Inspiration
A photographer can come home with hundreds of images and still miss one detail in the client’s brief. They photographed the kitchen, but every preparation shot is horizontal and the client asked for a vertical one. Wrap Check grew from that problem, the check you want to make before you call the shoot finished.
What it does
Wrap Check compares a selected shoot with a client brief or shot list. It turns the brief into a checklist, asks about unclear requirements, and finds possible matches. For each requirement, it reports Candidate found, Needs confirmation, or No qualifying candidate found. It explains close misses and lists files it could not read.
The photographer can review the findings, save a candidate album and report, then add follow-up photos and run the check again to see what changed. The photographer makes the final call.
How I built it
I built the Skill with UGREEN’s Skill Developer Package and the Uliya Skill Development Standard. Its SKILL.md coordinates UGREEN’s document, file, photo, and knowledge agents. The planned workflow reads the brief, inventories the selected JPEGs, checks metadata and photo search results, and sends a shortlist to a vision model for details that need visual judgment. The report records the findings so a later run can compare new photos with the previous check.
Challenges I ran into
A search result cannot prove that a required shot is missing. Files may be unreadable or unindexed, and a vision model may struggle with small details such as sign text. We designed the report to show those limits instead of presenting a guess as a fact. We also had to keep visual checks bounded so a large shoot would not require analyzing every image with the model.
Accomplishments that I'm proud of
The clearest part of Wrap Check is its explanation of near matches: “three preparation shots found, all horizontal” gives a photographer something useful to act on.
What I learned
Checking a shoot against a brief is different from finding photos by subject. A useful result has to account for conditions such as orientation, count, and readable text, while showing how much of the collection was actually checked. We also learned that a saved report makes the Skill more useful on a second run, when the photographer needs to know whether follow-up photos closed the gaps.
What’s next for Wrap Check
Hospitality photography is my first planned NAS test. If selected as a finalist, I'd like to test planted gaps and almost right images, add follow-up shots to check the comparison, and measure accuracy and time on the assigned UGREEN NAS. We’ll also verify local model routing, photo indexing, and album and report creation. Property and product shoots are possible next uses; they have not been tested yet.
Log in or sign up for Devpost to join the conversation.