images
Skill4 starsChanged 20 days agoMITView on GitHub
Set up or work on content images (png/jpg/webp) — remote-hosted originals, build-time optimization, responsive images, a post cover image, and image SEO (Google Discover / image sitemaps). Use when the user wants to add content images, host originals on a bucket/CDN (R2, S3…), add a post cover, cache optimized images in CI, or change how the <Image> component / Markdown images work.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- sitetiller/ai-visibility-skillstheir own version, first seen Nov 15, 2026
- aspose-words/agentic-net-examplestheir own version, first seen Sep 9, 2026
- samarth777/modal-skillstheir own version, first seen Feb 5, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- aspose-words/agentic-net-examples★ 4same name
- samarth777/modal-skills★ 1same name
- sitetiller/ai-visibility-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- analyticsSkill
- brandSkill
- og-imagesSkill
- redirectsSkill
- runSkill
- structured-dataSkill
- update-depsSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 44
Every image emits a schema.org ImageObject license JSON-LD (license, acquireLicensePage, creditText, copyrightNotice, creator) from …
- Distinctive: line 61
Use a full URL on the origin so it gets optimized.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Content images
Content images (photos: png/jpg/webp) are optimized at build by astro:assets: resized, re-encoded to AVIF with a WebP fallback, content-hashed into dist/_astro/. image.layout: "constrained" in astro.config.mjs makes them responsive by default. There's no runtime CDN transform.
Originals live in one of two places:
- Committed: a file in
src/images/, imported. Builds never fetch. - Remote origin: a bucket/CDN (e.g. an R2 bucket on a custom domain) set as
site.platform.imagesCDNHostinsrc/site.config.ts. Astro downloads and optimizes at build. It's independent ofsite.platform.deploy, and it's committed config rather than an env var so every build sees it, including host previews.
Key files:
src/site.config.ts:site.platform.imagesCDNHost, the one source for the origin. It also feedsimage.remotePatternsinastro.config.mjs.src/components/Image.astro: the only image component. Wrapsastro:assets<Picture>.src/lib/images.ts:publishedImageUrl(src), the URL of a 1200px WebP rendition under/_astro. Everything that hands out an image URL uses it.
The origin is never published
The origin is a build input. A leaked origin gets crawled in place of the site's own URLs, and anyone can hotlink the full-size originals. So JSON-LD (contentUrl, BlogPosting.image) and the image sitemap use publishedImageUrl, the CSP img-src doesn't list the origin, and npm run check:origin (in CI, after the build) fails if the origin host appears in any file in dist/.
Enable a remote origin
- Set
platform.imagesCDNHost: "https://images.example.com"(a full URL) insrc/site.config.ts. - Upload originals (e.g.
wrangler r2 object put <bucket>/<key> --file=… --remote) and reference them by full URL.
With imagesCDNHost: "", commit images to src/images/ and import them.
Authoring
<Image>: src is an imported asset or a full remote URL. A remote src without width and height is measured at build. priority sets eager loading + fetchpriority=high; use it on the LCP/hero image only.
<Image src="https://images.example.com/photo.jpg" alt="…" width={1200} height={675} priority />
<Image src="https://images.example.com/photo.jpg" alt="…" width={1200} height={675}>
A caption.
</Image>- A caption (the default slot) wraps the image in
<figure>+<figcaption>. - Every image emits a schema.org
ImageObjectlicense JSON-LD (license, acquireLicensePage, creditText, copyrightNotice, creator) fromsite.copyright.images.licenseandsite.organization. Override per image withauthor={{ name, url }}(credits a Person) and/orlicenseUrl="…", or drop it withschema={false}for a photo the site doesn't own (people's photos). - An image Astro can't optimize (off the origin) renders as a plain
<img>, and itscontentUrlis that URL.
Markdown  works with local files (./photo.jpg) and origin URLs. In .mdx, every content route maps img to src/components/MarkdownImage.astro (src/lib/content-components.ts): local and origin images go through <Image>, any other URL stays a plain <img> with no license JSON-LD (the CSP blocks it unless you add its host to img-src in src/headers/headers.config.ts). In .md, Astro optimizes them itself (WebP only, no license JSON-LD). A raw <img> in content is handled the same way (src/markdown/remark-img-tags.ts): in .mdx it goes through MarkdownImage with its attributes; in .md an origin or relative one becomes ![](), keeping only src, alt and title. An <img> inside a <picture> is left alone, as is one in a component or page: on the origin, the CSP blocks it and check:origin fails.
Post cover and Google Discover
A post's optional cover ({ image, alt } in blog frontmatter) is a real photo, separate from the OG card. It renders as the hero (LCP), and its published rendition becomes BlogPosting.image and the <image:loc> in the image sitemap (src/pages/sitemap.xml.ts). Google Discover reads these. og:image stays the OG card.
cover:
image: "https://images.example.com/photo.jpg"
alt: "Describe the photo"imageandaltare both required whencoveris set (schema-enforced).- For Discover, use a landscape photo at least 1200px wide, 16:9, not text-heavy (docs).
max-image-preview:largeis already set. - Use a full URL on the origin so it gets optimized.
- Without
cover,BlogPosting.imagefalls back to the OG card and there's no image-sitemap entry.
Reachability and CI caching
- Images on the configured origin are optimized in every build, so the origin must be reachable at build time.
- Images off the origin (or with
imagesCDNHost: "") render as plain<img>with no build-time fetch, so a down host can't fail the build. - Astro caches optimized images and fonts in
node_modules/.astro/assetsandnode_modules/.astro/fonts.npm ciwipes them, so CI restores both withactions/cache(.github/workflows/ci.yml;lighthouse.ymlrestores the same cache). Cache only these two: the rest ofnode_modules/.astrois cheap to rebuild. - Deploy hosts have their own build cache. Point it at the same two directories, and check they survive the host's install step, since
npm cideletesnode_modules.
Verify
npm run build: the cover is a<picture>with an AVIF<source>and a/_astro/*.webp<img>, each with a srcset of distinct sizes; the JSON-LDimageand/sitemap.xml's<image:loc>are a/_astro/*.webpURL;node_modules/.astro/assets/fills with avif and webp files.npm run check:originpasses.npm run check:sdvalidates the JSON-LD.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

