diff --git a/src/tui/diff-rows.test.ts b/src/tui/diff-rows.test.ts
index ee8e72149..9e5f958da 100644
--- a/src/tui/diff-rows.test.ts
+++ b/src/tui/diff-rows.test.ts
@@ -120,6 +120,64 @@ describe("diff transcript rows", () => {
     }, WIDE)
   })
 
+  test("a task/dispatch call paints a sentence, never the full spawn JSON (CL-5762)", async () => {
+    const brief = {
+      agent: "explore",
+      description: "map callers of leaveObserve",
+      prompt: "Find every call site of leaveObserve.\nReport paths and line numbers.",
+      intent: "explore",
+      maxTurns: 40,
+      success_criteria: ["list call sites", "note tests"],
+      do_not: ["edit code", "open PRs"],
+    }
+    const args = JSON.stringify(brief)
+    const row = toolCallRow({ name: "task", arguments: args })
+
+    // Structural: summary set, not raw args; detail expands with real newlines.
+    expect(row.summary).toBe("map callers of leaveObserve")
+    expect(row.verb).toBe("Explore")
+    expect(row.text).toBe(args) // clipboard still has raw; paint must not use it
+    expect(row.summary).not.toContain("success_criteria")
+    expect(row.summary).not.toContain("maxTurns")
+    // Expanded body uses real line breaks, not literal \\n escape sequences.
+    const detailPlain = (row.detail ?? [])
+      .map((line) => line.map((s) => s.text).join(""))
+      .join("\n")
+    expect(detailPlain).toContain("Find every call site of leaveObserve.")
+    expect(detailPlain).toContain("Report paths and line numbers.")
+    // A pretty-printed JSON dump would keep \\n inside the prompt string.
+    expect(detailPlain).not.toContain("\\n")
+    expect(detailPlain).toContain("list call sites")
+
+    await withTestRenderer(async (h) => {
+      const shell = createAppShell(h.renderer, shellOpts)
+      appendStreamRow(shell, row)
+      await settle(h)
+      const frame = h.captureCharFrame()
+      expect(frame).toContain("map callers of leaveObserve")
+      expect(frame).not.toContain('"maxTurns"')
+      expect(frame).not.toContain('"success_criteria"')
+      expect(frame).not.toContain(args.slice(0, 40))
+    }, WIDE)
+  })
+
+  test("a task without description still collapses — falls back to prompt, not raw JSON", () => {
+    const prompt = "Find every call site of leaveObserve and report them."
+    const args = JSON.stringify({
+      agent: "explore",
+      prompt,
+      intent: "explore",
+      success_criteria: ["list sites"],
+    })
+    const row = toolCallRow({ name: "task", arguments: args })
+    expect(row.summary).toBeDefined()
+    expect(row.summary!.length).toBeGreaterThan(0)
+    expect(row.summary).not.toContain("success_criteria")
+    expect(row.summary).not.toContain('"intent"')
+    // Paint layer must not fall through to raw text.
+    expect(row.summary).not.toBe(args)
+  })
+
   test("a write_file call paints the whole body as additions", async () => {
     await withTestRenderer(async (h) => {
       const shell = createAppShell(h.renderer, shellOpts)
diff --git a/src/tui/diff.ts b/src/tui/diff.ts
index 3a8fd78b3..b7e8bcf61 100644
--- a/src/tui/diff.ts
+++ b/src/tui/diff.ts
@@ -488,6 +488,15 @@ export function toolCallRow(input: ToolCallRowInput): StreamRow {
   // Identity of the sentence this call paints, not of its arguments: two calls
   // that read the same line are what a repeat looks like to the operator.
   const callKey = `${input.name} ${verb ?? ""} ${summary ?? ""}`
+  // Never leave `summary` unset when we have a verb or a summarised view —
+  // `undefined` makes the paint layer fall through to raw argument JSON
+  // (CL-5762). An empty string is fine: the verb alone names the call.
+  const paintSummary =
+    summary !== undefined
+      ? summary
+      : call !== null || summarised !== null
+        ? ""
+        : undefined
   return {
     role: "tool",
     text,
@@ -497,12 +506,7 @@ export function toolCallRow(input: ToolCallRowInput): StreamRow {
     ...(input.callId !== undefined ? { callId: input.callId } : {}),
     ...(diff !== null ? { diff } : {}),
     ...(verb !== undefined ? { verb } : {}),
-    // A summarised call may deliberately have no subject — its verb already
-    // names the whole call — and that blank must survive, or the row falls
-    // back to painting the raw arguments.
-    ...(summary !== undefined && (summary.length > 0 || summarised !== null)
-      ? { summary }
-      : {}),
+    ...(paintSummary !== undefined ? { summary: paintSummary } : {}),
     ...(stat !== undefined ? { stat } : {}),
     ...(detail !== undefined && detail.length > 0 ? { detail } : {}),
   }
diff --git a/src/tui/runner-host.test.ts b/src/tui/runner-host.test.ts
index 957e90ea4..190ab9d9d 100644
--- a/src/tui/runner-host.test.ts
+++ b/src/tui/runner-host.test.ts
@@ -74,6 +74,9 @@ describe("rowFromTranscriptEntry", () => {
       text: "{}",
       meta: "grep",
       verb: "Grep",
+      // Empty summary is intentional: without it the paint layer falls through
+      // to raw argument JSON (CL-5762). Verb alone names the call.
+      summary: "",
       pending: true,
       callKey: "grep Grep ",
       callId: "c",
diff --git a/src/tui/tool-args.ts b/src/tui/tool-args.ts
index a87f37a03..96c032af2 100644
--- a/src/tui/tool-args.ts
+++ b/src/tui/tool-args.ts
@@ -124,29 +124,73 @@ function isScalar(value: unknown): boolean {
 }
 
 /**
- * Scalar arguments as `key  value` pairs with their newlines intact — a shell
- * command or a prompt is written to be read as text, and pretty-printed JSON
- * would hand it back with its line breaks escaped.
+ * Scalar (or scalar-array) arguments as `key  value` pairs with their newlines
+ * intact — a shell command or a spawn prompt is written to be read as text, and
+ * pretty-printed JSON would hand it back with its line breaks escaped (CL-5762).
+ *
+ * Nested objects recurse one level so a task brief expands as fields rather than
+ * a JSON dump; deeper nesting collapses to a compact token.
  */
-function scalarDetail(args: Record<string, unknown>): readonly StyledBodyLine[] | null {
-  const entries = Object.entries(args)
-  if (entries.length === 0 || !entries.every(([, value]) => isScalar(value))) {
-    return null
-  }
+function fieldDetail(
+  args: Record<string, unknown>,
+  indent = 0,
+): readonly StyledBodyLine[] {
+  const pad = " ".repeat(indent)
   const lines: StyledBodyLine[] = []
-  for (const [key, value] of entries) {
-    const text = typeof value === "string" ? value : JSON.stringify(value)
-    const rows = (text ?? "null").split("\n")
-    rows.forEach((row, i) => {
-      lines.push(
-        i === 0
-          ? [
-              { text: `${key}: `, fg: UI.textDim },
-              { text: row, fg: UI.text },
-            ]
-          : [{ text: `${" ".repeat(key.length + 2)}${row}`, fg: UI.text }],
-      )
-    })
+  for (const [key, value] of Object.entries(args)) {
+    if (isScalar(value)) {
+      const text = typeof value === "string" ? value : JSON.stringify(value)
+      const rows = (text ?? "null").split("\n")
+      rows.forEach((row, i) => {
+        lines.push(
+          i === 0
+            ? [
+                { text: `${pad}${key}: `, fg: UI.textDim },
+                { text: row, fg: UI.text },
+              ]
+            : [{ text: `${pad}${" ".repeat(key.length + 2)}${row}`, fg: UI.text }],
+        )
+      })
+      continue
+    }
+    if (Array.isArray(value) && value.every(isScalar)) {
+      if (value.length === 0) {
+        lines.push([
+          { text: `${pad}${key}: `, fg: UI.textDim },
+          { text: "[]", fg: UI.text },
+        ])
+        continue
+      }
+      lines.push([{ text: `${pad}${key}:`, fg: UI.textDim }])
+      for (const item of value) {
+        const text = typeof item === "string" ? item : JSON.stringify(item)
+        for (const row of text.split("\n")) {
+          lines.push([{ text: `${pad}  - ${row}`, fg: UI.text }])
+        }
+      }
+      continue
+    }
+    if (typeof value === "object" && value !== null && !Array.isArray(value)) {
+      // One level of nesting is enough for a spawn brief; deeper stays compact.
+      if (indent === 0) {
+        lines.push([{ text: `${pad}${key}:`, fg: UI.textDim }])
+        lines.push(...fieldDetail(value as Record<string, unknown>, indent + 2))
+      } else {
+        lines.push([
+          { text: `${pad}${key}: `, fg: UI.textDim },
+          { text: "{…}", fg: UI.text },
+        ])
+      }
+      continue
+    }
+    // Arrays of objects, etc. — compact rather than a wall of JSON.
+    lines.push([
+      { text: `${pad}${key}: `, fg: UI.textDim },
+      {
+        text: Array.isArray(value) ? `[${value.length} items]` : "{…}",
+        fg: UI.text,
+      },
+    ])
   }
   return lines.slice(0, MAX_DETAIL_LINES)
 }
@@ -221,7 +265,9 @@ function subjectFor(
   args: Record<string, unknown>,
 ): string {
   const { summary } = summarizeToolArgs(name, raw)
-  if (!isArgumentList(args, summary)) return summary
+  // An empty formatter summary is not a subject — fall through to primarySubject
+  // so a task without description still paints its prompt rather than raw JSON.
+  if (summary.length > 0 && !isArgumentList(args, summary)) return summary
   return primarySubject(args) ?? summary
 }
 
@@ -246,7 +292,7 @@ export function toolArgsView(name: string, rawArgs: string): ToolArgsView | null
   // Its arguments are a query, not a subject: nobody reads a transcript for
   // the pagination cursor, so they belong behind the expand key or nowhere.
   if (args !== null && isMcpToolName(name)) {
-    return withDetail("", scalarDetail(args) ?? jsonDetail(args))
+    return withDetail("", fieldDetail(args))
   }
 
   if (args === null && raw.length <= INLINE_MAX && !raw.includes("\n")) return null
@@ -256,8 +302,10 @@ export function toolArgsView(name: string, rawArgs: string): ToolArgsView | null
     return summary.length === 0 ? null : withDetail(summary, jsonDetail(raw))
   }
   const subject = subjectFor(name, raw, args)
-  if (subject.length === 0) return null
-  return withDetail(subject, scalarDetail(args) ?? jsonDetail(args))
+  // Object args always get a summarised view — even with an empty subject the
+  // verb alone names the call and the body expands with real line breaks. A
+  // null return here is what used to dump raw argument JSON into the transcript.
+  return withDetail(subject, fieldDetail(args))
 }
 
 /**
diff --git a/src/tui/tool-formatter.test.ts b/src/tui/tool-formatter.test.ts
index 85912e669..715b2cd5f 100644
--- a/src/tui/tool-formatter.test.ts
+++ b/src/tui/tool-formatter.test.ts
@@ -285,6 +285,18 @@ describe("describeToolCall for task tool", () => {
     expect(result.summary).toBe("map all callers");
   });
 
+  test("task without description falls back to the prompt subject", () => {
+    const prompt = "Find every call site of leaveObserve and report them.";
+    const args = JSON.stringify({ agent: "explore", prompt, intent: "explore" });
+    const result = describeToolCall("task", args);
+    expect(result.display).toBe("Explore");
+    // ARG_VALUE_MAX = 48 with ellipsis when truncated
+    expect(result.summary.length).toBeLessThanOrEqual(48);
+    expect(result.full).toBe(prompt);
+    expect(result.summary.startsWith("Find every call site")).toBe(true);
+    expect(result.summary).not.toContain("intent");
+  });
+
   test("long description is abbreviated", () => {
     const long = "a".repeat(100);
     const args = JSON.stringify({ agent: "critique", description: long, prompt: "..." });
@@ -333,6 +345,19 @@ describe("task activity transcript lines", () => {
     expect(s.full).toBe("map callers of leaveObserve");
   });
 
+  test("summarizeToolArgs falls back to prompt when description is missing", () => {
+    const prompt = "Find every call site of leaveObserve and report them with paths.";
+    const s = summarizeToolArgs(
+      "task",
+      JSON.stringify({ agent: "explore", prompt, intent: "explore", maxTurns: 40 }),
+    );
+    expect(s.summary.length).toBeLessThanOrEqual(48);
+    expect(s.full).toBe(prompt);
+    expect(s.summary.startsWith("Find every call site")).toBe(true);
+    expect(s.summary).not.toContain("maxTurns");
+    expect(s.summary).not.toContain("intent");
+  });
+
   test("describeToolCall full keeps the untrimmed description for Ctrl+O", () => {
     const long = "a".repeat(80);
     const d = describeToolCall(
diff --git a/src/tui/tool-formatter.ts b/src/tui/tool-formatter.ts
index 4422375d4..5c2158551 100644
--- a/src/tui/tool-formatter.ts
+++ b/src/tui/tool-formatter.ts
@@ -132,16 +132,20 @@ export function describeToolCall(toolName: string, rawArgs: string): ToolCallDes
     if (!(taskParsed instanceof type.errors)) {
       const agentName = taskParsed.agent?.trim();
       const description = (taskParsed.description ?? "").trim();
+      // description is optional on spawn; the brief's prompt is the next best
+      // subject so the row never falls through to raw argument JSON.
+      const prompt = (taskParsed.prompt ?? "").trim();
+      const subject = description.length > 0 ? description : prompt;
       const display =
         agentName !== undefined && agentName.length > 0
           ? agentName[0]!.toUpperCase() + agentName.slice(1)
           : "Task";
-      // Collapsed row uses the abbreviated description; Alt+E expands to the full text.
+      // Collapsed row uses the abbreviated subject; Alt+E expands to the full text.
       return {
         display,
         role: "accent",
-        summary: description.length > 0 ? abbreviate(description, ARG_VALUE_MAX) : "",
-        full: description,
+        summary: subject.length > 0 ? abbreviate(subject, ARG_VALUE_MAX) : "",
+        full: subject,
         isShell: false,
       };
     }
@@ -183,7 +187,11 @@ const SearchFilesArgSchema = type({ pattern: "string", "path?": "string" });
 const WebSearchArgSchema = type({ query: "string" });
 const WebFetchArgSchema = type({ url: "string" });
 const ShellArgSchema = type({ command: "string" });
-const TaskArgSchema = type({ "agent?": "string", "description?": "string" });
+const TaskArgSchema = type({
+  "agent?": "string",
+  "description?": "string",
+  "prompt?": "string",
+});
 const WebSearchResultSchema = type({ results: "unknown[]" });
 const WebFetchResultSchema = type({ content: "string" });
 
@@ -234,14 +242,18 @@ export function summarizeToolArgs(toolName: string, rawArgs: string): ToolArgSum
     }
     case "task": {
       // Spawns carry a large structured brief (prompt, intent, criteria). The
-      // transcript only needs the short description; Alt+E still shows the
-      // full description text, not every spawn field.
+      // transcript only needs a short subject — prefer description, then prompt —
+      // so the row never dumps the whole JSON payload.
       const parsed = TaskArgSchema(obj);
       if (!(parsed instanceof type.errors)) {
         const desc = (parsed.description ?? "").trim();
         if (desc.length > 0) {
           return { summary: abbreviate(desc, ARG_VALUE_MAX), full: desc };
         }
+        const prompt = (parsed.prompt ?? "").trim();
+        if (prompt.length > 0) {
+          return { summary: abbreviate(prompt, ARG_VALUE_MAX), full: prompt };
+        }
       }
       return { summary: "", full: "" };
     }