece`, `trending on artstation`. These words pollute the latent space with low-quality amateur renders.\n2. **Enforcing Natural Texture & Imperfection**:\n   - Specify real-world micro-textures to kill plastic rendering:\n     - Skin: *\"Subtle skin pores, natural skin texture, unretouched, faint freckles, uneven lighting.\"*\n     - Environment: *\"Dust motes in light beams, worn edges on leather notebook, subtle motion blur.\"*\n\n### Step 3: Countering Stereotypical Demographic Drift\n1. **Active Debiasing**:\n   - Image models default to severe cultural and gender stereotypes (e.g. prompt \"software engineer\" produces exclusively young white or Asian men; \"nurse\" produces young women).\n   - Explicitly specify diverse, realistic demographic descriptors, ages (e.g. \"50-year-old female systems architect\"), clothing (realistic workwear, not futuristic spandex), and natural working environments.\n\n### Step 4: Technical Parameters & Engine Flags\n1. **Platform-Specific Flag Tuning**:\n   - *Midjourney v6*:\n     - Aspect Ratios: `--ar 16:9` (landscape\u002Fhero), `--ar 1:1` (avatar), `--ar 4:5` (social portrait).\n     - Stylize: `--s 50` to `--s 150` (restrained realism; default 250 is often overly cartoonish).\n     - Weird & Chaos: Keep `--c 0` to `--c 10` for reproducible commercial production.\n   - *Flux.1 (Dev \u002F Schnell)*:\n     - Responds best to natural language paragraphs over comma-separated tag soup.\n\n---\n\n## Deliverable Format: AI Image Prompt Specification (`IMAGE-PROMPTS.md`)\n\n```markdown\n# AI Image Generation Prompt Book: [Campaign \u002F Feature]\n\n## 1. Asset: Hero Visual for Developer Documentation\n- **Target Model**: Flux.1 Dev \u002F Midjourney v6\n- **Aspect Ratio**: 16:9 (`--ar 16:9`)\n- **Visual Genre**: Documentary editorial photograph\n\n### The Master Prompt\n> An authentic 35mm documentary photograph of a 42-year-old female lead infrastructure engineer debugging code late at night in a server room. She wears a navy wool sweater and glasses with thin titanium frames. Subtle facial expression of intense concentration. The background features server racks with out-of-focus amber and soft white blinking LED indicator lights. Soft rim lighting from terminal screens, moody atmospheric shadows, shallow depth of field, f\u002F2.0 aperture. Shot on Leica M6 with Kodak Portra 400 film grain, natural skin pores and subtle shadows, zero plastic smoothing, zero neon sci-fi holograms. --ar 16:9 --s 100 --v 6.0\n\n## 2. Negative Constraints & Exclusions\n- No glowing holographic HUDs floating in air.\n- No purple or magenta cyberpunk neon lighting.\n- No airbrushed smooth plastic mannequin skin.\n```\n\n---\n\n## Worked Example: Overhauling a Generic Tech Hero Visual\n\n- **Original Prompt**: \"Software engineer working on futuristic AI, hyperrealistic, 8k, glowing screens, masterpiece.\" -> *Produced a cartoonish neon cyborg.*\n- **Refined Craft Prompt**: Specified natural documentary photography, natural daylight from a warehouse window, real mechanical keyboard, and Kodak film grain.\n- **Result**: Photorealistic, credible editorial image used as the hero graphic for an enterprise Series B announcement.\n\n---\n\n## Verification Checklist\n\n- [ ] Prompt follows the 5-part anatomical structure (Subject, Context, Framing, Lighting, Camera).\n- [ ] Banned fluff words (\"hyperrealistic\", \"8k\", \"trending on artstation\") are 100% eliminated.\n- [ ] Natural skin texture, grain, and physical imperfections are explicitly instructed.\n- [ ] Stereotypical algorithmic drift is countered with specific demographic descriptors.\n- [ ] Technical flags (`--ar`, `--s`) match target generation engine.\n\n---\n\n## Anti-Patterns\n\n- **Tag Soup Prompting**: Stacking 40 disconnected keywords with commas hoping the model figures it out.\n- **Cyberpunk Defaulting**: Letting models inject purple neon glows and floating binary code into serious enterprise visuals.\n- **Neglecting Aspect Ratios**: Generating square images for a 16:9 website hero banner and cropping off critical subjects.\n","# image-prompts\n\n## Core Philosophy\nAuthoring AI image generation prompts (Midjourney, Flux.1, DALL-E 3, Stable Diffusion) is not typing random adjectives like \"hyperrealistic, 8k, award-winning trending on artstation\" into a text box. Generative image models respond to precise photographic parameters, lighting physics, historical art references, spatial composition rules, and deliberate debiasing techniques. High-craft image prompts construct an exact visual scene while actively counteracting the models' algorithmic drift toward plastic, oversaturated stereotypes.\n\n---\n\n## 4-Step Generative Image Prompt Architecture\n\n### Step 1: The 5-Part Architectural Prompt Anatomy\n1. **The Structural Prompt Formula**:\n   - `[Subject & Action]` + `[Environment & Context]` + `[Composition & Framing]` + `[Lighting & Atmosphere]` + `[Camera, Lens & Film Stock \u002F Style Engine]`\n2. **Component Breakdown**:\n   - *Subject*: Specific, concrete details (clothing texture, age, posture, expression, action).\n   - *Environment*: Exact time of day, architectural materials (brutalist raw concrete, wet asphalt, mahogany paneling).\n   - *Composition*: Rule of thirds, low-angle hero shot, Dutch tilt, wide-angle cinematic perspective.\n   - *Lighting*: Volumetric god rays, directional Rembrandt lighting, diffuse cloudy softbox, neon back-rim lighting.\n   - *Camera \u002F Medium*: 35mm photograph, Leica M11, f\u002F1.8 aperture, shallow depth of field, Kodak Portra 400 film grain.\n\n### Step 2: Eliminating Generative Clichés & Plastic Skin\n1. **The AI Slop Kill-List for Prompts**:\n   - Ban empty fluff words: `photorealistic`, `hyperrealistic`, `ultra-detailed`, `4K`, `8K`, `masterpiece`, `trending on artstation`. These words pollute the latent space with low-quality amateur renders.\n2. **Enforcing Natural Texture & Imperfection**:\n   - Specify real-world micro-textures to kill plastic rendering:\n     - Skin: *\"Subtle skin pores, natural skin texture, unretouched, faint freckles, uneven lighting.\"*\n     - Environment: *\"Dust motes in light beams, worn edges on leather notebook, subtle motion blur.\"*\n\n### Step 3: Countering Stereotypical Demographic Drift\n1. **Active Debiasing**:\n   - Image models default to severe cultural and gender stereotypes (e.g. prompt \"software engineer\" produces exclusively young white or Asian men; \"nurse\" produces young women).\n   - Explicitly specify diverse, realistic demographic descriptors, ages (e.g. \"50-year-old female systems architect\"), clothing (realistic workwear, not futuristic spandex), and natural working environments.\n\n### Step 4: Technical Parameters & Engine Flags\n1. **Platform-Specific Flag Tuning**:\n   - *Midjourney v6*:\n     - Aspect Ratios: `--ar 16:9` (landscape\u002Fhero), `--ar 1:1` (avatar), `--ar 4:5` (social portrait).\n     - Stylize: `--s 50` to `--s 150` (restrained realism; default 250 is often overly cartoonish).\n     - Weird & Chaos: Keep `--c 0` to `--c 10` for reproducible commercial production.\n   - *Flux.1 (Dev \u002F Schnell)*:\n     - Responds best to natural language paragraphs over comma-separated tag soup.\n\n---\n\n## Deliverable Format: AI Image Prompt Specification (`IMAGE-PROMPTS.md`)\n\n```markdown\n# AI Image Generation Prompt Book: [Campaign \u002F Feature]\n\n## 1. Asset: Hero Visual for Developer Documentation\n- **Target Model**: Flux.1 Dev \u002F Midjourney v6\n- **Aspect Ratio**: 16:9 (`--ar 16:9`)\n- **Visual Genre**: Documentary editorial photograph\n\n### The Master Prompt\n> An authentic 35mm documentary photograph of a 42-year-old female lead infrastructure engineer debugging code late at night in a server room. She wears a navy wool sweater and glasses with thin titanium frames. Subtle facial expression of intense concentration. The background features server racks with out-of-focus amber and soft white blinking LED indicator lights. Soft rim lighting from terminal screens, moody atmospheric shadows, shallow depth of field, f\u002F2.0 aperture. Shot on Leica M6 with Kodak Portra 400 film grain, natural skin pores and subtle shadows, zero plastic smoothing, zero neon sci-fi holograms. --ar 16:9 --s 100 --v 6.0\n\n## 2. Negative Constraints & Exclusions\n- No glowing holographic HUDs floating in air.\n- No purple or magenta cyberpunk neon lighting.\n- No airbrushed smooth plastic mannequin skin.\n```\n\n---\n\n## Worked Example: Overhauling a Generic Tech Hero Visual\n\n- **Original Prompt**: \"Software engineer working on futuristic AI, hyperrealistic, 8k, glowing screens, masterpiece.\" -> *Produced a cartoonish neon cyborg.*\n- **Refined Craft Prompt**: Specified natural documentary photography, natural daylight from a warehouse window, real mechanical keyboard, and Kodak film grain.\n- **Result**: Photorealistic, credible editorial image used as the hero graphic for an enterprise Series B announcement.\n\n---\n\n## Verification Checklist\n\n- [ ] Prompt follows the 5-part anatomical structure (Subject, Context, Framing, Lighting, Camera).\n- [ ] Banned fluff words (\"hyperrealistic\", \"8k\", \"trending on artstation\") are 100% eliminated.\n- [ ] Natural skin texture, grain, and physical imperfections are explicitly instructed.\n- [ ] Stereotypical algorithmic drift is countered with specific demographic descriptors.\n- [ ] Technical flags (`--ar`, `--s`) match target generation engine.\n\n---\n\n## Anti-Patterns\n\n- **Tag Soup Prompting**: Stacking 40 disconnected keywords with commas hoping the model figures it out.\n- **Cyberpunk Defaulting**: Letting models inject purple neon glows and floating binary code into serious enterprise visuals.\n- **Neglecting Aspect Ratios**: Generating square images for a 16:9 website hero banner and cropping off critical subjects.\n",{"name":5,"group":24,"description":7,"last_reviewed":25},"Implementation",1788652800000,"main","502b1a762b2c40a4ad7faa6297e2ebf7e4682efc","skills\u002Fimage-prompts",1,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskillary","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp","poorvithmp.com","Shipping tools for problems people notice too late.","2026-10-02 21:32:17",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add poorvith-mp\u002Fskills-design --skill image-prompts",{"tool":42,"command":43},"bunx","bunx skills add poorvith-mp\u002Fskills-design --skill image-prompts",{"tool":45,"command":46},"pnpm","pnpm dlx skills add poorvith-mp\u002Fskills-design --skill image-prompts","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fimage-prompts\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"evals\u002Frouting.yaml","REFERENCE","参考文档",1062,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fimage-prompts\u002Fevals\u002Frouting.yaml",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"image-prompts.skill","OTHER","其他",3676,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fimage-prompts\u002Fimage-prompts.skill",[63,70,76,82,88,95],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7906MDY214MTS6J24WC9HCM","youtube-thumbnail-maker","Moved to modernized catalog in skillary. Removed in v4.1.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-thumbnail-maker\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7906JSPSGETSSTXYTEDN5GT","youtube-popup-graphic","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-popup-graphic\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7906GYHJ8Z846MTA9FPEVE9","whimsy-injector","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fwhimsy-injector\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7906FK39C4JEE36EXD8R8ZC","visual-storyteller","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fvisual-storyteller\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7906D196H7PN42711NX46RE","ux-researcher","Moved to ux-research in skills-design. Removed in v4.1.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fux-researcher\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7906A5DJYRN1P5MQJPNN53K","ux-research","Plan and run usability studies and behavioural analysis, then turn findings into research-backed personas. Use when running usability tests, journey mapping, or observing user behavior.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-design\u002Fblob\u002Fmain\u002Fskills\u002Fux-research\u002FSKILL.md",[],[]]