operate-content-workflow

Import the local image as an asset, then configure the video generator with a configure_video_shot operation using the animate-still strategy and an animate-still-plan prompt. The operation validates the strategy, binds the image source, and selects a compatible model mode atomically.

Create two image generator nodes where the last frame uses image-to-image with the first frame as its sole reference, then submit one configure_video_shot operation with a first-last-frame-plan. The operation binds both images as firstFrame and lastFrame to an interpolation-capable video model.

Use a Prompt node only when one verbatim text fragment must be shared by at least two consumers and updated in a single edit, such as a brand voice block. Single-use prompts belong directly on the generator's prompt or promptPlan field.

Inspect the readiness, diagnostics, and capabilities views to classify the cause: invalid graph topology, unsupported mode or ratio, provider rejection, or reference mismatch. Repair the graph or adjust the configuration, then retry only after confirming inputs remain valid.

Yes. Executing run with action prepare opens the plugin's global confirmation dialog, and generation only starts after the user approves it because runs may consume quota. Edit batches apply atomically without per-operation confirmation.

No. All workflow changes must go through the content_creation_search and content_creation_execute tools, which validate and apply revision-bound batches atomically. Direct file edits and calls to media providers outside the plugin are not permitted.