feat(iace): complete CE risk assessment — LLM tech-file generation, multi-format export, TipTap editor

Phase 1: Fix completeness gates G23 (require verified/rejected mitigations) and G09 (audit trail check)
Phase 2: LLM-based tech-file section generation with 19 German prompts and RAG enrichment
Phase 3: Multi-format document export (PDF/Excel/DOCX/Markdown/JSON)
Phase 4: Company profile → IACE data flow with auto component/classification creation
Phase 5: TipTap WYSIWYG editor replacing textarea for tech-file sections
Phase 6: User journey tests, developer portal API reference, updated documentation

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ai-compliance-sdk/internal/iace/completeness_test.go

@@ -33,7 +33,7 @@ func TestCompletenessCheck_EmptyContext(t *testing.T) {
 	// With nil project, most gates fail. However, some auto-pass:
 	// G06 (AI classification): auto-passes when HasAI=false
 	// G22 (critical/high mitigated): auto-passes when no critical/high assessments exist
-	// G23 (mitigations verified): auto-passes when no mitigations with status "implemented"
+	// G23 (mitigations verified): auto-passes when no mitigations (empty list)
 	// G42 (AI documents): auto-passes when HasAI=false
 	// That gives 4 required gates passing even with empty context.
 	if result.PassedRequired != 4 {
@@ -89,7 +89,8 @@ func TestCompletenessCheck_MinimalValidProject(t *testing.T) {
 			{ID: uuid.New(), ProjectID: projectID, SectionType: "risk_assessment_report"},
 			{ID: uuid.New(), ProjectID: projectID, SectionType: "hazard_log_combined"},
 		},
-		HasAI: false,
+		HasAI:                    false,
+		PatternMatchingPerformed: true,
 	}
 
 	result := checker.Check(ctx)
@@ -376,11 +377,11 @@ func TestCompletenessCheck_G23_MitigationsVerified(t *testing.T) {
 			wantG23Passed: false,
 		},
 		{
-			name: "planned mitigations pass G23 (not yet implemented)",
+			name: "planned mitigations fail G23 (not yet verified)",
 			mitigations: []Mitigation{
 				{HazardID: hazardID, Status: MitigationStatusPlanned},
 			},
-			wantG23Passed: true,
+			wantG23Passed: false,
 		},
 		{
 			name: "rejected mitigations pass G23",
@@ -390,12 +391,20 @@ func TestCompletenessCheck_G23_MitigationsVerified(t *testing.T) {
 			wantG23Passed: true,
 		},
 		{
-			name: "mix of verified planned rejected passes G23",
+			name: "mix of verified planned rejected fails G23",
 			mitigations: []Mitigation{
 				{HazardID: hazardID, Status: MitigationStatusVerified},
 				{HazardID: hazardID, Status: MitigationStatusPlanned},
 				{HazardID: hazardID, Status: MitigationStatusRejected},
 			},
+			wantG23Passed: false,
+		},
+		{
+			name: "mix of verified and rejected passes G23",
+			mitigations: []Mitigation{
+				{HazardID: hazardID, Status: MitigationStatusVerified},
+				{HazardID: hazardID, Status: MitigationStatusRejected},
+			},
 			wantG23Passed: true,
 		},
 	}
@@ -422,6 +431,48 @@ func TestCompletenessCheck_G23_MitigationsVerified(t *testing.T) {
 	}
 }
 
+func TestCompletenessCheck_G09_PatternMatchingPerformed(t *testing.T) {
+	checker := NewCompletenessChecker()
+
+	tests := []struct {
+		name          string
+		performed     bool
+		wantG09Passed bool
+	}{
+		{
+			name:          "pattern matching not performed fails G09",
+			performed:     false,
+			wantG09Passed: false,
+		},
+		{
+			name:          "pattern matching performed passes G09",
+			performed:     true,
+			wantG09Passed: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			ctx := &CompletenessContext{
+				Project:                  &Project{MachineName: "Test"},
+				PatternMatchingPerformed: tt.performed,
+			}
+
+			result := checker.Check(ctx)
+
+			for _, g := range result.Gates {
+				if g.ID == "G09" {
+					if g.Passed != tt.wantG09Passed {
+						t.Errorf("G09 Passed = %v, want %v", g.Passed, tt.wantG09Passed)
+					}
+					return
+				}
+			}
+			t.Error("G09 gate not found in results")
+		})
+	}
+}
+
 func TestCompletenessCheck_G24_ResidualRiskAccepted(t *testing.T) {
 	checker := NewCompletenessChecker()