diff --git a/docs/tools-reference.md b/docs/tools-reference.md
index 78ab32a..ffa030c 100644
--- a/docs/tools-reference.md
+++ b/docs/tools-reference.md
@@ -85,6 +85,14 @@ subsystem oracles. Test scenes are **synthesized** via `Age.Engine/Sys4/ScriptAs
 | `play [--boot] [--state <f>] [--save-state <f>] <SCENE.BIN…> [0xADDR=VAL…]` | ★ Cross-scene **state runner**: run a scene sequence carrying persistent globals. `--boot` first runs the 9 `*INIT` data scripts (real skill/item/unit/map/stage state). `--state`/`--save-state` load/persist a JSON snapshot. | `GameSession`; **executes call-script**. |
 | `sweep [--boot] [0xADDR=VAL…]` | Corpus-scale run. **With call-script execution on: 284/297 exit, 13 STEP-LIMIT** (input/state-gated ADV scenes spin headless once subroutine global-writes drive their loops — state divergence, not a bug; 0 depth-cap/unresolved). **With seeds = a story-state explorer**: reports which scenes' dialogue changes ±seed (e.g. form flag `0xa57=1` → 34/297 scenes). | |
 
+**`--trace [--trace-file <path>] [--trace-steps]`** (on `run`/`play`/`sweep`): stream the engine's own
+diagnostic events over the `Age.Engine.Diagnostics.ITraceSink` seam — scene/subroutine frame enter+exit
+(indented by call depth), call-script dispatch with resolved name, and the final halt+step count — to
+console or a file. Add `--trace-steps` for per-instruction opcode/arg + stub-op detail (high volume;
+gated). Absent ⇒ no tracing (`NullTraceSink`, byte-identical run). This is an **engine** fact stream:
+frontends consume it instead of reimplementing a diagnostic `IHost`. Example: `play SC0000.BIN --trace`
+shows `» SC0000.BIN (enter, TopScene)` → `call-script 0xee =INPUTNAME.BIN (resolved)` → `halt: …`.
+
 **Godot frontend** (`S:/Godot/Godot_v4.7…`; project = `godot/`). Toolchain: `godot --headless --path godot
 --import` → `dotnet build godot/Himegari.csproj` → `godot [--headless] --path godot [-- <userargs>]`.
 Plays the real bytecode with call-script execution on (subroutines run live). `--headless` can't render
diff --git a/engine/Age.Cli/Program.cs b/engine/Age.Cli/Program.cs
index ffbf242..08efaee 100644
--- a/engine/Age.Cli/Program.cs
+++ b/engine/Age.Cli/Program.cs
@@ -1,6 +1,8 @@
 using System.Text.Json;
 using System.Text.RegularExpressions;
+using Age.Engine.Diagnostics;
 using Age.Engine.Hosting;
+using Age.Engine.Model;
 using Age.Engine.Sys4;
 using Age.Engine.Vm;
 
@@ -9,13 +11,25 @@ var table = OpcodeTableJson.Load(Paths.OpcodesJson);
 // `trace` stays provider-less on purpose (the base-ISA offset oracle).
 var provider = Sys4ScriptProvider.Load(table);
 
+// --trace [--trace-file <path>] [--trace-steps] → a TextTraceSink to console or file; else inert.
+static ITraceSink BuildSink(string[] a, OpcodeTable tbl)
+{
+    if (!a.Contains("--trace")) return NullTraceSink.Instance;
+    bool steps = a.Contains("--trace-steps");
+    int fi = Array.IndexOf(a, "--trace-file");
+    TextWriter w = (fi >= 0 && fi + 1 < a.Length)
+        ? new StreamWriter(a[fi + 1]) { AutoFlush = true }
+        : Console.Out;
+    return new TextTraceSink(w, tbl, steps);
+}
+
 if (args.Length == 0) { Console.WriteLine("usage: run <file> | trace <out.json>"); return 1; }
 
 if (args[0] == "run")
 {
     var script = Sys4Loader.Load(args[1], table);
     var runHost = new CaptureHost();
-    var vm = new VirtualMachine(script, table, runHost, null, provider);
+    var vm = new VirtualMachine(script, table, runHost, null, provider, BuildSink(args, table));
     vm.Run();
     Console.WriteLine($"{Path.GetFileName(args[1])}: {vm.Steps} steps, {vm.Emitted.Count} show-text, {vm.CallScriptDispatches} call-scripts (halt: {vm.HaltReason})");
     foreach (var (off, text, scr) in vm.Emitted.Take(30)) Console.WriteLine($"  [{scr} 0x{off:x}] {text}");
@@ -100,7 +114,7 @@ if (args[0] == "play")
     foreach (var name in scenes)
     {
         var script = Sys4Loader.Load(scripts[name.ToUpperInvariant()], table);
-        var r = session.RunScene(script, table, new CaptureHost(), null, provider);
+        var r = session.RunScene(script, table, new CaptureHost(), null, provider, BuildSink(args, table));
         totalLines += r.Emitted.Count;
         Console.WriteLine($"  {name,-14} {r.Emitted.Count,4} lines, {r.Steps,7} steps (halt: {r.Halt})");
     }
@@ -165,7 +179,7 @@ if (args[0] == "sweep")
     foreach (var name in names)
     {
         var session = Fresh();
-        var r = session.RunScene(Sys4Loader.Load(scripts[name], table), table, new CaptureHost(), null, provider);
+        var r = session.RunScene(Sys4Loader.Load(scripts[name], table), table, new CaptureHost(), null, provider, BuildSink(args, table));
         var halt = r.Halt ?? "null";
         haltDist[halt] = haltDist.GetValueOrDefault(halt) + 1;
         totalLines += r.Emitted.Count;
diff --git a/engine/Age.Engine/Vm/VirtualMachine.cs b/engine/Age.Engine/Vm/VirtualMachine.cs
index 87989af..6426db4 100644
--- a/engine/Age.Engine/Vm/VirtualMachine.cs
+++ b/engine/Age.Engine/Vm/VirtualMachine.cs
@@ -226,7 +226,9 @@ public sealed class VirtualMachine
             case "play-bgm":   _host.PlayBgm(Read(a[0])); return pc + 1;
             case "play-voice": _host.PlayVoice(Read(a[0])); return pc + 1;
             default:
-                _sink.Emit(TraceEvent.Stub(op, pc)); return pc + 1;
+                // Stub is per-instruction frequency (the VM handles ~30 ops; the rest hit here, e.g.
+                // 0x258/0x259 stmt markers appear en masse), so gate it with Step — else --trace floods.
+                if (_sink.TracingSteps) _sink.Emit(TraceEvent.Stub(op, pc)); return pc + 1;
         }
     }
 }