Shoot-to-Delivery

From "the shoot is done" to a client delivery that provably matches the contract: picks, contract, invoice and model releases checked on the NAS before anything leaves the studio.

Comment

Inspiration

We run a small photo business: shoots for clients, a camera shop catalogue, and an AI "company" of agents that already hands finished work to its owner with a README and a checksum list. Every photographer knows the last mile hurts: re-reading the contract, counting picks, and wondering whether the person in frame 23 ever signed a release.

What it does

Shoot-to-Delivery is a UGOS Skill for photographers and small studios. Say "Prepare the delivery for the Kovacs wedding." It finds the shoot and the photos you tagged Pick. It reads the contract, the invoice and the model releases, and shows one gap report: image count, format, resolution, deadline, people without a release or with a release that does not cover the agreed use, technical warnings, and payment. Nothing is written until you choose. Then it hashes the originals, copies the picks into a new delivery folder, verifies every copy and saves DELIVERY-NOTE.md. It can also create a client album.

How we built it

We built it with the official UGOS Skill Creation Tool (uliya-skill-kit 0.7.3). The parent Skill only orchestrates six self-contained delegations to file-agent, photo-agent, document-agent and knowledge-agent, and passes each step only the real results the experts return. All five delegation states are handled. Every write waits for platform approval, and the Skill stops on name collisions.

Challenges

Release checks are only as good as face naming, so unnamed people are flagged for a manual look instead of guessed. Excluding flagged photos can create a new shortfall, so the Skill recomputes the count and asks again.

What we learned

The value is in the cross-check between domains (photos × documents × files), not in any single capability.

What's next

Finalist NAS testing with a prepared shoot, contract, invoice and signed releases. We will then tune the vision prompts and the extraction of Hungarian and English contracts.

Built With

  • markdown
  • ug-document
  • ug-filemgr
  • ug-filesystem
  • ug-knowledge
  • ug-photo
  • ugos
  • uliya-skill-kit
Share this project:

Updates

Submission history