generated: '2026-08-29'
method: generated
source: openapi/superb-ai-mlops-platform-openapi.json
api: Superb AI MLOps Platform API
base_url: https://api.bdai.superb-ai.com
note: >-
  Four packaged Agent Skills covering the marquee flows of the Superb AI MLOps Platform API.
  Every operationId referenced in every skill was verified against the published OpenAPI document
  before it was written — none is invented. Each skill also carries the cross-cutting rules from
  conventions/superb-ai-conventions.yml (cursor pagination, Idempotency-Key on bulk POSTs,
  branch-on-error.code, unheadered 429 back-off) and flags the operations the contract itself
  declares irreversible. Superb AI publishes no AGENTS.md or skills of its own; these are
  generated by API Evangelist from the contract.
skills:
- name: superb-ai-ingest-image-dataset
  file: superb-ai-ingest-image-dataset.md
  summary: >-
    Create an image dataset and upload assets through the two-phase presigned upload, then verify
    and optionally embed them for similarity search.
  operation_count: 7
- name: superb-ai-label-and-version-project
  file: superb-ai-label-and-version-project.md
  summary: >-
    Stand up a labeling project over a dataset — classes, attributes, members — move assets in,
    assign and review in bulk, and freeze a version.
  operation_count: 13
- name: superb-ai-export-training-data
  file: superb-ai-export-training-data.md
  summary: >-
    Export a frozen version as a COCO, YOLO or Delta bundle and download it inside the 7-day
    lifecycle window, handling the implicit X-From-Cache idempotency layer.
  operation_count: 6
- name: superb-ai-train-and-deploy-model
  file: superb-ai-train-and-deploy-model.md
  summary: >-
    Train on a frozen version, deploy the resulting model to an inference endpoint, run
    predictions, and use the ZERO foundation models with no training at all.
  operation_count: 15