ugc-product-profile
Skill4 starsChanged 5 months agoMITView on GitHub
Step 1 of the ugc-ad pipeline. Analyze a product photo and/or product URL into a structured product profile (classification, dimensions, packaging, exact usage steps) and a clean product.png reference. Use standalone when you need a reusable product profile, or as the first step of a UGC ad. Brand-agnostic.
- Sends data out
In this repository7 other files it gives its agents.
- ugc-adSkill
- ugc-base-characterSkill
- ugc-briefSkill
- ugc-enhanceSkill
- ugc-multicut-scriptSkill
- ugc-storyboard-sheetSkill
- ugc-videoSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
If PRODUCT_URL points at an image, OR the product page has a hero image URL: call media_import_url(url=<image-url>) → media_id.
- Very specific: line 18
If PRODUCT_IMAGE is a local file:
- Distinctive: line 47
Do NOT invent facts. If the URL/image is ambiguous, state the assumption in the profile.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
ugc-product-profile
Turns raw product input into a structured profile + a clean reference image the rest of the UGC pipeline locks onto.
Inputs
PRODUCT_IMAGE(path) and/orPRODUCT_URL(https) — at least one required.DEST(path) — the project folder to write into.
Steps
- Acquire the product image as a Higgsfield media_id.
- If
PRODUCT_URLpoints at an image, OR the product page has a hero image URL: callmedia_import_url(url=<image-url>)→media_id. - If
PRODUCT_IMAGEis a local file:- In an Apps-UI client (Claude desktop/web): call
media_upload_widgetso the user picks the file; use the returnedmedia_id. - In Claude Code / CLI (filesystem access): call
media_upload(filename, content_type)for a presignedupload_url+media_id, thencurl -X PUT -H "Content-Type: <type>" --data-binary @<file> "<upload_url>", thenmedia_confirm(media_id, type:"image"). (Validated 2026-06-22.)
- In an Apps-UI client (Claude desktop/web): call
- Save a copy as
DEST/product.png(download the imported/source image withcurl -fsSL <url> -o "$DEST/product.png"when a URL is available).
- If
- Extract product facts. From the URL page content (fetch it) and/or the image, determine:
- classification (what kind of product it is),
- physical dimensions / form factor,
- packaging details (container, label, colors, logo placement),
- exact usage steps (how a person physically uses it — the motion the actor will perform).
Write
DEST/product-profile.mdwith this structure:# Product Profile: <name> - Classification: ... - Dimensions / form factor: ... - Packaging: ... - Brand cues (colors, logo, label): ... - Exact usage steps: 1. ... 2. ... - product media_id: <media_id>
Output
DEST/product-profile.md + DEST/product.png + the product media_id. Report the media_id back to the caller.
Notes
- Do NOT invent facts. If the URL/image is ambiguous, state the assumption in the profile.
- This step does no paid video generation.
media_import_urland a page fetch are the only external calls.
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.

