Inspiration
A video project can look ready until you open the audio folder and find two voiceovers with no clear approval. The client brief says “use the approved voiceover,” but the files are named voiceover-v1.wav and voiceover-v2.wav. Guessing can mean cutting the reel twice. We wanted editors to catch that kind of gap before they start editing.
What it does
Shoot Handoff Assistant checks a client brief against the files in a UGREEN AI NAS project folder. Ask, “Is the Candle Launch Reel ready for editing?” and it returns a checklist with Found, Missing, or Needs review for each required asset. It shows matching file paths and explains what needs clarification. If two voiceovers match and approval is unclear, it lists both and asks the editor which one the client approved.
How we built it
We used Codex and the official Uliya Skill Kit to build a read-only skill. The workflow uses UGOS file capabilities to locate the project, find the brief, and inventory the folder. It reads plain-text briefs through the file expert and uses the document expert to extract requirements from supported document formats. Uliya compares those requirements with the returned file inventory and builds the checklist in chat.
Challenges we ran into
We had to define what counts as enough evidence for each status. A newer filename gives no evidence of client approval, and an unlisted subfolder can hide the file we’re looking for. We made approval uncertainty and incomplete inventory explicit reasons to return Needs review. The skill uses Missing only when the project inventory is complete.
Accomplishments that we're proud of
The official Uliya checker returned static-pass with zero errors and warnings. Four simulated tests passed, covering an approved file, a missing file, two versions with unclear approval, and an incomplete inventory. In the last case, the skill identified the exact Audio/ folder it couldn’t inspect and told the editor what to check next.
What we learned
An editor needs to know why an asset has a particular status. Showing the exact path, the approval evidence, or the folder that couldn’t be checked turns a readiness result into something the editor can act on.
What's next for Shoot Handoff Assistant
Our next step is live validation on UGREEN AI NAS. We’ll rerun the core tests and verify how the workflow handles read permissions, real file paths, and incomplete listings. We also want to test a larger handoff with several required assets, where one brief can produce a mix of Found, Missing, and Needs review.
Built With
- markdown
- openai-codex
- ugos
- ugreen-ai-nas
- uliya
- uliya-skill-kit
