agentleFS
astriaaiastriaai/skills/virtual-try-on/SKILL.md

virtual-try-on

Skill2 starsChanged 26 days agoMITView on GitHub

Generate on-model fashion photos by putting a supplied garment, shoes, bag, jewelry, or accessory on a supplied or selected person. Use for "put this dress on a model", "show me wearing this", virtual try-on, or the same outfit and model in several locations. Product-only photography uses product-photoshoot. Respect an explicitly chosen provider; styling advice and prompt writing do not authorize generation.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use astria-api for schemas, reference creation and tokens, model discovery, HTTPS media transport, idempotency, workspace scope, and …

    line 12
  2. Very specific:

    Use supplied person and product references.

    line 17
  3. Very specific:

    If the user says "on a model" without specifying a person, select an appropriate available public model reference through list_references …

    line 21
  4. Distinctive:

    Ask only for a missing input or decision that changes the requested outcome.

    line 26
  5. Distinctive:

    Create new references only for supplied subjects that lack a reusable reference.

    line 30
  6. Very specific:

    Call list_models and use the explicit model or catalog default.

    line 34
  7. Very specific:

    Check every returned prompt with get_prompt using its ID.

    line 50
  8. Distinctive:

    When inspection is available, compare identity, garment details, and placement with the sources.

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Virtual try-on

Generate the requested still photographs with Astria, keeping the chosen person and wearable item consistent. A generated try-on illustrates appearance; do not infer measurements or promise physical fit.

Use astria-api for schemas, reference creation and tokens, model discovery, HTTPS media transport, idempotency, workspace scope, and completion handling.

Resolve the person and item

Use supplied person and product references. Resolve named references through list_references and inspect them with get_reference, using the returned class names. Reuse the same person and item IDs throughout the set.

If the user says "on a model" without specifying a person, select an appropriate available public model reference through list_references with gallery: true; inspect it before use. If they want a newly invented model, avatar can supply one within that requested scope. "Show me wearing this" requires usable photos or a saved reference of that person; do not substitute a stranger. Ask only for a missing input or decision that changes the requested outcome.

Media must be accessible HTTPS URLs. Use the host's upload capability or the optional authenticated CLI for local inputs; a filename is not a reference URL. Create new references only for supplied subjects that lack a reusable reference.

Generate the requested views

Call list_models and use the explicit model or catalog default. Compose generate_images.text with both reference tokens, such as <faceid:123:1> woman wearing <faceid:456:1> dress, using the actual returned IDs and class names. Describe the item's visible cut, fabric, color, and design along with the requested pose and setting; do not invent hidden product views.

For different locations, submit a distinct prompt per location with the same person and product tokens. Preserve requested count and aspect ratio. If these are unspecified, make one portrait-oriented image. Use a fresh idempotency key per distinct submission and retain it for an identical retry.

A request for a video should use astria-api's supported video workflow with these same references; do not silently deliver still images instead.

Finish the set

Check every returned prompt with get_prompt using its ID. Present the requested completed assets and clearly identify failures or pending work. When inspection is available, compare identity, garment details, and placement with the sources. Do not claim exact fit or submit unrequested correction jobs.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.