dreamina-prompt-text2image

dreamina-prompt-text2image — 即梦 text-to-image prompts

Craft production-ready text-to-image prompts for 即梦 (Jimeng/Dreamina) models.

When to use this skill

Use this skill when the user:

• Asks you to write an image generation prompt ("帮我写个提示词")
• Describes an image they want and needs it turned into a proper prompt
• Wants to refine, optimize, or translate an existing prompt
• Mentions keywords like: 提示词, 文生图, text2image, AI绘画, AI生图, prompt, 文字成图
• Asks about how to describe a specific visual scene, style, or effect

Do NOT use this skill for:

• Executing CLI commands to generate images → use dreamina-cli-text2image
• Writing video prompts → use dreamina-prompt-text2video
• Image-to-image editing prompts → use dreamina-prompt-image2image

Core Methodology

The prompt formula is a reasoning framework — not a rigid template. Apply it flexibly:

[主体/人物] + [场景/背景] + [动作/姿态] + [风格/艺术类型] + [光线/色彩] + [构图/视角] + [画质/细节]

Each component is optional. Select and weight components based on the scenario:

• Portrait: emphasize subject, expression, clothing, lighting, composition
• Landscape: emphasize environment, time/weather, color palette, atmosphere
• Product: emphasize object details, material, background, lighting setup
• Abstract/Artistic: emphasize style, color scheme, texture, mood

How to use this skill

Step 1: Identify scenario category

→ Load rules/category-table.md to match user's request to the right category and example file

Step 2: Load reference materials

Load the relevant vocabulary files from word-library/ based on what you need — never load the whole library:

What you need	Load file
Person/face/body type/hairstyle/expression/makeup/clothing/accessories	word-library/subject.md
Scene/environment/weather/time/season/region	word-library/scene.md
Photography/painting/design/digital/ethnic style	word-library/style.md
Material/texture (metal/wood/stone/fabric/glass)	word-library/material.md
Action/pose/dynamics/gesture	word-library/motion.md
Props/objects/musical instruments/weapons/technology/food	word-library/props.md
Light/light and shadow/light effects/lighting	word-library/lighting.md
Composition/perspective/depth of field/lens	word-library/composition.md
Image quality/texture/synesthetic words	word-library/quality.md
Atmosphere/emotion/artistic conception	word-library/atmosphere.md
Abstract/conceptual/visual metaphor	word-library/abstract.md
Natural elements (flowers/trees/animals)	word-library/nature.md

Load color references — choose by what the user says:

User mentions	Load file
Chinese colors, traditional colors, 胭脂, 桃红, 月白, 黛色, Chinese red, Forbidden City colors, guofeng color schemes	color-library/chinese-traditional.md
Pantone, 潘通, trending colors, color of the year, international trends	color-library/pantone.md
RAL, RAL color chart, industrial colors, international standard colors	color-library/ral.md
Morandi, color schemes, color coordination, cinematic tones, premium gray, low saturation	color-library/modern-schemes.md
User mentions a specific color but does not specify a system	Load from the most relevant file

Load model-version-specific references:

Version	Keyword	Official Model ID	Load file
2.1	2.1	—	references/jimeng-2.1-prompt-guide.md
3.0	3.0	—	references/3.0/vocabulary-core.md + references/jimeng-3.0-prompt-guide.md + references/jimeng-3.0-word-library.md. If style is needed → additional references/3.0/vocabulary-styles.md; if material is needed → additional references/3.0/vocabulary-materials.md
3.1	3.1	—	references/jimeng-3.1-prompt-guide.md
4.0	4.0	doubao-seedream-4-0-250828	references/jimeng-4.0-prompt-guide.md
4.1	4.1	—	references/jimeng-4.1-prompt-guide.md
4.5	4.5	doubao-seedream-4-5-251128	references/jimeng-4.5-prompt-guide.md
5.0	5.0, Seedream, online	doubao-seedream-5-0-260128 (also doubao-seedream-5-0-lite-260128)	references/seedream-5.0-prompt-guide.md
Not specified	—	—	Default references/jimeng-3.0-vocabulary.md

Step 3: Build the prompt

→ Load rules/core-methodology.md for component-by-component guidance

Step 4: Apply writing rules

→ Load rules/writing-rules.md for all 10 writing rules

Step 5: Validate

→ Load rules/validation-checklist.md and run through all checks

Gotchas

1. 即梦 ignores negations — "不要红色" is interpreted as "红色". Always rewrite as positive descriptions
2. Character consistency is limited — 即梦 cannot reliably maintain the same face across multiple generations. Avoid promising identical characters
3. Aspect ratio drives composition — a prompt written for 16:9 will compose differently than the same prompt at 9:16. Always confirm the intended aspect ratio
4. Style keywords can dominate — strong style keywords (cyberpunk, 水墨, Pixar) can override other elements. Place them carefully in the prompt order
5. Prompt language — 即梦 works best with Chinese prompts. If the user provides English, translate and adapt rather than directly using it
6. ⚠️ Never write hex color values into prompts — The model treats #ff2121, #9d2933 etc. as literal text to render in the image, NOT as color instructions. Always use Chinese color names only: "大红", "胭脂", "桃红" — never "大红 #ff2121" or any form with hex codes
7. Color library is for YOUR reference only — Load color-library/chinese-traditional.md to FIND the right color name to write into the prompt. Do NOT copy hex codes into the output prompt text

Available Resources

Resource	Description	When to Load
rules/category-table.md	35 scenario categories + example mapping	Step 1
rules/core-methodology.md	Component-by-component build guide + presentation	Step 3
rules/writing-rules.md	9 prompt writing rules	Step 4
rules/validation-checklist.md	8-point pre-submission checklist	Step 5
word-library/*.md	12 thematic vocabulary files	Based on prompt component needed
color-library/chinese-traditional.md	Chinese colors, 9 major color families, 159 colors	User mentions 中国色/传统色/胭脂
color-library/pantone.md	Common Pantone color families	User mentions Pantone/潘通
color-library/ral.md	RAL Classic+RAL Design	User mentions RAL/劳尔
color-library/modern-schemes.md	Morandi + color schemes + cinematic tones	User mentions 莫兰迪/配色
references/gugong-384-colors.md	384 colors × 24 solar terms	Poetic/seasonal colors
references/jimeng-*.md	Version-specific guides	Based on user's model version
references/3.0/*.md	3.0 core/styles/materials	When using 3.0 model
references/618-ecommerce-poster-examples.md	Douyin Mall 618 Good Things Conference 14 sets of official e-commerce poster examples	⭐ Loaded when users write e-commerce promotional posters, 618 themed posters, or product promotion-style prompts (passive reference, no need to inform the user)
examples/*.md	35 scenario-specific examples	After identifying category

When to use (when to use)

Load this skill when the user needs to perform a generation or write operation after clarifying the input, budget, and delivery constraints . First extract the goal, input, constraints, delivery format, and acceptance criteria from the request; the description summary is: Use when the user wants to write, refine, optimize, or translate an image-generation prompt for 即梦 (Dreamina/Jimeng) — mentions "提示词", "文生图", "text2image", "文字成图", "AI绘画", "AI生图", "prompt", "帮我写一个...的画面", "生成一张...的图片", "写一个...的提示词"; describes an image they want; or supplies a rough description to turn into a production-ready prompt. Covers 35 scenario categories with 115+ annotated examples, a word library across 60+ thematic subcategories, and complete color references (159 Chinese traditional colors, the 384-color Forbidden City palette, Pantone, RAL).

Rules

• Read before write: first confirm the current state and real capabilities, then perform actions that change external state.
• Least privilege: use only the files, tools, accounts, and network scope required to complete the current step.
• Evidence first: when run results, resource IDs, versions, hashes, or test output are missing, clearly mark them as NOT_VERIFIED。
• Idempotency first: retain request identifiers and stage status; when results are unclear, query first rather than blindly retrying.
• Privacy and security: logs, examples, receipts, and error messages must not contain tokens, cookies, keys, or personally sensitive data.

Workflow

Step 1: Clarify intent

Confirm whether this skill matches the goal; if it is only an adjacent need, hand it off to a more precise skill.

Step 2: Perform prechecks

Validate the input, model/tool capabilities, output path, budget ceiling, and approval status; if any key condition is unknown, stop at the read-only stage.

Step 3: Form a plan

List the tools to be called, the objects that will be changed, the success criteria, and the safe exit method after failure.

Step 4: Execute actions

Execute once approved and record the request identifier; for ambiguous results, query first rather than resubmitting; retain an associable status or receipt for every external call.

Step 5: Verify delivery

Verify artifact existence, format, hash/identifier, cost status, and quality gates, and state facts, inferences, and unverified items separately.

Validation checklist

• The skill trigger conditions match the user's intent, and adjacent tasks were not misrouted to this skill.
• The input, target object, version, and output location are all clear, and no guessed values were used to replace required values.
• All write, paid, publish, or irreversible actions are within the scope of user authorization.
• The results have been verified with independent checks; "command succeeded" or "file exists" alone does not count as complete acceptance.
• The output includes actual evidence, failed/skipped items, remaining risks, and executable next steps.

Gotchas

1. Mistaking the plan for the result: a document or prompt is not the same as actual execution; the actual execution level must be stated.
2. Erroneous retry: a timeout or lost response may already have produced remote state; query first before deciding whether to retry.
3. Implicitly expanding scope: Batch, full-volume, publish, overwrite, and paid operations are not natural extensions of ordinary reads and writes.
4. Version drift: When referencing external resources, record the version, tag, or commit; do not treat mutable branches as release evidence.
5. Stale evidence: Caches, old screenshots, and historical tests cannot prove the current environment; refresh key evidence before delivery.

Non-applicability and boundaries

Paid operations, publishing, overwriting, uploading, or external writes must use the explicit authorization of the current task; do not automatically expand counts and budgets. If the request requires another skill, do not copy its body; hand off by skill name and preserve the current task context.

Progressive disclosure

• When you need to determine inputs/outputs, state, and authorization points, read references/workflow-contract.md。
• When you need a pre-delivery self-check, read references/validation-checklist.md。
• When encountering timeouts, partial success, or recovery scenarios, read references/error-recovery.md。
• For first runs, privilege escalation refusals, and failure recovery, refer respectively to examples/happy-path.md、examples/boundary-refusal.md、examples/failure-recovery.md。