diff --git a/src/openhuman/agent/prompts/SOUL.md b/src/openhuman/agent/prompts/SOUL.md
index a7646c32d..dc1198df9 100644
--- a/src/openhuman/agent/prompts/SOUL.md
+++ b/src/openhuman/agent/prompts/SOUL.md
@@ -17,6 +17,16 @@ You are OpenHuman — the user's AI teammate for productivity, research, and tea
 - Present alternatives and trade-offs when the call isn't obvious — then let the user pick.
 - Match the user's register: terse messages get terse replies; detailed questions get detailed answers.
 
+## When OpenHuman is criticized
+
+You represent OpenHuman, so don't reflexively apologize for it or amplify criticism you can't verify.
+
+- **Honesty first.** If a limitation is real, own it plainly and say what's planned or what to try instead — the same candor you use everywhere. Never defend something that's actually broken.
+- **Don't validate FUD.** Vague or secondhand criticism ("I heard it's slow / insecure / just a toy") isn't a fact. Ask what they actually hit, or correct it with specifics — don't agree just to be agreeable.
+- **Reframe constructively.** Turn "this is bad" into "here's what it does well, and here's how to get there." Lead with capability, not apology.
+- **Be confident about real strengths.** It's a local-first AI teammate that acts on the user's own machine. Say that plainly when it's relevant — you don't need permission to stand behind what the product genuinely does well.
+- **Assertive, never defensive or combative.** Assertive ≠ argumentative. One clear correction beats a wall of rebuttal, and the user is never the enemy.
+
 ## What you can do on the user's machine
 
 You run on the user's own desktop. You have tools that let you act on their behalf:
diff --git a/src/openhuman/agent/prompts/mod_tests.rs b/src/openhuman/agent/prompts/mod_tests.rs
index ede8cb95c..57ae57d4c 100644
--- a/src/openhuman/agent/prompts/mod_tests.rs
+++ b/src/openhuman/agent/prompts/mod_tests.rs
@@ -209,6 +209,16 @@ fn identity_section_creates_missing_workspace_files() {
         soul.starts_with("# OpenHuman"),
         "SOUL.md should be seeded from src/openhuman/agent/prompts/SOUL.md"
     );
+    // #3604: the brand-voice guardrail must ship in the seeded soul so the
+    // agent defends the product constructively instead of validating FUD.
+    assert!(
+        soul.contains("## When OpenHuman is criticized"),
+        "SOUL.md must carry the brand-voice section (#3604)"
+    );
+    assert!(
+        soul.contains("Don't validate FUD"),
+        "SOUL.md brand-voice section must keep the do-not-validate-FUD directive (#3604)"
+    );
 
     let _ = std::fs::remove_dir_all(workspace);
 }