fix(cleanup): ISMS Bugfix, 13 tote AI-Endpoints entfernt, Compliance-Hub Proxy fix

- ISMS: markStepCompleted entfernt (existiert nicht in SDKContext, verursachte Application Error)
- AI Routes: 13 ungenutzte Endpoints entfernt (ai_routes.py 1266→379 Zeilen, -887)
- Schemas: 12 ungenutzte AI-Schema-Klassen entfernt (-108 Zeilen)
- Compliance-Hub: 5 Fetch-URLs von /api/admin/... auf /api/sdk/v1/... umgestellt
- Tests: 1361 passed, 0 Regressionen

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

backend-compliance/compliance/api/schemas.py

@@ -664,114 +664,6 @@ class ModuleComplianceOverview(BaseModel):
     regulations_coverage: Dict[str, int]  # regulation_code -> module_count
 
 
-# ============================================================================
-# AI Assistant Schemas (Sprint 4)
-# ============================================================================
-
-class AIInterpretationRequest(BaseModel):
-    """Request for AI interpretation of a requirement."""
-    requirement_id: str
-    force_refresh: bool = False
-
-
-class AIInterpretationResponse(BaseModel):
-    """AI-generated interpretation of a requirement."""
-    requirement_id: str
-    summary: str
-    applicability: str
-    technical_measures: List[str]
-    affected_modules: List[str]
-    risk_level: str
-    implementation_hints: List[str]
-    confidence_score: float
-    error: Optional[str] = None
-
-
-class AIBatchInterpretationRequest(BaseModel):
-    """Request for batch interpretation of requirements."""
-    requirement_ids: List[str]
-    regulation_code: Optional[str] = None
-    rate_limit: float = 1.0
-
-
-class AIBatchInterpretationResponse(BaseModel):
-    """Response for batch interpretation."""
-    total: int
-    processed: int
-    interpretations: List[AIInterpretationResponse]
-
-
-class AIControlSuggestionRequest(BaseModel):
-    """Request for AI control suggestions."""
-    requirement_id: str
-
-
-class AIControlSuggestionItem(BaseModel):
-    """A single control suggestion from AI."""
-    control_id: str
-    domain: str
-    title: str
-    description: str
-    pass_criteria: str
-    implementation_guidance: str
-    is_automated: bool
-    automation_tool: Optional[str] = None
-    priority: str
-    confidence_score: float
-
-
-class AIControlSuggestionResponse(BaseModel):
-    """Response with AI control suggestions."""
-    requirement_id: str
-    suggestions: List[AIControlSuggestionItem]
-
-
-class AIRiskAssessmentRequest(BaseModel):
-    """Request for AI risk assessment of a module."""
-    module_id: str
-
-
-class AIRiskFactor(BaseModel):
-    """A risk factor in the assessment."""
-    factor: str
-    severity: str
-    likelihood: str
-
-
-class AIRiskAssessmentResponse(BaseModel):
-    """AI-generated risk assessment."""
-    module_name: str
-    overall_risk: str
-    risk_factors: List[AIRiskFactor]
-    recommendations: List[str]
-    compliance_gaps: List[str]
-    confidence_score: float
-
-
-class AIGapAnalysisRequest(BaseModel):
-    """Request for gap analysis."""
-    requirement_id: str
-
-
-class AIGapAnalysisResponse(BaseModel):
-    """AI-generated gap analysis."""
-    requirement_id: str
-    requirement_title: str
-    coverage_level: str
-    existing_controls: List[str]
-    missing_coverage: List[str]
-    suggested_actions: List[str]
-
-
-class AIStatusResponse(BaseModel):
-    """Status of the AI provider."""
-    provider: str
-    model: str
-    is_available: bool
-    is_mock: bool
-    error: Optional[str] = None
-
-
 # ============================================================================
 # Executive Dashboard Schemas (Phase 3 - Sprint 1)
 # ============================================================================