Alibaba's Wan Adds 'Agent Skills' for Sharing Video Workflows

An era of "using" workflows rather than "building" them

Instead of writing a prompt from scratch every time you want to make a video, what if you could just grab a recipe someone else already built? Alibaba's video generation platform Wan unveiled exactly this on its official account on August 14 (UTC). It's called "Agent Skills." Users can browse community-built workflows and use them right away, or create their own combinations and share them with the world.

What's changed

Screenshots of the release show a new "Skills" tab added to the Wan platform's left-hand menu, alongside the existing Generate, Canvas, Playground, and Assets tabs. At the top, Explore and Skills sub-tabs sit side by side, separating a browsing view for skills made by other users from a management view for skills the user has uploaded themselves. The official announcement stated that the feature "turns ideas into reusable creative tools, from short drama generators to music video generators."

A skill is less a single finished prompt set and more an automated procedure with predefined conditions and steps. For example, opening a "short drama generator" skill likely runs repetitive tasks — character setup, scene transitions, dialogue insertion — in a pre-arranged sequence. However, the announcement did not detail the specific internal logic of any individual skill or how difficult it is to build one.

What Wan has been building recently

Wan previously released its video generation model Wan3.0 in public beta on August 6, touting native 30-second video generation and "reality-grade" rendering. That release also introduced the Omni-Reference feature, which accepts text, images, audio, and video as well as documents and spreadsheets as reference input. Then on August 11, Wan released the open-source character animation model Wan-Animate-2, adding multi-character animation and real-time streaming generation.

The input UI for Wan3.0 was also shown alongside this announcement. The Omni input panel supports 1080P resolution, Smart Ratio, and 30-second duration options, with feature buttons listed for Reference to Video, First&Last Frame, Video Edit, Image Generation, and Agent. Skills appears to be a higher-level layer that ties these individual generation features together into reusable workflows.

How to try it

Clicking the "Skills" tab in the left-hand menu of the Wan platform brings up a list of skills registered by the community. The announcement described two paths — "Explore skills" for browsing skills, and "Upload your own" for registering a personal workflow.

1. Go to the Wan platform and click Skills in the left-hand menu.
2. In the Explore tab, browse the list of published skills, such as short drama generators and music video generators.
3. Selecting a desired skill takes you to a generation screen with that workflow already applied.
4. If you've built your own combination, you can share it with other users via the Upload path.

Who can try what

The announcement did not specify country-by-country usage conditions or pricing tiers. However, a "Premium" label and a separate credit (✦) indicator appear at the top of the screen, suggesting a tiered usage structure similar to Wan's existing services.

The range of possible uses can be inferred. A creator making short web dramas every week could pull up a short drama skill with character setup and scene transitions already built in, and simply swap in a new script. A musician could layer a song and lyric concept onto a music video generator skill to produce a draft. Repetitive work that once required designing a new prompt each time can now be reduced to a few clicks.

What this means

Competition among video generation models is increasingly shifting from image quality and length toward the question of "who can make repetitive workflows easier to automate." Rather than a battle over individual model performance, the Skills feature Wan just launched functions more like a distribution structure — letting one person's prompt-combination know-how be directly reused by others. Following Wan3.0's native 30-second generation and Wan-Animate-2's real-time streaming animation, this latest feature suggests Wan is aiming to grow not just the model's raw performance but the entire workflow ecosystem built on top of it.