メニュー
Generate and edit images for ArtStroy articles using the Gemini image generation API. Use when creating cover images, OG images, inline illustrations, or editing existing images for article assets.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Pull weekly traffic data from Plausible Analytics for artstroy.net, produce a structured digest with top articles, category breakdown, referrer mix, and trend signals. Posted to the root company issue weekly.
Cross-post published ArtStroy articles to dev.to and Hashnode with correct canonical URLs, metadata mapping, and formatting. Handles API integration for automated posting and creates manual-assist drafts for platforms without API access.
Scan ArtStroy npm/bun dependencies for known vulnerabilities. Ranks CVEs by CVSS severity, creates vuln-fix issues with deadlines, and verifies fixes post-deploy. Run monthly and on any new dependency addition.
Manage and verify the ArtStroy CI/CD pipeline. Covers the GitHub Actions deploy workflow, SSH-based Hetzner deployment, Docker Compose orchestration, and post-deploy health checks. Used by DevOps Engineer for deploy operations and pipeline maintenance.
Manage ArtStroy's Docker Compose configuration, Hetzner server provisioning, environment variable management, and infrastructure-as-code practices. Used by DevOps Engineer for all infrastructure changes.
Compile new ArtStroy articles into a Markdown newsletter draft ready for Buttondown. Runs weekly; sends draft to human for final review before dispatch.
| name | image-gen |
| description | Generate and edit images for ArtStroy articles using the Gemini image generation API. Use when creating cover images, OG images, inline illustrations, or editing existing images for article assets. |
Generate or edit images using the Gemini image generation API (GEMINI_API_KEY from .paperclip.yaml inputs.env).
The generation script is invoked via uv:
Generate a new image:
uv run scripts/generate_image.py \
--prompt "your image description" \
--filename "output.png" \
--resolution 1K
Edit an existing image:
uv run scripts/generate_image.py \
--prompt "edit instructions" \
--filename "output.png" \
-i "/path/to/input.png" \
--resolution 2K
Multi-image composition (up to 14 source images):
uv run scripts/generate_image.py \
--prompt "combine these into one scene" \
--filename "output.png" \
-i img1.png -i img2.png -i img3.png
With a specific aspect ratio:
uv run scripts/generate_image.py \
--prompt "portrait photo" \
--filename "output.png" \
--aspect-ratio 9:16
Resolutions: 1K (default), 2K, 4K
Aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9
Omit --aspect-ratio to let the model choose freely. Use it when exact dimensions are required (avatars, covers, OG images).
| Asset | Format | Dimensions | Aspect Ratio | Notes |
|---|---|---|---|---|
| Cover image | AVIF | 1920×1080 | 16:9 | No text; dark aesthetic preferred |
| OG image | PNG | 1200×630 | ~1.91:1 | Must include article title text |
| Inline illustration | PNG | up to 2K | free | Match article's visual theme |
Generate at the correct aspect ratio, then convert/resize as needed:
./imgs/cover.avif./imgs/og.pngUse timestamps in filenames for intermediate outputs:
2026-05-25-14-30-00-cover-draft.png
Final outputs must match the contract paths exactly:
src/content/articles/{slug}/imgs/cover.avif
src/content/articles/{slug}/imgs/og.png
Requires GEMINI_API_KEY env var. This is declared in .paperclip.yaml inputs.env and injected by the Paperclip runtime at execution — do not hardcode or print it.
The script prints a MEDIA: line pointing to the saved file. Report the path from that line. Do not read the image back into the conversation — just report where it was saved.
./imgs/. One retry is reasonable; more than two retries should escalate to human review.