Hunts open on a notebook. A hunt is now a folder in Explorer — Notebook, Results, Flow and Entities views behind one switcher, plus its report and chats — and it opens on the notebook: the steps as an ordered list of cells, each saying what it found (12 rows, No rows, Failed) with the output underneath. Sections group cells, notes sit between them, and the flow graph is one click away rather than the landing page. See Hunts.
Fix failed steps in place. A failed cell shows its error with Fix with Scout, Edit and Retry; cells that failed for the same reason say so. Scout's fix is a new cell that replaces the original (v2 of N), so the record stays intact. Scout's schema probes fold into a Discovery strip.
Verdicts are a log. The verdict cell offers Confirmed, Nothing found, Inconclusive and False positive, with the server's guard rails explained inline. Recording a verdict no longer closes the hunt: keep hunting, record again, and every entry stays. Scout adds draft verdicts as the hunt progresses.
Needs you, in order. The notebook's right pane lists what is waiting — failed cells, checkpoints, approvals, the verdict — ranked by the server, with Next to walk through them and Approve & run on checkpoints from the list. The hunts list and the triage lane use the same ranking.
Edit cells where they are. Query cells edit inline with the same editor as the query workspace; insert above or below, move up or down, archive, or delete a cell that never ran. Keyboard: j / k, a / b, ⇧⏎. New cells appear the moment you click.
The dock is a rail of views. One view at a time — Workspace, Chats, Queries, Library, Scout — instead of a stack of panels. Workspace is a tree: every hunt is a folder that expands into its notebook, flow, results, entities, evidence, report and chats, with pins, needs-you counts and a find field; Stores and Schedules sit beneath. Every view ends with an Ask Scout strip. See Explorer overview.
Library, for this hunt. Inside a hunt the dock's Library view leads with Suggested for this hunt, each template with Add to hunt.
Hunt rows in the Activity Feed show a per-step Cells strip, a derived state (Checkpoint waiting, N failed, Needs verdict…), a Needs you count, and whether a hunt with a verdict is open or closed.
Closed hunts and hunts you cannot edit say so in the notebook, with Reopen hunt for the owner.
The hunt header counts cells, matching the notebook; the flow's node cards say what a step found rather than that it ran, and a dashed v2 edge links a replaced step to its replacement.
Results opens even before any step has produced rows, and lands on the cell you came from; Show scaffold and Show archived reveal rows the grid hides by default.
The notebook's side pane is the same resizable panel the entity explorer uses, and becomes a sheet on narrow screens.
Flagging a result row from the flag column now writes one organization-wide flagged tag, so every analyst sees the same flags. Previously the first click was rejected.
Starting a new chat from the dock always opens a fresh conversation and no longer steers the dock to an empty Scout view.
Template suggestions appear again when you type in a new query cell.