Img Gen Prompts logo

Img Gen Prompts

Community
NanmiCoder
img-gen-prompts

Search oip-visual-v2 for traceable prompt-image pairs and open local result galleries. Use when finding, inspecting, comparing, browsing, or copying real image-generation references or exact source prompts. After references are selected, adapt them into a derived bilingual prompt when requested; use img-gen-taste first when the visual direction is still unclear.

Overview

PublisherNanmiCoder
Repositoryopen-image-prompts
Skill nameimg-gen-prompts
Stars
356
Forks
45
Bundled files
4
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 4 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by NanmiCoder on GitHub. Read the source before you install it.

Installation

Install the Img Gen Prompts AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/NanmiCoder/open-image-prompts/tree/main/skills/img-gen-prompts
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/NanmiCoder/open-image-prompts.git /tmp/open-image-prompts
mkdir -p .claude/skills
cp -r /tmp/open-image-prompts/skills/img-gen-prompts .claude/skills/img-gen-prompts
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Img Gen Prompts in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Img Gen Prompts on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Img Gen Prompts is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Img Gen Prompts

Act as the evidence and retrieval layer. Use only the latest paired oip-visual-v2 archive. Every recommended record must include a real image, unchanged source prompt, source URL, author, stable ID, and explainable match. Never fall back to legacy/V1 data or generic web image search.

Locate and verify the archive

Resolve this Skill's directory and run its bundled script by absolute path. It finds the checkout from the current directory, a symlinked Skill path, or OIP_REPO_ROOT; copied installations outside the checkout must set that variable.

bash
python3 <absolute-skill-directory>/scripts/oip.py status

Stop unless ready is true and the active taxonomy is exactly oip-visual-v2. The dataset is not part of the repository: db/prompts.db.gz and images/ are downloaded from GitHub Releases. If the archive, local images, or gallery dependencies are missing, run npm run setup from the repository root before continuing.

If the script reports that the repository cannot be found, this Skill was installed as a copy outside a checkout. Point it at one, or create one:

bash
git clone https://github.com/NanmiCoder/open-image-prompts.git
cd open-image-prompts
npm run data:pull        # add :db to skip the multi-gigabyte image packs
export OIP_REPO_ROOT="$PWD"

Search and retrieval work without the image packs; only local image preview needs them, and the gallery falls back to each record's source URL. This Skill is read-only: never activate a taxonomy, label records, publish data, or modify the database.

Search the user's actual intent

bash
python3 <absolute-skill-directory>/scripts/oip.py search \
  --intent "雨夜霓虹小巷里的电影感人像" --limit 8

python3 <absolute-skill-directory>/scripts/oip.py get --id <tweet_id>

Pass the user's wording without embellishing it with guessed styles, subjects, or negative constraints. Inspect parsed_intent, match_reasons, tags, and absolute_image_paths; open representative images before judging quality. Read references/retrieval-contract.md when interpreting constraints or an empty result.

Treat results as exact matches. related_results is a separate fallback channel that may omit exactly one declared aesthetic constraint. Never present a related reference as an exact match: disclose its missing_constraints and recommend only the visual decisions it actually demonstrates. Related references are already restricted to image-confirmed visual evidence and exclude unrequested screenshots, document graphics, heavy text, multi-panel layouts, and watermarks.

For each recommendation return:

  • stable ID, author, and source URL;
  • representative image path;
  • whether the match is exact or related;
  • any declared missing_constraints;
  • unchanged source_prompt;
  • two or three meaningful match reasons;
  • what visual decision is reusable and what should not be copied.

Translations and derived prompts are aids, never source prompts. If no relevant result exists in either channel, report the corpus gap instead of inventing provenance. The tool already performs the only allowed one-facet fallback; never manually relax the subject, use, explicit style, scene, user invariants, or forbidden constraints.

Some archive records contain commentary or post copy rather than an executable generation prompt. Label that distinction clearly and do not present such text as a reusable original prompt.

Adapt selected references

When requested, synthesize the user's selected coherent references into a new English prompt and a Simplified Chinese equivalent. Preserve any creative_spec invariants and avoid list supplied by $img-gen-taste. Keep the new prompt, rationale with source IDs, and original prompts clearly separated. Do not copy a complete source prompt, identity, brand, protected character, or unique scene.

Let the user choose visually

When the user wants to browse or compare results, read references/gallery-session.md, create a local result session, and open only the URL returned by the command.

bash
python3 <absolute-skill-directory>/scripts/oip.py gallery \
  --intent "克制的暗调香水广告" --lang zh-Hans --limit 8

The gallery must preserve search rank and keep “copy original prompt” byte-for-byte faithful. If url_usable is false, run the returned start_command before presenting the URL as openable.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Img Gen Prompts AI skill do?

Search oip-visual-v2 for traceable prompt-image pairs and open local result galleries. Use when finding, inspecting, comparing, browsing, or copying real image-generation references or exact source prompts. After references are selected, adapt them into a derived bilingual prompt when requested; use img-gen-taste first when the visual direction is still unclear.

Why use Img Gen Prompts on TypingMind?

Because you install it once and use it with any model. Img Gen Prompts is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Img Gen Prompts in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/NanmiCoder/open-image-prompts/tree/main/skills/img-gen-prompts. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Img Gen Prompts?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Img Gen Prompts?

As many as you like. As long as a model supports skills, you can use Img Gen Prompts with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Img Gen Prompts AI skill free?

Yes. It is published on GitHub by NanmiCoder under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇