diff --git a/apps/api/src/services/awooop_ansible_check_mode_service.py b/apps/api/src/services/awooop_ansible_check_mode_service.py
index df512db8..10fb9271 100644
--- a/apps/api/src/services/awooop_ansible_check_mode_service.py
+++ b/apps/api/src/services/awooop_ansible_check_mode_service.py
@@ -483,6 +483,99 @@ def _build_apply_result_payload(result: AnsibleRunResult) -> tuple[str, dict[str
     return status, output, dry_run_result, error
 
 
+def _build_auto_repair_execution_receipt(
+    claim: AnsibleCheckModeClaim,
+    result: AnsibleRunResult,
+    *,
+    apply_op_id: str,
+) -> dict[str, Any]:
+    success = result.returncode == 0
+    return {
+        "incident_id": claim.incident_id,
+        "playbook_id": str(claim.catalog_id or "")[:36] or "ansible",
+        "playbook_name": f"Ansible controlled apply: {claim.apply_playbook_path}"[:200],
+        "success": success,
+        "executed_steps": [
+            f"candidate:{claim.source_candidate_op_id}",
+            f"check_mode:{claim.op_id}",
+            f"apply:{apply_op_id}",
+            f"catalog:{claim.catalog_id}",
+            f"returncode:{result.returncode}",
+        ],
+        "error_message": None if success else _tail(result.stderr or result.stdout, 2000),
+        "triggered_by": "ansible_controlled_apply",
+        "similarity_score": None,
+        "risk_level": str(claim.risk_level or ""),
+        "execution_time_ms": result.duration_ms,
+    }
+
+
+async def _record_auto_repair_execution_receipt(
+    claim: AnsibleCheckModeClaim,
+    result: AnsibleRunResult,
+    *,
+    apply_op_id: str,
+    project_id: str,
+) -> bool:
+    receipt = _build_auto_repair_execution_receipt(
+        claim,
+        result,
+        apply_op_id=apply_op_id,
+    )
+    try:
+        async with get_db_context(project_id) as db:
+            inserted = await db.execute(
+                text("""
+                    INSERT INTO auto_repair_executions (
+                        incident_id,
+                        playbook_id,
+                        playbook_name,
+                        success,
+                        executed_steps,
+                        error_message,
+                        triggered_by,
+                        similarity_score,
+                        risk_level,
+                        execution_time_ms
+                    )
+                    SELECT
+                        :incident_id,
+                        :playbook_id,
+                        :playbook_name,
+                        :success,
+                        CAST(:executed_steps AS jsonb),
+                        :error_message,
+                        :triggered_by,
+                        :similarity_score,
+                        :risk_level,
+                        :execution_time_ms
+                    WHERE NOT EXISTS (
+                        SELECT 1
+                        FROM auto_repair_executions existing
+                        WHERE existing.incident_id = :incident_id
+                          AND existing.triggered_by = :triggered_by
+                          AND existing.executed_steps::text LIKE :apply_op_id_needle
+                    )
+                    RETURNING id
+                """),
+                {
+                    **receipt,
+                    "executed_steps": json.dumps(receipt["executed_steps"], ensure_ascii=False),
+                    "apply_op_id_needle": f"%{apply_op_id}%",
+                },
+            )
+            return inserted.scalar() is not None
+    except Exception as exc:
+        logger.warning(
+            "ansible_auto_repair_execution_receipt_failed",
+            incident_id=claim.incident_id,
+            catalog_id=claim.catalog_id,
+            apply_op_id=apply_op_id,
+            error=str(exc),
+        )
+        return False
+
+
 async def claim_pending_check_modes(
     *,
     project_id: str = "awoooi",
@@ -888,6 +981,12 @@ async def run_controlled_apply_for_claim(
                 "op_id": apply_op_id,
             },
         )
+    receipt_written = await _record_auto_repair_execution_receipt(
+        claim,
+        result,
+        apply_op_id=apply_op_id,
+        project_id=project_id,
+    )
 
     logger.info(
         "ansible_controlled_apply_completed",
@@ -898,6 +997,7 @@ async def run_controlled_apply_for_claim(
         catalog_id=claim.catalog_id,
         returncode=result.returncode,
         timed_out=result.timed_out,
+        auto_repair_receipt_written=receipt_written,
     )
     return result
 
diff --git a/apps/api/tests/test_awooop_truth_chain_service.py b/apps/api/tests/test_awooop_truth_chain_service.py
index 821bcd26..5656b8b4 100644
--- a/apps/api/tests/test_awooop_truth_chain_service.py
+++ b/apps/api/tests/test_awooop_truth_chain_service.py
@@ -12,7 +12,10 @@ from src.services.awooop_ansible_audit_service import (
     record_ansible_decision_audit,
 )
 from src.services.awooop_ansible_check_mode_service import (
+    AnsibleCheckModeClaim,
+    AnsibleRunResult,
     _automation_operation_log_incident_id,
+    _build_auto_repair_execution_receipt,
     build_ansible_apply_command,
     build_ansible_check_mode_claim_input,
     build_ansible_check_mode_command,
@@ -1442,6 +1445,40 @@ def test_ansible_apply_command_uses_controlled_apply_without_check(tmp_path: Pat
     assert str(known_hosts) in spec.command[-1]
 
 
+def test_ansible_controlled_apply_builds_auto_repair_receipt() -> None:
+    claim = AnsibleCheckModeClaim(
+        op_id="check-op-1",
+        source_candidate_op_id="candidate-op-1",
+        incident_id="INC-20260627-NODE110",
+        catalog_id="ansible:110-devops",
+        playbook_path="infra/ansible/playbooks/110-devops.yml",
+        apply_playbook_path="infra/ansible/playbooks/110-devops.yml",
+        inventory_hosts=("host_110",),
+        risk_level="medium",
+        input_payload={"controlled_apply_allowed": True},
+    )
+    result = AnsibleRunResult(
+        returncode=0,
+        stdout="ok",
+        stderr="",
+        duration_ms=1234,
+    )
+
+    receipt = _build_auto_repair_execution_receipt(
+        claim,
+        result,
+        apply_op_id="apply-op-1",
+    )
+
+    assert receipt["incident_id"] == "INC-20260627-NODE110"
+    assert receipt["playbook_id"] == "ansible:110-devops"
+    assert receipt["success"] is True
+    assert receipt["triggered_by"] == "ansible_controlled_apply"
+    assert "apply:apply-op-1" in receipt["executed_steps"]
+    assert receipt["risk_level"] == "medium"
+    assert receipt["execution_time_ms"] == 1234
+
+
 def test_ansible_claim_query_limits_recent_candidate_backlog() -> None:
     source = inspect.getsource(claim_pending_check_modes)