feat(webapp): completed tool calls leave no row in the chat

The answer is the prose and the cards, never the input/output plumbing:
a landed call disappears (the pill was its only trace), and only a
FAILED call keeps its error row. Gallery example reworked accordingly.

Author: kathiekiwi

apps/webapp/app/components/dashboard-agent/DashboardAgentMessages.tsx

@@ -106,10 +106,11 @@ function blocksFor(part: UIMessage["parts"][number]): unknown[] | null {
  * Everything the panel styles itself is handled here; the rest falls through to
  * the shared `renderPart` so agent output still looks the same across the app.
  * The differences: text is always the rendered markdown (no raw toggle) at the
- * dashboard's default size, and an in-flight tool call is a pending pill rather
- * than a tool row streaming its input JSON — the input is the agent's business,
- * and watching it arrive character by character only to have it replaced by a
- * card is noise. Once the call lands, the part renders exactly as it always did.
+ * dashboard's default size, and tool calls never show their mechanics — while
+ * running they are a pending pill ("Reading the queue…"), and once they land
+ * they leave NO row at all: the answer is the prose and the cards, not the
+ * input/output plumbing. The one exception is a FAILED call, which keeps its
+ * error row — a silent failure would read as the agent ignoring the question.
  * Citations are handled a level up, where a run of them can be grouped into one
  * row.
  */
@@ -127,8 +128,12 @@ function renderDashboardPart(part: UIMessage["parts"][number], i: number) {
     return p.text ? <ChatText key={i} text={p.text} /> : null;
   }
 
-  if (type.startsWith("tool-") && IN_FLIGHT_TOOL_STATES.has(p.state ?? "")) {
-    return <ChatPendingTool key={i} label={`${toolPendingLabel(type.slice(5))}…`} />;
+  if (type.startsWith("tool-")) {
+    if (IN_FLIGHT_TOOL_STATES.has(p.state ?? "")) {
+      return <ChatPendingTool key={i} label={`${toolPendingLabel(type.slice(5))}…`} />;
+    }
+    if (p.state === "output-error") return renderPart(part, i);
+    return null;
   }
 
   return renderPart(part, i);

apps/webapp/app/routes/storybook.agent-ui/manifest.ts

@@ -284,10 +284,9 @@ export const MANIFEST: GallerySection[] = [
     group: "messages",
   },
   {
-    sectionId: "messages-tool-expanded",
-    title: "Tool call expanded on its output",
+    sectionId: "messages-tool-completed",
+    title: "Completed tool call — no row, just the answer",
     group: "messages",
-    expandText: "output",
   },
   {
     sectionId: "messages-error-retry",

apps/webapp/app/routes/storybook.agent-ui/route.tsx

@@ -736,7 +736,9 @@ const STATES: Record<string, React.ReactNode> = {
   "messages-reasoning": <MessageHarness chatId={demoId("investigate-streaming")} />,
   "messages-tool-in-flight": <MessageHarness chatId={demoId("base-tool-in-flight")} />,
   "messages-tool-pending-pills": <PendingPillsHarness />,
-  "messages-tool-expanded": <MessageHarness chatId={demoId("base-tool-in-flight")} />,
+  // The same chat whose pending pill is shown above, but with the call landed:
+  // the tool row is gone entirely — only the prose (and any cards) remain.
+  "messages-tool-completed": <MessageHarness chatId={demoId("docs-answer")} />,
   "messages-error-retry": <MessageHarness chatId={demoId("base-error-retry")} withError />,
   // Two turns only: the resumed chat's third turn is a live `chart` block, and
   // the real AgentChart has no environment to query outside a project route.