Skip to content
Back to search

open-image-prompts

Open, local-first visual prompt archive with traceable prompt-image references and installable Agent Skills.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/NanmiCoder/open-image-prompts with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/NanmiCoder/open-image-prompts'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/nanmicoder--open-image-prompts/badge.svg)](/listings/nanmicoder--open-image-prompts)

HTML

<a href="/listings/nanmicoder--open-image-prompts"><img src="/api/listings/nanmicoder--open-image-prompts/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

List and install both Skills:

npx skills add NanmiCoder/open-image-prompts --list
npx skills add NanmiCoder/open-image-prompts -g

`img-gen-taste` works immediately from its bundled style cards. `img-gen-prompts` uses this repository's public SQLite archive and fetched images (`npm run data:pull` downloads both):

export OIP_REPO_ROOT="$PWD"  # PowerShell: $env:OIP_REPO_ROOT = (Get-Location)
npm run status

A ready checkout reports `"active_taxonomy_version": "oip-visual-v2"` and `"ready": true`.

Example:

python3 skills/img-gen-prompts/scripts/oip.py search \
  --intent "vintage city travel poster" \
  --limit 5

Search keeps exact matches in `results`. When exact coverage is sparse, a separate `related_results` channel may provide image-confirmed references that miss exactly one declared aesthetic preference. It never adds a vector database, model download, API key, or Python dependency. Run the labeled, bilingual 72-query regression benchmark (including visually reviewed related references) with:

npm run test:retrieval

On Windows, replace `python3` with `py -3` or use the Skill through a compatible Agent.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 199 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <p align="center">
  2. <img src="./assets/readme/hero.svg" width="100%" alt="Open Image Prompts">
  3. </p>
  4. <p align="center">
  5. <strong>English</strong> · <a href="./README.zh-CN.md">简体中文</a>
  6. </p>
  7. <p align="center">
  8. <a href="https://openimages.ajiang.me/"><img src="./assets/readme/screenshot-grid.png" width="100%" alt="Open Image Prompts — gallery, model collection, prompt detail and site index"></a>
  9. </p>
  10. # Open Image Prompts
  11. An open, local-first visual prompt archive with two installable Agent Skills:
  12. - `img-gen-taste` turns a rough brief into a clear art direction.
  13. - `img-gen-prompts` retrieves traceable prompt-image references and opens a local comparison gallery.
  14. Working through a coding agent? [AGENTS.md](./AGENTS.md) is the condensed setup,
  15. port, and Skill contract.
  16. The public dataset contains **23,665 source prompts**, **39,097 images**, **47,200 translations**, **275,474 active v2 prompt labels**, and a closed taxonomy of **185 visual labels**. Labeling models, backfill tools, provider configuration, test runs, error logs, and other labeling-process records are not included. These counts are checked against `data/public-corpus.json` by `npm run verify:docs`.
  17. Dataset assets ship through [GitHub Releases](https://github.com/NanmiCoder/open-image-prompts/releases) instead of Git LFS: the repository clone stays small, and `scripts/fetch_dataset.py` downloads the SQLite archive (~80 MB) plus optional monthly image packs (~4.3 GB total) with sha256 verification. See `data/dataset-manifest.json` for the exact asset list.
  18. **Videos.** The published packs carry images and each video's poster frame, not the video files themselves (2,061 clips would add roughly 8.8 GB). A video row points at `images/<tweet_id>/video_<index>.mp4`; when that file is absent the gallery falls back to the original X URL, which mainland networks cannot reach, expires over time and blocks hotlinked playback — so a fresh clone shows the poster but cannot play the clip, and the gallery says exactly that instead of failing silently.
  19. Pull the clips yourself when you want offline playback. `scripts/fetch_videos.py` reads every video row, downloads the clip from its original source and writes it to exactly the path the database points at — row by row, so a clip can never land in another tweet's directory:
  20. python3 scripts/fetch_videos.py --dry-run # plan only
  21. python3 scripts/fetch_videos.py --limit 20 # smoke-test a few clips
  22. python3 scripts/fetch_videos.py # everything still missing (~8.8 GB)
  23. It needs reachable access to `video.twimg.com`: run it on a machine outside mainland China (a VPS or workstation, or with a VPN) or hand it a proxy with `--proxy http://host:port` (it also honors `HTTPS_PROXY`). Re-running costs nothing — files that are already valid are skipped, interrupted downloads resume from the kept `.part`, and `--report failures.json` records what still failed so a later run can retry just those. `AGENTS.md` documents the same flow as an agent setup step.
  24. ## Repository vs. dataset assets
  25. This repository tracks the application code, frontend, API, Skills, taxonomy, and small dataset indexes. It intentionally does **not** commit the SQLite database or image files into Git history. Runtime data is downloaded from GitHub Releases into your local checkout:
  26. ```text
  27. open-image-prompts/
  28. ├── db/prompts.db.gz # SQLite dataset archive from Releases, gitignored
  29. ├── images/ # extracted image packs from Releases, gitignored
  30. ├── .oip/runtime/prompts.db # expanded read-only runtime SQLite, gitignored
  31. ├── data/dataset-manifest.json
  32. ├── data/public-corpus.json
  33. └── web/dims.json
  34. ```
  35. If you only clone the repository without fetching the dataset, you have the code but not the local prompt/image corpus needed for full preview and retrieval. From the repository root, run:
  36. ```bash
  37. npm run data:pull # downloads DB + all image packs and verifies sha256
  38. ```
  39. To download only the DB and skip the multi-gigabyte image packs:
  40. ```bash
  41. npm run data:pull:db
  42. # or
  43. python3 scripts/fetch_dataset.py --db-only
  44. ```
  45. DB-only mode supports search/retrieval while the gallery falls back to original source image URLs when local images are absent. Full local image preview requires the image packs.
  46. When the dataset updates, Git commits normally only change small files such as `data/dataset-manifest.json`, `data/public-corpus.json`, and `web/dims.json`. The large `prompts.db.gz` and `images-YYYY-MM.tar.gz` files are published as Release assets. Re-run `npm run data:pull` to download the new DB and only the image packs whose sha256 changed.
  47. ## One-click start
  48. Install [Git](https://git-scm.com/downloads) and [Node.js](https://nodejs.org/) 20.19+ or 22.12+, then clone the repository:
  49. ```bash
  50. git clone https://github.com/NanmiCoder/open-image-prompts.git
  51. cd open-image-prompts
  52. ```
  53. Start on macOS or Linux:
  54. ```bash
  55. ./start.sh
  56. ```
  57. Start on Windows:
  58. ```bat
  59. start.bat
  60. ```
  61. You can also double-click `start.bat` in File Explorer. The launcher installs [uv](https://docs.astral.sh/uv/) when needed, creates a compatible Python environment, downloads the dataset from GitHub Releases, installs the frontend packages, and starts both services. Open the local URL printed in the terminal. To skip the multi-gigabyte image packs (the gallery then falls back to original source URLs), set `OIP_FETCH_SKIP_IMAGES=1` before starting.
  62. The first start expands the compressed SQLite archive into the ignored `.oip/runtime/` directory. Later starts reuse the Python environment while refreshing locked dependencies.
  63. ## Dataset assets
  64. The Git repository does not store the large dataset files directly. A clone gives you the app code, Skills, public metadata, and `data/dataset-manifest.json`. The SQLite database and image packs must be downloaded from GitHub Releases before the full local gallery can run.
  65. Download the complete dataset:
  66. ```bash
  67. npm run data:pull
  68. ```
  69. This command reads `data/dataset-manifest.json`, downloads the release assets, verifies their sha256 hashes, and places them in the paths expected by the app:
  70. - `db/prompts.db.gz` is the compressed public SQLite database.
  71. - `images/` receives the extracted monthly image packs from `images-YYYY-MM.tar.gz`.
  72. - `.oip/packs/` stores local extraction markers so unchanged packs are skipped on the next run.
  73. To download only the database and let the gallery fall back to original source image URLs:
  74. ```bash
  75. npm run data:pull:db
  76. ```
  77. These generated files are intentionally ignored by Git. Dataset releases may update often, but repository commits stay small: Git tracks code and lightweight metadata, while GitHub Releases carry `prompts.db.gz` and the image archives.
  78. ## Run with Docker
  79. Docker provides a Linux-isolated runtime with Node.js 22 and Python 3. The image build runs the public data checks, API/frontend tests, lint, and production build before producing the runtime image:
  80. ```bash
  81. docker build -t open-image-prompts .
  82. docker run --rm --name open-image-prompts -p 4173:4173 open-image-prompts
  83. ```
  84. Open <http://localhost:4173>. The API remains loopback-only inside the container and is exposed only through the frontend proxy. The image runs as the unprivileged `node` user and includes a `/health` health check.
  85. The build downloads the SQLite archive from GitHub Releases (network access to github.com is required) and serves images through source-URL fallback. The same commands work with Docker Desktop on Windows/macOS and Docker Engine on Linux.
  86. ## Install the Skills
  87. List and install both Skills:
  88. ```bash
  89. npx skills add NanmiCoder/open-image-prompts --list
  90. npx skills add NanmiCoder/open-image-prompts -g
  91. ```
  92. `img-gen-taste` works immediately from its bundled style cards. `img-gen-prompts` uses this repository's public SQLite archive and fetched images (`npm run data:pull` downloads both):
  93. ```bash
  94. export OIP_REPO_ROOT="$PWD" # PowerShell: $env:OIP_REPO_ROOT = (Get-Location)
  95. npm run status
  96. ```
  97. A ready checkout reports `"active_taxonomy_version": "oip-visual-v2"` and `"ready": true`.
  98. Example:
  99. ```bash
  100. python3 skills/img-gen-prompts/scripts/oip.py search \
  101. --intent "vintage city travel poster" \
  102. --limit 5
  103. ```
  104. Search keeps exact matches in `results`. When exact coverage is sparse, a
  105. separate `related_results` channel may provide image-confirmed references that
  106. miss exactly one declared aesthetic preference. It never adds a vector
  107. database, model download, API key, or Python dependency. Run the labeled,
  108. bilingual 72-query regression benchmark (including visually reviewed related
  109. references) with:
  110. ```bash
  111. npm run test:retrieval
  112. ```
  113. On Windows, replace `python3` with `py -3` or use the Skill through a compatible Agent.
  114. ## Public data boundary
  115. The public DB deliberately contains only product runtime data:
  116. - source prompts and source URLs;
  117. - image records for the full public corpus;
  118. - bilingual translations;
  119. - active `oip-visual-v2` prompt/image labels;
  120. - the public taxonomy and FTS search index.
  121. It does **not** contain labeling candidates, model/provider settings, run IDs, leases, model rationales, error paths, evaluation tables, or legacy label assignments.
  122. See [DATASET.md](./DATASET.md), [DATA_LICENSE.md](./DATA_LICENSE.md), and the machine-readable [public corpus manifest](./data/public-corpus.json).
  123. ## Validate a checkout
  124. ```bash
  125. uv sync --locked
  126. npm --prefix web ci
  127. npm test
  128. npm run lint
  129. npm run build
  130. npm run status
  131. ```
  132. The API and Skill open SQLite in read-only immutable mode. Every service binds `127.0.0.1` by default and never starts a labeling job. The frontend starts at port `5173` and moves to the next free port when it is taken, printing the URL it actually serves; set `OIP_WEB_HOST`/`OIP_WEB_PORT` to pin them, in which case a port collision fails loudly instead of drifting. The Skill's gallery bridge behaves the same way around port `4173`.
  133. ## License
  134. Application code and Skill instructions are available under the [MIT License](./LICENSE). Dataset licensing and third-party-content boundaries are documented separately in [DATA_LICENSE.md](./DATA_LICENSE.md).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.