Watch & React Workflow

The Watch & React workflow is for situations where you want to monitor nodes over time rather than analyze a specific recorded moment. Instead of recording and reviewing, you set up watches and the agent observes the game as it runs.

When to use watches instead of recordings

Use recordings when...	Use watches when...
The bug already happened	You want to catch the bug as it happens
You need frame-precise timing	Coarse timing is sufficient
You want to scrub backward in time	You want real-time monitoring
The bug is intermittent and hard to trigger	The bug is happening continuously
Post-mortem analysis	Live investigation

The basic loop

1. spatial_snapshot          — get current frame number
2. spatial_watch (add)       — register nodes to monitor
3. [game runs, things happen]
4. spatial_delta             — check what changed
5. [repeat 3-4 as needed]
6. spatial_watch (remove)    — clean up

Example: Monitoring player health during combat

You are debugging a health regen system that seems to not work correctly during certain combat conditions.

Step 1: Set up the watches

I want to monitor the player health, shield, and position during this combat encounter.

Let me set up watches on the relevant properties. What is the node path to the player?

Just Player. It has health and shield as exported vars.

Baseline at frame 500. Now setting up the watches.

Watch created (watch_id: w_c8d9e0). Monitoring Player.position, .velocity, .health, and .shield. Start the combat encounter whenever you are ready.

Step 2: Check changes during combat

I just took some hits and health does not seem to be regenerating. Check it.

Delta from frame 500 to frame 847 (5.8 seconds):\n\n- health: 100 → 35 (3 damage events)\n- shield: 50 → 0 (fully depleted at frame 612)\n- position moved about 8 meters total\n\nNo regeneration events. Health only went down. Shield depleted at frame 612 and has not recovered. Expected regen at 5 hp/sec for 5 seconds would have added 25 hp — that should have brought health from 35 to 60 but health is still 35.

Keep watching. I am out of combat now — wait 10 seconds and check again.

Delta from frame 847 to frame 1447 (10 seconds): health stayed at 35, no regen. Shield also still at 0. Position barely moved (standing still). The regen is definitely not triggering despite being out of combat for 10 seconds.

Step 3: Narrow down the cause

Check the health regen controller node.

HealthRegenController state: regen_active=false, in_combat=true, regen_timer=0.0. Even though you are standing still with no enemies nearby, in_combat is still true. The combat exit timer is not running — regen_timer is stuck at 0. The HealthRegenController.in_combat flag is not being cleared when combat ends.

The agent found that the in_combat flag never resets — likely a missing call to set it to false when the last enemy dies. The fix is in the combat state manager.

Example: Watching enemy patrol behavior

You suspect enemies are teleporting slightly during patrol waypoint transitions.

Watch Enemy_0 position and velocity with a high threshold so I only see big changes.

Watch created (watch_id: w_f1a2b3). I will poll with min_distance_change: 0.5 to filter small movements.

It has been patrolling for 20 seconds. Check.

Found a suspicious event at frame 1102: position jumped from [4.2, 0.0, 8.1] to [4.2, 0.0, 8.7] in one frame — a 0.6m teleport with velocity=0. That is not smooth movement. At frame 1101 velocity was [0,0,3.2] and at frame 1103 velocity is [0,0,2.8]. The waypoint transition at frame 1102 snapped the position instead of continuing movement smoothly.

Watching multiple nodes simultaneously

Set up watches on related nodes to catch interactions:

{ "action": "add", "watch": { "node": "Player", "track": ["position", "health"] } }
{ "action": "add", "watch": { "node": "Boss", "track": ["position", "velocity", "phase", "health"] } }
{ "action": "add", "watch": { "node": "Boss/AttackHitbox", "track": ["monitoring"] } }
{ "action": "add", "watch": { "node": "BossArena/HazardTiles", "track": ["monitoring"] } }

Now a single spatial_delta call shows changes across all four nodes — you can see the Boss switch phase at a certain health threshold, the AttackHitbox become active, and the HazardTiles respond, all in one response.

Delta poll frequency

You control how often to poll. The agent does not poll automatically — you ask for a delta when you want it.

Typical patterns:

• Quick check: poll once after a specific event ("I just died, check the delta")
• Periodic: poll every 10-30 seconds during a long session
• Event-triggered: poll after specific in-game events ("I just finished wave 3, check it")

Poll when you need the next delta. The watch remains session state between calls, and each delta advances the baseline. Do not infer zero runtime cost from a lack of network polling.

Cleaning up

{ "action": "clear" }

Always clear watches when moving to a new investigation. Stale watches from a previous investigation add noise to subsequent deltas.

Or delete a specific watch:

{ "action": "remove", "watch_id": "w_c8d9e0" }