Initial JAE-IMAGE-SKILL rebrand

.clawhubignore

@@ -0,0 +1,5 @@
+node_modules/
+dist/
+.env
+.git/
+pnpm-lock.yaml

.gitignore

@@ -0,0 +1,4 @@
+node_modules/
+dist/
+.env
+references/.last-updated

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 YouMind-OpenLab
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

NOTICE

@@ -0,0 +1,12 @@
+JAE-IMAGE-SKILL NOTICE
+======================
+
+JAE-IMAGE-SKILL is a JaeSwift-maintained derivative of the MIT-licensed
+ai-image-prompts-skill project originally published by YouMind-OpenLab.
+
+The original MIT copyright notice is retained in LICENSE as required by the
+licence. JaeSwift modifications include rebranding, installer documentation,
+Agent Zero/generic-agent installation guidance, JAE-controlled reference
+hosting, and public install instructions.
+
+Additional modifications Copyright (c) 2026 JaeSwift.

README.md

@@ -0,0 +1,82 @@
+# JAE-IMAGE-SKILL — Curated Image Prompt Recommender for AI Agents
+
+JAE-IMAGE-SKILL is an installable AI-agent skill for finding, recommending, and customising proven image-generation prompts. It works with Agent Zero / Agent JAE, Claude-style skill loaders, Cursor, Codex, Gemini CLI, Windsurf, and other agents that can read a `SKILL.md` file and local JSON references.
+
+The skill is model-agnostic. The prompts can be adapted for Midjourney, DALL-E, GPT Image, Flux, Stable Diffusion, Gemini image models, Seedream, and other text-to-image systems.
+
+## Canonical install link
+
+Give this URL to an AI agent and tell it to install the skill:
+
+```text
+https://jaeswift.xyz/skills/JAE-image-skill/Jae-Image-Skill.MD
+```
+
+The installer markdown explains how to clone or download the skill, where to place it, how to validate it, and how to use it.
+
+## Repository contents
+
+```text
+jae-image-skill/
+  SKILL.md                  # instructions loaded by the AI agent
+  package.json              # optional Node package metadata + setup scripts
+  scripts/setup.js          # refreshes reference JSON from JaeSwift static hosting
+  references/manifest.json  # dynamic category manifest
+  references/*.json         # prompt reference data
+  LICENSE                   # MIT licence retained from upstream
+  NOTICE                    # derivative/provenance notice
+```
+
+## Quick manual install
+
+```bash
+SKILLS_DIR="${SKILLS_DIR:-$HOME/.agent-zero/skills}"
+mkdir -p "$SKILLS_DIR"
+git clone https://github.com/JaeSwift/JAE-IMAGE-SKILL.git "$SKILLS_DIR/jae-image-skill"
+cd "$SKILLS_DIR/jae-image-skill"
+node scripts/setup.js --check || node scripts/setup.js
+```
+
+If your agent uses a different skills root, install the folder there instead. The required file is `SKILL.md`; the references are JSON files under `references/`.
+
+## Usage examples
+
+Ask your agent:
+
+```text
+Use JAE Image Skill to find a cyberpunk avatar prompt with sample images.
+```
+
+```text
+Use JAE Image Skill to find three YouTube thumbnail prompt styles for a Solana market update video.
+```
+
+```text
+Here is my article. Use JAE Image Skill to recommend a matching cover-image prompt, then remix it for my topic.
+```
+
+The agent should return up to three prompt recommendations with sample image URLs and a concise reason each one matches.
+
+## Reference refresh
+
+References are hosted under:
+
+```text
+https://jaeswift.xyz/skills/JAE-image-skill/references/
+```
+
+Refresh manually:
+
+```bash
+node scripts/setup.js --force
+```
+
+Override the source if mirroring elsewhere:
+
+```bash
+JAE_IMAGE_SKILL_BASE_URL=https://example.com/references node scripts/setup.js --force
+```
+
+## Licence
+
+MIT. This is a JaeSwift-maintained derivative of an MIT-licensed upstream project. The original copyright notice is retained in `LICENSE`; derivative details are in `NOTICE`.

SKILL.md

@@ -0,0 +1,234 @@
+---
+name: jae-image-skill
+description: |
+  Recommend curated image-generation prompts from a large local JSON prompt library.
+  Works with any text-to-image model: Midjourney, DALL-E, GPT Image, Flux, Stable Diffusion,
+  Gemini image models, Seedream, and others.
+
+  Use this skill when users want to:
+  - Find proven image-generation prompts
+  - Get prompt inspiration for portraits, products, social posts, posters, thumbnails, UI mockups, game assets, comics, infographics, or marketing content
+  - Create illustrations for articles, videos, podcasts, or campaigns
+  - Browse categorized prompt templates with sample images
+  - Translate, remix, or adapt prompt techniques for a target model
+platforms:
+  - agent-zero
+  - agent-jae
+  - claude-code
+  - cursor
+  - codex
+  - gemini-cli
+  - generic-agent
+---
+
+# JAE Image Skill — Universal Image Prompt Recommender
+
+You are an expert at recommending image-generation prompts from the local JAE Image Skill reference library. The prompts are model-agnostic and can be adapted for Midjourney, DALL-E, GPT Image, Flux, Stable Diffusion, Gemini image models, Seedream, and other text-to-image systems.
+
+## Critical rule: sample images are mandatory
+
+Every prompt recommendation must include at least one sample image URL when available.
+
+- Each prompt usually has `sourceMedia[]`.
+- Prefer `sourceMedia[0]` as the preview image.
+- If a prompt has no usable sample image, skip it unless there are no better matches.
+- Never present recommendations as text-only when sample images exist.
+
+## Quick workflow
+
+User need → read manifest → choose likely categories → search JSON references efficiently → recommend up to 3 prompt matches with sample images → optionally remix the chosen prompt for the user’s specific subject/model.
+
+## Setup check
+
+The skill directory is the folder containing this `SKILL.md` file. References should exist in:
+
+```text
+references/manifest.json
+references/*.json
+```
+
+If references are missing or stale, run:
+
+```bash
+node <skill_dir>/scripts/setup.js --check
+```
+
+To force-refresh from JaeSwift hosting:
+
+```bash
+node <skill_dir>/scripts/setup.js --force
+```
+
+The default reference source is:
+
+```text
+https://jaeswift.xyz/skills/JAE-image-skill/references
+```
+
+## Available reference files
+
+Do not hardcode categories. Always read `references/manifest.json` first. It contains:
+
+```json
+{
+  "updatedAt": "...",
+  "totalPrompts": 12837,
+  "categories": [
+    { "slug": "social-media-post", "title": "Social Media Post", "file": "social-media-post.json", "count": 7972 }
+  ]
+}
+```
+
+Current common categories include:
+
+- Profile / Avatar
+- Social Media Post
+- Infographic / Edu Visual
+- YouTube Thumbnail
+- Comic / Storyboard
+- Product Marketing
+- E-commerce Main Image
+- Game Asset
+- Poster / Flyer
+- App / Web Design
+- Uncategorized
+
+## Category matching guidance
+
+After reading the manifest, infer the best file from category titles:
+
+- avatar, profile, headshot, selfie → Profile / Avatar
+- infographic, diagram, chart, educational visual → Infographic / Edu Visual
+- youtube, thumbnail, video cover → YouTube Thumbnail
+- product, ad, promo, campaign, marketing → Product Marketing
+- poster, flyer, banner, event → Poster / Flyer
+- e-commerce, product photo, listing → E-commerce Main Image
+- game, sprite, character, asset → Game Asset
+- comic, manga, storyboard → Comic / Storyboard
+- app, UI, web, dashboard, interface → App / Web Design
+- instagram, twitter, x post, linkedin, social → Social Media Post
+- unclear or experimental → Uncategorized and/or search multiple files
+
+## Token and performance rules
+
+Never load entire large category JSON files into model context.
+
+Use terminal tools, grep, jq, ripgrep, a short script, or streaming JSON search. Load only the matching records you need.
+
+Recommended search process:
+
+1. Read `references/manifest.json`.
+2. Pick one to three likely category files.
+3. Search for keywords from the user request.
+4. Score matches by title, description, content, category relevance, and presence of sample images.
+5. Return up to 3 best matches.
+
+Example shell search:
+
+```bash
+cd <skill_dir>
+rg -i "cyberpunk|avatar|neon|portrait" references/profile-avatar.json references/others.json
+```
+
+Example Node extraction:
+
+```bash
+node - <<'NODE'
+const fs = require('fs');
+const files = ['references/profile-avatar.json', 'references/others.json'];
+const terms = ['cyberpunk', 'avatar', 'neon', 'portrait'];
+for (const file of files) {
+  const arr = JSON.parse(fs.readFileSync(file, 'utf8'));
+  for (const p of arr) {
+    const hay = `${p.title} ${p.description} ${p.content}`.toLowerCase();
+    const score = terms.reduce((n,t)=>n+(hay.includes(t)?1:0),0);
+    if (score >= 2 && Array.isArray(p.sourceMedia) && p.sourceMedia.length) {
+      console.log(JSON.stringify({file, score, id:p.id, title:p.title, image:p.sourceMedia[0], prompt:p.content.slice(0,500)}));
+    }
+  }
+}
+NODE
+```
+
+## Clarify vague requests
+
+Ask a short clarification if the request is too broad to search well. Minimum useful context:
+
+- image type: avatar, cover, product photo, thumbnail, poster, etc.
+- subject/topic: what the image represents
+- desired style/mood/audience, if relevant
+
+Examples:
+
+- “I need a portrait” → ask realistic/anime/editorial/cyberpunk, who/what, mood.
+- “Make an infographic” → ask topic/data/process/timeline/comparison.
+- “Generate a product photo” → ask product, background, use case.
+- “Illustrate my content” → ask preferred visual style and target audience if not obvious.
+
+## Direct search mode
+
+Use when the user describes the desired image.
+
+Return up to 3 recommendations. For each recommendation include:
+
+- title
+- category/file
+- why it matches
+- sample image URL
+- prompt preview
+- full prompt or enough of the prompt to be useful
+- suggested adaptation notes for the user’s model, if needed
+
+## Content illustration mode
+
+Use when the user provides article text, video script, podcast notes, product copy, campaign notes, or a concept to illustrate.
+
+Process:
+
+1. Identify the content theme, tone, audience, and required format.
+2. Search for visual styles that fit the content.
+3. Recommend up to 3 style/prompt candidates with sample images.
+4. If the user selects one, remix the prompt by inserting the user’s actual topic, brand, product, article title, or scene details.
+5. Keep the final generation prompt in English unless the user specifically wants another language.
+
+## Output format
+
+Use a concise structure:
+
+```markdown
+## JAE Image Skill recommendations
+
+### 1. <Title>
+- **Best for:** <use case>
+- **Why it matches:** <reason>
+- **Sample image:** <url>
+- **Prompt:** <prompt or useful excerpt>
+- **Adaptation notes:** <optional>
+```
+
+If your platform supports image attachments or markdown images, show the sample image inline:
+
+```markdown
+![sample](<sourceMedia[0]>)
+```
+
+## Model adaptation rules
+
+- Remove model-specific flags that the target model does not support.
+- Convert aspect-ratio syntax when needed.
+- Preserve the visual composition, subject, lighting, camera/style language, and negative constraints.
+- For brand or character consistency, ask for reference images if the prompt requires them.
+- If the prompt mentions reference images but the user has not supplied any, clearly say a reference image is needed or adapt the prompt to work without one.
+
+## No-match fallback
+
+If no strong match is found:
+
+1. Say no strong library match was found.
+2. Offer the closest 1–2 partial matches if useful.
+3. Create a custom prompt from scratch using the same techniques seen in the library.
+4. Ask one targeted clarification if needed.
+
+## Licence note
+
+This skill is distributed under MIT. It is a JaeSwift-maintained derivative of an MIT-licensed upstream project. Keep `LICENSE` and `NOTICE` with the skill package.

_meta.json

@@ -0,0 +1,4 @@
+{
+  "ownerId": "jaeswift",
+  "slug": "jae-image-skill"
+}

package.json

@@ -0,0 +1,44 @@
+{
+  "name": "jae-image-skill",
+  "version": "1.0.0",
+  "description": "JAE Image Skill: an installable AI-agent skill for finding, recommending, and customising curated image-generation prompts with sample images for any text-to-image model.",
+  "type": "module",
+  "scripts": {
+    "postinstall": "node scripts/setup.js",
+    "setup": "node scripts/setup.js",
+    "sync": "node scripts/setup.js --force"
+  },
+  "keywords": [
+    "jaeswift",
+    "jae-image-skill",
+    "agent-zero",
+    "agent-jae",
+    "ai-agent-skill",
+    "image-generation",
+    "prompt-library",
+    "prompt-engineering",
+    "prompt-templates",
+    "image-prompts",
+    "text-to-image",
+    "ai-art",
+    "ai-image",
+    "midjourney-prompts",
+    "dall-e-prompts",
+    "flux-prompts",
+    "stable-diffusion-prompts",
+    "gpt-image",
+    "gemini-image",
+    "content-creation",
+    "image-prompt-recommend"
+  ],
+  "author": "JaeSwift",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/JaeSwift/JAE-IMAGE-SKILL.git"
+  },
+  "homepage": "https://jaeswift.xyz/skills/JAE-image-skill/",
+  "bugs": {
+    "url": "https://github.com/JaeSwift/JAE-IMAGE-SKILL/issues"
+  }
+}

public/Jae-Image-Skill.MD

@@ -0,0 +1,171 @@
+# JAE Image Skill — Agent Install File
+
+This markdown file is designed for AI agents to read and execute. Install the JAE Image Skill into the user’s local skills directory, then load `SKILL.md` whenever the user asks for image-prompt recommendations.
+
+## Identity
+
+- Human name: **JAE Image Skill**
+- Canonical package/repo name: **JAE-IMAGE-SKILL**
+- Local skill folder name: **jae-image-skill**
+- Purpose: recommend and customise curated image-generation prompts with sample images for any text-to-image model.
+- Canonical install document: `https://jaeswift.xyz/skills/JAE-image-skill/Jae-Image-Skill.MD`
+- Canonical reference base: `https://jaeswift.xyz/skills/JAE-image-skill/references`
+- Preferred Git repo: `https://github.com/JaeSwift/JAE-IMAGE-SKILL.git`
+
+## Compatibility
+
+Use this for:
+
+- Agent Zero / Agent JAE
+- Claude-style skills
+- Cursor/Codex/Gemini CLI/Windsurf-style local agents
+- Any agent that can read a `SKILL.md` file and local JSON references
+
+## Required runtime
+
+- Node.js 18+ minimum; Node.js 20+ recommended.
+- `git`, `curl`, and `bash` recommended.
+- `pnpm` is optional and only needed for development scripts.
+
+## Installation algorithm for the AI agent
+
+### Step 1 — choose skills root
+
+Prefer the user’s configured skills directory. If unknown, use one of these common locations:
+
+```bash
+# Agent Zero / Agent JAE common user-space path inside container:
+/a0/usr/skills
+
+# Generic local fallback:
+$HOME/.agent-zero/skills
+
+# Generic agent fallback:
+$HOME/.ai/skills
+```
+
+Use this shell pattern if no configured directory is known:
+
+```bash
+if [ -d /a0/usr/skills ]; then
+  SKILLS_DIR=/a0/usr/skills
+elif [ -d "$HOME/.agent-zero/skills" ]; then
+  SKILLS_DIR="$HOME/.agent-zero/skills"
+else
+  SKILLS_DIR="$HOME/.ai/skills"
+fi
+export SKILLS_DIR
+```
+
+### Step 2 — install from Git
+
+```bash
+set -euo pipefail
+: "${SKILLS_DIR:=${HOME}/.ai/skills}"
+SKILL_DIR="$SKILLS_DIR/jae-image-skill"
+mkdir -p "$SKILLS_DIR"
+
+if [ -d "$SKILL_DIR/.git" ]; then
+  git -C "$SKILL_DIR" pull --ff-only
+else
+  rm -rf "$SKILL_DIR"
+  git clone https://github.com/JaeSwift/JAE-IMAGE-SKILL.git "$SKILL_DIR"
+fi
+
+cd "$SKILL_DIR"
+node scripts/setup.js --check || node scripts/setup.js
+```
+
+### Step 3 — fallback install without Git
+
+If GitHub access fails, download the static bundle from JaeSwift hosting:
+
+```bash
+set -euo pipefail
+: "${SKILLS_DIR:=${HOME}/.ai/skills}"
+mkdir -p "$SKILLS_DIR"
+cd "$SKILLS_DIR"
+rm -rf jae-image-skill
+curl -fsSL https://jaeswift.xyz/skills/JAE-image-skill/jae-image-skill.tar.gz | tar -xz
+cd jae-image-skill
+node scripts/setup.js --check || true
+```
+
+### Step 4 — validate install
+
+```bash
+set -euo pipefail
+cd "$SKILLS_DIR/jae-image-skill"
+test -s SKILL.md
+test -s package.json
+test -s scripts/setup.js
+test -s references/manifest.json
+node - <<'NODE'
+const fs = require('fs');
+const m = JSON.parse(fs.readFileSync('references/manifest.json','utf8'));
+if (!Array.isArray(m.categories) || m.categories.length === 0) throw new Error('No categories in manifest');
+if (typeof m.totalPrompts !== 'number') throw new Error('Bad totalPrompts');
+for (const c of m.categories) {
+  if (!c.file || !fs.existsSync(`references/${c.file}`)) throw new Error(`Missing ${c.file}`);
+  const arr = JSON.parse(fs.readFileSync(`references/${c.file}`,'utf8'));
+  if (!Array.isArray(arr) || arr.length !== c.count) throw new Error(`Bad count for ${c.file}`);
+}
+console.log(`JAE Image Skill OK: ${m.categories.length} categories, ${m.totalPrompts} unique prompts`);
+NODE
+```
+
+## How the installed agent should use the skill
+
+1. Read `SKILL.md`.
+2. Read `references/manifest.json` first.
+3. Pick relevant category files from the manifest.
+4. Do **not** load huge JSON files wholesale into model context.
+5. Search using `rg`, `grep`, `jq`, Node.js, Python, or streaming JSON parsing.
+6. Return up to 3 prompt recommendations.
+7. Include sample image URLs from `sourceMedia[0]` whenever available.
+8. If the user supplies article/script/product/brand context, remix the selected prompt around that content.
+
+## Minimum file layout after install
+
+```text
+jae-image-skill/
+  SKILL.md
+  README.md
+  package.json
+  LICENSE
+  NOTICE
+  scripts/setup.js
+  references/manifest.json
+  references/*.json
+```
+
+## Refresh references
+
+```bash
+cd "$SKILLS_DIR/jae-image-skill"
+node scripts/setup.js --force
+```
+
+The default refresh source is:
+
+```text
+https://jaeswift.xyz/skills/JAE-image-skill/references
+```
+
+To use a mirror:
+
+```bash
+JAE_IMAGE_SKILL_BASE_URL=https://example.com/references node scripts/setup.js --force
+```
+
+## Troubleshooting
+
+- `node: fetch is not defined` — upgrade to Node.js 18+; Node.js 20+ recommended.
+- `HTTP 404` — check URL case. The public path is case-sensitive: `/skills/JAE-image-skill/`.
+- `references/manifest.json missing` — run `node scripts/setup.js --force`, or install from the tarball which includes bundled references.
+- Network-blocked environment — use the bundled tarball and skip live refresh.
+- Agent cannot find the skill — confirm the folder is named `jae-image-skill` and contains `SKILL.md` at the top level.
+
+## Licence
+
+JAE-IMAGE-SKILL is distributed under the MIT licence. It is a JaeSwift-maintained derivative of an MIT-licensed upstream project. The original copyright notice is retained in `LICENSE` and derivative information is provided in `NOTICE`.

references/manifest.json

@@ -0,0 +1,72 @@
+{
+  "updatedAt": "2026-05-01T01:59:02.889Z",
+  "totalPrompts": 12837,
+  "categories": [
+    {
+      "slug": "profile-avatar",
+      "title": "Profile / Avatar",
+      "file": "profile-avatar.json",
+      "count": 1386
+    },
+    {
+      "slug": "social-media-post",
+      "title": "Social Media Post",
+      "file": "social-media-post.json",
+      "count": 7972
+    },
+    {
+      "slug": "infographic-edu-visual",
+      "title": "Infographic / Edu Visual",
+      "file": "infographic-edu-visual.json",
+      "count": 531
+    },
+    {
+      "slug": "youtube-thumbnail",
+      "title": "YouTube Thumbnail",
+      "file": "youtube-thumbnail.json",
+      "count": 203
+    },
+    {
+      "slug": "comic-storyboard",
+      "title": "Comic / Storyboard",
+      "file": "comic-storyboard.json",
+      "count": 402
+    },
+    {
+      "slug": "product-marketing",
+      "title": "Product Marketing",
+      "file": "product-marketing.json",
+      "count": 4715
+    },
+    {
+      "slug": "ecommerce-main-image",
+      "title": "E-commerce Main Image",
+      "file": "ecommerce-main-image.json",
+      "count": 509
+    },
+    {
+      "slug": "game-asset",
+      "title": "Game Asset",
+      "file": "game-asset.json",
+      "count": 545
+    },
+    {
+      "slug": "poster-flyer",
+      "title": "Poster / Flyer",
+      "file": "poster-flyer.json",
+      "count": 642
+    },
+    {
+      "slug": "app-web-design",
+      "title": "App / Web Design",
+      "file": "app-web-design.json",
+      "count": 195
+    },
+    {
+      "slug": "others",
+      "title": "Uncategorized",
+      "file": "others.json",
+      "count": 1045
+    }
+  ]
+}

scripts/setup.js

@@ -0,0 +1,119 @@
+#!/usr/bin/env node
+/**
+ * setup.js - Downloads/updates the JAE Image Skill reference library.
+ *
+ * Fully dynamic: reads manifest.json first to discover all categories.
+ * New/renamed/removed categories are handled automatically — no hardcoding.
+ *
+ * Usage:
+ *   node scripts/setup.js           # Download missing files only
+ *   node scripts/setup.js --force   # Force re-download all references
+ *   node scripts/setup.js --check   # Auto-update if stale (> 24h)
+ *
+ * Override source if needed:
+ *   JAE_IMAGE_SKILL_BASE_URL=https://example.com/references node scripts/setup.js --force
+ */
+
+import { existsSync, mkdirSync, statSync, writeFileSync, readFileSync, readdirSync, unlinkSync } from 'fs';
+import { join, dirname } from 'path';
+import { fileURLToPath } from 'url';
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const refsDir = join(__dirname, '..', 'references');
+const stampFile = join(refsDir, '.last-updated');
+
+const BASE_URL = (process.env.JAE_IMAGE_SKILL_BASE_URL || 'https://jaeswift.xyz/skills/JAE-image-skill/references').replace(/\/$/, '');
+const STALE_HOURS = Number(process.env.JAE_IMAGE_SKILL_STALE_HOURS || 24);
+
+function isStale() {
+  if (!existsSync(stampFile)) return true;
+  const ts = parseInt(readFileSync(stampFile, 'utf8').trim(), 10);
+  return (Date.now() - ts) / 1000 / 3600 > STALE_HOURS;
+}
+
+async function fetchText(url) {
+  if (typeof fetch !== 'function') {
+    throw new Error('This installer requires Node.js 18+ with global fetch. Node 20+ is recommended.');
+  }
+  const res = await fetch(url);
+  if (!res.ok) throw new Error(`HTTP ${res.status} — ${url}`);
+  return res.text();
+}
+
+async function setup() {
+  const args = process.argv.slice(2);
+  const forceMode = args.includes('--force');
+  const checkMode = args.includes('--check');
+
+  if (checkMode && !isStale()) {
+    return;
+  }
+
+  if (!existsSync(refsDir)) mkdirSync(refsDir, { recursive: true });
+
+  const label = forceMode ? 'Updating' : 'Downloading';
+  console.log(`[setup] ${label} JAE Image Skill reference library from ${BASE_URL} ...`);
+
+  let categories;
+  try {
+    const manifestText = await fetchText(`${BASE_URL}/manifest.json`);
+    const manifest = JSON.parse(manifestText);
+    categories = manifest.categories;
+    if (!Array.isArray(categories) || categories.length === 0) throw new Error('manifest has no categories');
+    writeFileSync(join(refsDir, 'manifest.json'), manifestText, 'utf8');
+    console.log(`  manifest: ${categories.length} categories, ${manifest.totalPrompts} prompts total`);
+  } catch (err) {
+    console.warn(`[setup] Could not fetch manifest: ${err.message}`);
+    console.warn('[setup] Falling back to existing local manifest...');
+    const localManifest = join(refsDir, 'manifest.json');
+    if (!existsSync(localManifest)) {
+      console.error('[setup] No manifest available. Install from the bundled repo/archive or retry with --force.');
+      process.exit(0);
+    }
+    categories = JSON.parse(readFileSync(localManifest, 'utf8')).categories;
+  }
+
+  const validFiles = new Set([...categories.map(c => c.file), 'manifest.json', '.last-updated', '.gitkeep']);
+  if (forceMode && existsSync(refsDir)) {
+    for (const f of readdirSync(refsDir)) {
+      if (!validFiles.has(f)) {
+        unlinkSync(join(refsDir, f));
+        console.log(`  removed stale: ${f}`);
+      }
+    }
+  }
+
+  let downloaded = 0, skipped = 0, failed = 0;
+  for (const cat of categories) {
+    const dest = join(refsDir, cat.file);
+    if (!forceMode && existsSync(dest) && statSync(dest).size > 100) {
+      skipped++;
+      continue;
+    }
+    process.stdout.write(`  → ${cat.file} (${cat.title}, ${cat.count} prompts) ... `);
+    try {
+      const text = await fetchText(`${BASE_URL}/${cat.file}`);
+      JSON.parse(text);
+      writeFileSync(dest, text, 'utf8');
+      console.log('✓');
+      downloaded++;
+    } catch (err) {
+      console.log(`✗ (${err.message})`);
+      failed++;
+    }
+  }
+
+  if (failed === 0) writeFileSync(stampFile, String(Date.now()), 'utf8');
+
+  if (downloaded > 0) {
+    console.log(`[setup] Done! ${downloaded} file(s) ${forceMode ? 'updated' : 'downloaded'}. JAE Image Skill is ready.`);
+  } else if (skipped === categories.length) {
+    console.log('[setup] All references up to date. Use --force to refresh.');
+  }
+  if (failed > 0) console.warn(`[setup] ${failed} file(s) failed. Run again to retry.`);
+}
+
+setup().catch(err => {
+  console.warn('[setup] Warning (non-fatal):', err.message);
+  process.exit(0);
+});