Get Run Review
The full-review manifest (plan A5): source pages via the EXISTING thumbnail endpoint, the per-shape result payloads (INLINE from the run’s own stored result — no artifact or intermediates listing, so no workspace path reaches a client), and the projected trace.
Authorizations
Path Parameters
Response
Successful Response
Everything the full run review renders (frontend spec §4): chrome (the run card), the source surfaces (pages OR sheets — never both), the per-shape result pane payload, and the projected trace.
Deliberately NOT here: any listing of the run's workspace files. The review shows the run's own output and derived views of it; raw workspace browsing would put storage paths in front of a customer.
RunCard-shaped (snake_case): the drawer/board consume this.
Per-shape result payloads for the review's result pane (frontend RunOutput). Only the fields the run's shape produced are populated.
Page surfaces (pdf/image docs). EMPTY for spreadsheet runs — no page images exist for xlsx, matching production; the source pane is sheets.
Sheet surfaces (spreadsheet docs): per-sheet block HTML + cell-range bounds for the sheet-tab source pane. EMPTY for page docs.
Workspace filename → original input filename. The sandbox anonymizes the staged document to .; this maps those names back for display.