diff --git a/fleetd/src/main/java/dev/ltms/fleet/member/OpenCodeLauncher.java b/fleetd/src/main/java/dev/ltms/fleet/member/OpenCodeLauncher.java
index 28dd5b2..006da0d 100644
--- a/fleetd/src/main/java/dev/ltms/fleet/member/OpenCodeLauncher.java
+++ b/fleetd/src/main/java/dev/ltms/fleet/member/OpenCodeLauncher.java
@@ -794,9 +794,16 @@ public final class OpenCodeLauncher extends HerdrPeerLauncher {
             String requestedProvider = requestedParts == null ? null : requestedParts[0];
             String requestedId = requestedParts == null ? cfg.model() : requestedParts[1];
             boolean idMatches = requestedId.equals(actual.id());
-            // Compare the provider ONLY when the profile actually asked for one — a bare model name
-            // (no "/") is a match on id alone, regardless of which provider opencode resolved it to.
-            boolean providerMatches = requestedProvider == null || requestedProvider.equals(actual.provider());
+            // Compare the provider ONLY when BOTH sides have one. requestedProvider == null covers
+            // a bare profile model with no "/" — the profile never asked for a specific provider.
+            // actual.provider() == null covers opencode's model JSON having an id but no providerID
+            // (a real shape parseModel accepts) — that is missing evidence, not a contradiction, and
+            // acceptance rule 4 says missing evidence is UNKNOWN, never a mismatch. Narrowing the
+            // provider comparison this way keeps the id comparison (the part that actually caught the
+            // xf bug) fully intact — a genuine id mismatch is still caught either way (fleetd #175
+            // review round 2).
+            boolean providerMatches = requestedProvider == null || actual.provider() == null
+                    || requestedProvider.equals(actual.provider());
             if (idMatches && providerMatches) {
                 return;
             }
diff --git a/fleetd/src/test/java/dev/ltms/fleet/member/OpenCodeLauncherTest.java b/fleetd/src/test/java/dev/ltms/fleet/member/OpenCodeLauncherTest.java
index f74abf0..df70228 100644
--- a/fleetd/src/test/java/dev/ltms/fleet/member/OpenCodeLauncherTest.java
+++ b/fleetd/src/test/java/dev/ltms/fleet/member/OpenCodeLauncherTest.java
@@ -932,6 +932,53 @@ class OpenCodeLauncherTest {
         assertTrue(exhausted.isEmpty(), "id and provider both match → never a mismatch: " + exhausted);
     }
 
+    /**
+     * Incomplete evidence, not THE TRAP's provider mismatch: opencode's {@code model} JSON had an
+     * {@code id} but no {@code providerID} at all (a real shape {@code parseModel} accepts — see
+     * {@code OpenCodeSessionDiscoveryTest}). The id matches; the provider dimension is simply
+     * unknown, not contradicted. A provider-prefixed profile must NOT be quarantined on this —
+     * that would quarantine on incomplete evidence, which acceptance rule 4 forbids.
+     */
+    @Test
+    void aMissingProviderIdInTheEvidenceIsUnknownNotAMismatchWhenTheIdMatches(
+            @TempDir Path configRoot, @TempDir Path discRoot) throws Exception {
+        List<String> exhausted = new ArrayList<>();
+        ExhaustionSink sink = (target, reason) -> exhausted.add(reason);
+        FleetConfig.Profile cfg = opencodeCfg("openai/gpt-5.6-terra", null, null);
+        PeerHandle handle = serviceWithSink(new FakeHerdr(), configRoot, discRoot, cfg, sink)
+                .spawn(new SpawnRequest(null, "/work/dir", null));
+        OpenCodeSessionDiscoveryTest.writeRecord(discRoot, "ses_x", "/work/dir", 1000L,
+                "{\"id\":\"gpt-5.6-terra\"}");
+
+        assertEquals("ses_x", handle.agentSessionId());
+        assertTrue(exhausted.isEmpty(),
+                "id matches and provider is simply unknown (absent), never a mismatch: " + exhausted);
+    }
+
+    /**
+     * The other half of the same fix: a missing {@code providerID} must NOT blind the check to a
+     * genuine id mismatch. This is what proves the fix narrows the comparison rather than switching
+     * the whole check off whenever {@code providerID} happens to be absent.
+     */
+    @Test
+    void aMissingProviderIdInTheEvidenceStillCatchesARealIdMismatch(
+            @TempDir Path configRoot, @TempDir Path discRoot) throws Exception {
+        List<String> exhausted = new ArrayList<>();
+        ExhaustionSink sink = (target, reason) -> exhausted.add(reason);
+        FleetConfig.Profile cfg = opencodeCfg("openai/gpt-5.6-terra", null, null);
+        PeerHandle handle = serviceWithSink(new FakeHerdr(), configRoot, discRoot, cfg, sink)
+                .spawn(new SpawnRequest(null, "/work/dir", null));
+        OpenCodeSessionDiscoveryTest.writeRecord(discRoot, "ses_x", "/work/dir", 1000L,
+                "{\"id\":\"gpt-5.6-sol\"}");
+
+        assertEquals("ses_x", handle.agentSessionId());
+        assertEquals(1, exhausted.size(),
+                "the id genuinely differs, so this must still quarantine even with providerID absent: "
+                        + exhausted);
+        assertTrue(exhausted.get(0).contains("gpt-5.6-sol") && exhausted.get(0).contains(cfg.model()),
+                "reports both the requested and actual model: " + exhausted.get(0));
+    }
+
     /**
      * THE TRAP, row 3: a profile that names no provider prefix (bare {@code "deepseek-v4-flash"})
      * must match on id alone — the profile never asked for a specific provider, so opencode