diff --git a/tests/phase3/test_decision_framework.py b/tests/phase3/test_decision_framework.py
new file mode 100644
index 00000000..ff8a7997
--- /dev/null
+++ b/tests/phase3/test_decision_framework.py
@@ -0,0 +1,358 @@
+"""
+Phase 3: Decision Framework Tests
+Tests for distributed decision making, voting systems, and consensus algorithms
+"""
+
+import pytest
+import asyncio
+import json
+from datetime import datetime, timedelta
+from unittest.mock import Mock, AsyncMock
+from typing import Dict, List, Any
+
+# Mock imports for testing
+class MockDecisionEngine:
+    def __init__(self):
+        self.decisions = {}
+        self.votes = {}
+        
+    async def make_decision(self, decision_data: Dict[str, Any]) -> Dict[str, Any]:
+        decision_id = decision_data.get('decision_id', 'test_decision')
+        self.decisions[decision_id] = decision_data
+        return {
+            'decision_id': decision_id,
+            'status': 'completed',
+            'result': decision_data.get('proposal', 'approved'),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+    
+    async def submit_vote(self, vote_data: Dict[str, Any]) -> Dict[str, Any]:
+        vote_id = vote_data.get('vote_id', 'test_vote')
+        self.votes[vote_id] = vote_data
+        return {
+            'vote_id': vote_id,
+            'status': 'recorded',
+            'timestamp': datetime.utcnow().isoformat()
+        }
+
+class MockConsensusAlgorithm:
+    def __init__(self):
+        self.consensus_results = {}
+        
+    async def achieve_consensus(self, participants: List[str], proposal: Dict[str, Any]) -> Dict[str, Any]:
+        consensus_id = f"consensus_{len(self.consensus_results)}"
+        self.consensus_results[consensus_id] = {
+            'participants': participants,
+            'proposal': proposal,
+            'result': 'consensus_reached'
+        }
+        return {
+            'consensus_id': consensus_id,
+            'status': 'consensus_reached',
+            'agreement': True,
+            'timestamp': datetime.utcnow().isoformat()
+        }
+
+class TestDecisionEngine:
+    """Test the decision engine functionality"""
+    
+    def setup_method(self):
+        self.decision_engine = MockDecisionEngine()
+    
+    @pytest.mark.asyncio
+    async def test_make_decision(self):
+        """Test basic decision making"""
+        decision_data = {
+            'decision_id': 'test_decision_001',
+            'proposal': 'test_proposal',
+            'priority': 'high'
+        }
+        
+        result = await self.decision_engine.make_decision(decision_data)
+        
+        assert result['decision_id'] == 'test_decision_001'
+        assert result['status'] == 'completed'
+        assert result['result'] == 'test_proposal'
+        assert 'timestamp' in result
+    
+    @pytest.mark.asyncio
+    async def test_submit_vote(self):
+        """Test vote submission"""
+        vote_data = {
+            'vote_id': 'test_vote_001',
+            'voter_id': 'agent_001',
+            'vote': 'approve',
+            'decision_id': 'test_decision_001'
+        }
+        
+        result = await self.decision_engine.submit_vote(vote_data)
+        
+        assert result['vote_id'] == 'test_vote_001'
+        assert result['status'] == 'recorded'
+        assert 'timestamp' in result
+    
+    @pytest.mark.asyncio
+    async def test_decision_with_complex_data(self):
+        """Test decision making with complex data"""
+        decision_data = {
+            'decision_id': 'complex_decision_001',
+            'proposal': {
+                'action': 'resource_allocation',
+                'resources': ['cpu', 'memory', 'storage'],
+                'amounts': {'cpu': 50, 'memory': 2048, 'storage': 100}
+            },
+            'participants': ['agent_001', 'agent_002', 'agent_003'],
+            'deadline': (datetime.utcnow() + timedelta(hours=1)).isoformat()
+        }
+        
+        result = await self.decision_engine.make_decision(decision_data)
+        
+        assert result['decision_id'] == 'complex_decision_001'
+        assert result['status'] == 'completed'
+        assert 'timestamp' in result
+
+class TestConsensusAlgorithm:
+    """Test consensus algorithm functionality"""
+    
+    def setup_method(self):
+        self.consensus = MockConsensusAlgorithm()
+    
+    @pytest.mark.asyncio
+    async def test_achieve_consensus(self):
+        """Test basic consensus achievement"""
+        participants = ['agent_001', 'agent_002', 'agent_003']
+        proposal = {
+            'action': 'system_update',
+            'version': '1.0.0',
+            'description': 'Update system to new version'
+        }
+        
+        result = await self.consensus.achieve_consensus(participants, proposal)
+        
+        assert result['status'] == 'consensus_reached'
+        assert result['agreement'] is True
+        assert 'consensus_id' in result
+        assert 'timestamp' in result
+    
+    @pytest.mark.asyncio
+    async def test_consensus_with_single_agent(self):
+        """Test consensus with single participant"""
+        participants = ['agent_001']
+        proposal = {'action': 'test_action'}
+        
+        result = await self.consensus.achieve_consensus(participants, proposal)
+        
+        assert result['status'] == 'consensus_reached'
+        assert result['agreement'] is True
+    
+    @pytest.mark.asyncio
+    async def test_consensus_with_complex_proposal(self):
+        """Test consensus with complex proposal"""
+        participants = ['agent_001', 'agent_002', 'agent_003', 'agent_004']
+        proposal = {
+            'action': 'policy_change',
+            'policy': {
+                'name': 'resource_allocation_policy',
+                'rules': [
+                    {'rule': 'priority_based', 'weight': 0.6},
+                    {'rule': 'fair_share', 'weight': 0.4}
+                ],
+                'effective_date': datetime.utcnow().isoformat()
+            }
+        }
+        
+        result = await self.consensus.achieve_consensus(participants, proposal)
+        
+        assert result['status'] == 'consensus_reached'
+        assert result['agreement'] is True
+        assert 'consensus_id' in result
+
+class TestVotingSystem:
+    """Test voting system functionality"""
+    
+    def setup_method(self):
+        self.decision_engine = MockDecisionEngine()
+        self.votes = {}
+    
+    @pytest.mark.asyncio
+    async def test_majority_voting(self):
+        """Test majority voting mechanism"""
+        votes = [
+            {'voter_id': 'agent_001', 'vote': 'approve'},
+            {'voter_id': 'agent_002', 'vote': 'approve'},
+            {'voter_id': 'agent_003', 'vote': 'reject'}
+        ]
+        
+        # Simulate majority voting
+        approve_votes = sum(1 for v in votes if v['vote'] == 'approve')
+        total_votes = len(votes)
+        majority_threshold = total_votes // 2 + 1
+        
+        result = {
+            'decision': 'approve' if approve_votes >= majority_threshold else 'reject',
+            'vote_count': {'approve': approve_votes, 'reject': total_votes - approve_votes},
+            'threshold': majority_threshold
+        }
+        
+        assert result['decision'] == 'approve'
+        assert result['vote_count']['approve'] == 2
+        assert result['vote_count']['reject'] == 1
+        assert result['threshold'] == 2
+    
+    @pytest.mark.asyncio
+    async def test_weighted_voting(self):
+        """Test weighted voting mechanism"""
+        votes = [
+            {'voter_id': 'agent_001', 'vote': 'approve', 'weight': 3},
+            {'voter_id': 'agent_002', 'vote': 'reject', 'weight': 1},
+            {'voter_id': 'agent_003', 'vote': 'approve', 'weight': 2}
+        ]
+        
+        # Calculate weighted votes
+        approve_weight = sum(v['weight'] for v in votes if v['vote'] == 'approve')
+        reject_weight = sum(v['weight'] for v in votes if v['vote'] == 'reject')
+        total_weight = approve_weight + reject_weight
+        
+        result = {
+            'decision': 'approve' if approve_weight > reject_weight else 'reject',
+            'weighted_count': {'approve': approve_weight, 'reject': reject_weight},
+            'total_weight': total_weight
+        }
+        
+        assert result['decision'] == 'approve'
+        assert result['weighted_count']['approve'] == 5
+        assert result['weighted_count']['reject'] == 1
+        assert result['total_weight'] == 6
+    
+    @pytest.mark.asyncio
+    async def test_unanimous_voting(self):
+        """Test unanimous voting mechanism"""
+        votes = [
+            {'voter_id': 'agent_001', 'vote': 'approve'},
+            {'voter_id': 'agent_002', 'vote': 'approve'},
+            {'voter_id': 'agent_003', 'vote': 'approve'}
+        ]
+        
+        # Check for unanimity
+        all_approve = all(v['vote'] == 'approve' for v in votes)
+        
+        result = {
+            'decision': 'approve' if all_approve else 'reject',
+            'unanimous': all_approve,
+            'vote_count': len(votes)
+        }
+        
+        assert result['decision'] == 'approve'
+        assert result['unanimous'] is True
+        assert result['vote_count'] == 3
+
+class TestAgentLifecycleManagement:
+    """Test agent lifecycle management"""
+    
+    def setup_method(self):
+        self.agents = {}
+        self.agent_states = {}
+    
+    @pytest.mark.asyncio
+    async def test_agent_registration(self):
+        """Test agent registration in decision system"""
+        agent_data = {
+            'agent_id': 'agent_001',
+            'capabilities': ['decision_making', 'voting'],
+            'status': 'active',
+            'join_time': datetime.utcnow().isoformat()
+        }
+        
+        self.agents[agent_data['agent_id']] = agent_data
+        
+        assert agent_data['agent_id'] in self.agents
+        assert self.agents[agent_data['agent_id']]['status'] == 'active'
+        assert 'decision_making' in self.agents[agent_data['agent_id']]['capabilities']
+    
+    @pytest.mark.asyncio
+    async def test_agent_status_update(self):
+        """Test agent status updates"""
+        agent_id = 'agent_002'
+        self.agents[agent_id] = {
+            'agent_id': agent_id,
+            'status': 'active',
+            'last_update': datetime.utcnow().isoformat()
+        }
+        
+        # Update agent status
+        self.agents[agent_id]['status'] = 'busy'
+        self.agents[agent_id]['last_update'] = datetime.utcnow().isoformat()
+        
+        assert self.agents[agent_id]['status'] == 'busy'
+        assert 'last_update' in self.agents[agent_id]
+    
+    @pytest.mark.asyncio
+    async def test_agent_removal(self):
+        """Test agent removal from decision system"""
+        agent_id = 'agent_003'
+        self.agents[agent_id] = {
+            'agent_id': agent_id,
+            'status': 'active'
+        }
+        
+        # Remove agent
+        del self.agents[agent_id]
+        
+        assert agent_id not in self.agents
+
+# Integration tests
+class TestDecisionIntegration:
+    """Integration tests for decision framework"""
+    
+    @pytest.mark.asyncio
+    async def test_end_to_end_decision_process(self):
+        """Test complete decision making process"""
+        decision_engine = MockDecisionEngine()
+        consensus = MockConsensusAlgorithm()
+        
+        # Step 1: Create decision proposal
+        decision_data = {
+            'decision_id': 'integration_test_001',
+            'proposal': 'test_proposal',
+            'participants': ['agent_001', 'agent_002']
+        }
+        
+        # Step 2: Make decision
+        decision_result = await decision_engine.make_decision(decision_data)
+        
+        # Step 3: Achieve consensus
+        consensus_result = await consensus.achieve_consensus(
+            decision_data['participants'],
+            {'action': decision_data['proposal']}
+        )
+        
+        # Verify results
+        assert decision_result['status'] == 'completed'
+        assert consensus_result['status'] == 'consensus_reached'
+        assert decision_result['decision_id'] == 'integration_test_001'
+    
+    @pytest.mark.asyncio
+    async def test_multi_agent_coordination(self):
+        """Test coordination between multiple agents"""
+        agents = ['agent_001', 'agent_002', 'agent_003']
+        decision_engine = MockDecisionEngine()
+        
+        # Simulate coordinated decision making
+        decisions = []
+        for i, agent in enumerate(agents):
+            decision_data = {
+                'decision_id': f'coord_test_{i}',
+                'agent_id': agent,
+                'proposal': f'proposal_{i}',
+                'coordinated_with': [a for a in agents if a != agent]
+            }
+            result = await decision_engine.make_decision(decision_data)
+            decisions.append(result)
+        
+        # Verify all decisions were made
+        assert len(decisions) == len(agents)
+        for decision in decisions:
+            assert decision['status'] == 'completed'
+
+if __name__ == '__main__':
+    pytest.main([__file__])
diff --git a/tests/phase4/test_autonomous_decision_making.py b/tests/phase4/test_autonomous_decision_making.py
new file mode 100644
index 00000000..51b991fe
--- /dev/null
+++ b/tests/phase4/test_autonomous_decision_making.py
@@ -0,0 +1,532 @@
+"""
+Phase 4: Autonomous Decision Making Tests
+Tests for autonomous systems, learning, and adaptation
+"""
+
+import pytest
+import asyncio
+import json
+from datetime import datetime, timedelta
+from unittest.mock import Mock, AsyncMock
+from typing import Dict, List, Any, Optional
+
+# Mock imports for testing
+class MockAutonomousEngine:
+    def __init__(self):
+        self.policies = {}
+        self.decisions = []
+        self.learning_data = {}
+        self.performance_metrics = {}
+        
+    async def make_autonomous_decision(self, context: Dict[str, Any]) -> Dict[str, Any]:
+        """Make autonomous decision based on context"""
+        decision_id = f"auto_decision_{len(self.decisions)}"
+        decision = {
+            'decision_id': decision_id,
+            'context': context,
+            'action': self._determine_action(context),
+            'reasoning': self._generate_reasoning(context),
+            'confidence': self._calculate_confidence(context),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+        self.decisions.append(decision)
+        return decision
+    
+    def _determine_action(self, context: Dict[str, Any]) -> str:
+        """Determine action based on context"""
+        if context.get('system_load', 0) > 0.8:
+            return 'scale_resources'
+        elif context.get('error_rate', 0) > 0.1:
+            return 'trigger_recovery'
+        elif context.get('task_queue_size', 0) > 100:
+            return 'allocate_more_agents'
+        else:
+            return 'maintain_status'
+    
+    def _generate_reasoning(self, context: Dict[str, Any]) -> str:
+        """Generate reasoning for decision"""
+        return f"Based on system metrics: load={context.get('system_load', 0)}, errors={context.get('error_rate', 0)}"
+    
+    def _calculate_confidence(self, context: Dict[str, Any]) -> float:
+        """Calculate confidence in decision"""
+        # Simple confidence calculation based on data quality
+        has_metrics = all(key in context for key in ['system_load', 'error_rate'])
+        return 0.9 if has_metrics else 0.6
+
+class MockLearningSystem:
+    def __init__(self):
+        self.experience_buffer = []
+        self.performance_history = []
+        self.adaptations = {}
+        
+    async def learn_from_experience(self, experience: Dict[str, Any]) -> Dict[str, Any]:
+        """Learn from experience"""
+        experience_id = f"exp_{len(self.experience_buffer)}"
+        learning_data = {
+            'experience_id': experience_id,
+            'experience': experience,
+            'lessons_learned': self._extract_lessons(experience),
+            'performance_impact': self._calculate_impact(experience),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+        self.experience_buffer.append(learning_data)
+        return learning_data
+    
+    def _extract_lessons(self, experience: Dict[str, Any]) -> List[str]:
+        """Extract lessons from experience"""
+        lessons = []
+        if experience.get('success', False):
+            lessons.append("Action was successful")
+        if experience.get('performance_gain', 0) > 0:
+            lessons.append("Performance improved")
+        return lessons
+    
+    def _calculate_impact(self, experience: Dict[str, Any]) -> float:
+        """Calculate performance impact"""
+        return experience.get('performance_gain', 0.0)
+    
+    async def adapt_behavior(self, adaptation_data: Dict[str, Any]) -> Dict[str, Any]:
+        """Adapt behavior based on learning"""
+        adaptation_id = f"adapt_{len(self.adaptations)}"
+        adaptation = {
+            'adaptation_id': adaptation_id,
+            'type': adaptation_data.get('type', 'parameter_adjustment'),
+            'changes': adaptation_data.get('changes', {}),
+            'expected_improvement': adaptation_data.get('expected_improvement', 0.1),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+        self.adaptations[adaptation_id] = adaptation
+        return adaptation
+
+class MockPolicyEngine:
+    def __init__(self):
+        self.policies = {
+            'resource_management': {
+                'max_cpu_usage': 0.8,
+                'max_memory_usage': 0.85,
+                'auto_scale_threshold': 0.7
+            },
+            'error_handling': {
+                'max_error_rate': 0.05,
+                'retry_attempts': 3,
+                'recovery_timeout': 300
+            },
+            'task_management': {
+                'max_queue_size': 1000,
+                'task_timeout': 600,
+                'priority_weights': {'high': 1.0, 'normal': 0.5, 'low': 0.2}
+            }
+        }
+        
+    async def evaluate_policy_compliance(self, decision: Dict[str, Any]) -> Dict[str, Any]:
+        """Evaluate if decision complies with policies"""
+        compliance_score = self._calculate_compliance(decision)
+        violations = self._find_violations(decision)
+        
+        return {
+            'decision_id': decision.get('decision_id'),
+            'compliance_score': compliance_score,
+            'violations': violations,
+            'approved': compliance_score >= 0.8 and len(violations) == 0,
+            'timestamp': datetime.utcnow().isoformat()
+        }
+    
+    def _calculate_compliance(self, decision: Dict[str, Any]) -> float:
+        """Calculate policy compliance score"""
+        # Simplified compliance calculation
+        base_score = 1.0
+        if decision.get('action') == 'scale_resources':
+            # Check resource management policy
+            base_score -= 0.1  # Small penalty for resource scaling
+        return max(0.0, base_score)
+    
+    def _find_violations(self, decision: Dict[str, Any]) -> List[str]:
+        """Find policy violations"""
+        violations = []
+        context = decision.get('context', {})
+        
+        # Check resource limits
+        if context.get('system_load', 0) > self.policies['resource_management']['max_cpu_usage']:
+            violations.append("CPU usage exceeds policy limit")
+        
+        return violations
+
+class TestAutonomousEngine:
+    """Test autonomous decision making engine"""
+    
+    def setup_method(self):
+        self.autonomous_engine = MockAutonomousEngine()
+    
+    @pytest.mark.asyncio
+    async def test_autonomous_decision_making(self):
+        """Test basic autonomous decision making"""
+        context = {
+            'system_load': 0.9,
+            'error_rate': 0.02,
+            'task_queue_size': 50,
+            'active_agents': 5
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        assert decision['action'] == 'scale_resources'
+        assert decision['confidence'] > 0.5
+        assert 'reasoning' in decision
+        assert 'timestamp' in decision
+    
+    @pytest.mark.asyncio
+    async def test_decision_with_high_error_rate(self):
+        """Test decision making with high error rate"""
+        context = {
+            'system_load': 0.4,
+            'error_rate': 0.15,
+            'task_queue_size': 30,
+            'active_agents': 3
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        assert decision['action'] == 'trigger_recovery'
+        assert 'error_rate' in decision['reasoning']
+    
+    @pytest.mark.asyncio
+    async def test_decision_with_task_queue_pressure(self):
+        """Test decision making with task queue pressure"""
+        context = {
+            'system_load': 0.6,
+            'error_rate': 0.03,
+            'task_queue_size': 150,
+            'active_agents': 4
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        assert decision['action'] == 'allocate_more_agents'
+    
+    @pytest.mark.asyncio
+    async def test_decision_with_normal_conditions(self):
+        """Test decision making with normal conditions"""
+        context = {
+            'system_load': 0.5,
+            'error_rate': 0.02,
+            'task_queue_size': 25,
+            'active_agents': 4
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        assert decision['action'] == 'maintain_status'
+        assert decision['confidence'] > 0.8
+
+class TestLearningSystem:
+    """Test learning and adaptation system"""
+    
+    def setup_method(self):
+        self.learning_system = MockLearningSystem()
+    
+    @pytest.mark.asyncio
+    async def test_learning_from_successful_experience(self):
+        """Test learning from successful experience"""
+        experience = {
+            'action': 'scale_resources',
+            'success': True,
+            'performance_gain': 0.15,
+            'context': {'system_load': 0.9}
+        }
+        
+        learning_result = await self.learning_system.learn_from_experience(experience)
+        
+        assert learning_result['experience_id'].startswith('exp_')
+        assert 'lessons_learned' in learning_result
+        assert learning_result['performance_impact'] == 0.15
+        assert 'Action was successful' in learning_result['lessons_learned']
+    
+    @pytest.mark.asyncio
+    async def test_learning_from_failure(self):
+        """Test learning from failed experience"""
+        experience = {
+            'action': 'scale_resources',
+            'success': False,
+            'performance_gain': -0.05,
+            'context': {'system_load': 0.9}
+        }
+        
+        learning_result = await self.learning_system.learn_from_experience(experience)
+        
+        assert learning_result['experience_id'].startswith('exp_')
+        assert learning_result['performance_impact'] == -0.05
+    
+    @pytest.mark.asyncio
+    async def test_behavior_adaptation(self):
+        """Test behavior adaptation based on learning"""
+        adaptation_data = {
+            'type': 'threshold_adjustment',
+            'changes': {'scale_threshold': 0.75, 'error_threshold': 0.08},
+            'expected_improvement': 0.1
+        }
+        
+        adaptation = await self.learning_system.adapt_behavior(adaptation_data)
+        
+        assert adaptation['type'] == 'threshold_adjustment'
+        assert adaptation['expected_improvement'] == 0.1
+        assert 'scale_threshold' in adaptation['changes']
+    
+    @pytest.mark.asyncio
+    async def test_experience_accumulation(self):
+        """Test accumulation of experiences over time"""
+        experiences = [
+            {'action': 'scale_resources', 'success': True, 'performance_gain': 0.1},
+            {'action': 'allocate_agents', 'success': True, 'performance_gain': 0.05},
+            {'action': 'trigger_recovery', 'success': False, 'performance_gain': -0.02}
+        ]
+        
+        for exp in experiences:
+            await self.learning_system.learn_from_experience(exp)
+        
+        assert len(self.learning_system.experience_buffer) == 3
+        assert all(exp['experience_id'].startswith('exp_') for exp in self.learning_system.experience_buffer)
+
+class TestPolicyEngine:
+    """Test policy engine for autonomous decisions"""
+    
+    def setup_method(self):
+        self.policy_engine = MockPolicyEngine()
+    
+    @pytest.mark.asyncio
+    async def test_policy_compliance_evaluation(self):
+        """Test policy compliance evaluation"""
+        decision = {
+            'decision_id': 'test_decision_001',
+            'action': 'scale_resources',
+            'context': {
+                'system_load': 0.7,
+                'error_rate': 0.03,
+                'task_queue_size': 50
+            }
+        }
+        
+        compliance = await self.policy_engine.evaluate_policy_compliance(decision)
+        
+        assert compliance['decision_id'] == 'test_decision_001'
+        assert 'compliance_score' in compliance
+        assert 'violations' in compliance
+        assert 'approved' in compliance
+        assert 'timestamp' in compliance
+    
+    @pytest.mark.asyncio
+    async def test_policy_violation_detection(self):
+        """Test detection of policy violations"""
+        decision = {
+            'decision_id': 'test_decision_002',
+            'action': 'scale_resources',
+            'context': {
+                'system_load': 0.9,  # Exceeds policy limit
+                'error_rate': 0.03,
+                'task_queue_size': 50
+            }
+        }
+        
+        compliance = await self.policy_engine.evaluate_policy_compliance(decision)
+        
+        assert len(compliance['violations']) > 0
+        assert any('CPU usage' in violation for violation in compliance['violations'])
+    
+    @pytest.mark.asyncio
+    async def test_policy_approval(self):
+        """Test policy approval for compliant decisions"""
+        decision = {
+            'decision_id': 'test_decision_003',
+            'action': 'maintain_status',
+            'context': {
+                'system_load': 0.5,
+                'error_rate': 0.02,
+                'task_queue_size': 25
+            }
+        }
+        
+        compliance = await self.policy_engine.evaluate_policy_compliance(decision)
+        
+        assert compliance['approved'] is True
+        assert compliance['compliance_score'] >= 0.8
+
+class TestSelfCorrectionMechanism:
+    """Test self-correction mechanisms"""
+    
+    def setup_method(self):
+        self.autonomous_engine = MockAutonomousEngine()
+        self.learning_system = MockLearningSystem()
+        self.policy_engine = MockPolicyEngine()
+    
+    @pytest.mark.asyncio
+    async def test_automatic_error_correction(self):
+        """Test automatic error correction"""
+        # Simulate error condition
+        context = {
+            'system_load': 0.9,
+            'error_rate': 0.12,  # High error rate
+            'task_queue_size': 50
+        }
+        
+        # Make initial decision
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        # Simulate error in execution
+        error_experience = {
+            'action': decision['action'],
+            'success': False,
+            'error_type': 'resource_exhaustion',
+            'performance_gain': -0.1
+        }
+        
+        # Learn from error
+        learning_result = await self.learning_system.learn_from_experience(error_experience)
+        
+        # Adapt behavior
+        adaptation_data = {
+            'type': 'resource_threshold_adjustment',
+            'changes': {'scale_threshold': 0.8},
+            'expected_improvement': 0.15
+        }
+        
+        adaptation = await self.learning_system.adapt_behavior(adaptation_data)
+        
+        # Verify self-correction
+        assert decision['action'] == 'trigger_recovery'
+        assert learning_result['experience_id'].startswith('exp_')
+        assert adaptation['type'] == 'resource_threshold_adjustment'
+    
+    @pytest.mark.asyncio
+    async def test_performance_optimization(self):
+        """Test performance optimization through learning"""
+        # Initial performance
+        initial_context = {
+            'system_load': 0.7,
+            'error_rate': 0.05,
+            'task_queue_size': 80
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(initial_context)
+        
+        # Simulate successful execution with performance gain
+        success_experience = {
+            'action': decision['action'],
+            'success': True,
+            'performance_gain': 0.2
+        }
+        
+        learning_result = await self.learning_system.learn_from_experience(success_experience)
+        
+        # Adapt to optimize further
+        adaptation_data = {
+            'type': 'performance_optimization',
+            'changes': {'aggressive_scaling': True},
+            'expected_improvement': 0.1
+        }
+        
+        adaptation = await self.learning_system.adapt_behavior(adaptation_data)
+        
+        # Verify optimization
+        assert learning_result['performance_impact'] == 0.2
+        assert adaptation['type'] == 'performance_optimization'
+    
+    @pytest.mark.asyncio
+    async def test_goal_oriented_behavior(self):
+        """Test goal-oriented autonomous behavior"""
+        # Define goals
+        goals = {
+            'primary_goal': 'maintain_system_stability',
+            'secondary_goals': ['optimize_performance', 'minimize_errors'],
+            'constraints': ['resource_limits', 'policy_compliance']
+        }
+        
+        # Simulate goal-oriented decision making
+        context = {
+            'system_load': 0.6,
+            'error_rate': 0.04,
+            'task_queue_size': 60,
+            'goals': goals
+        }
+        
+        decision = await self.autonomous_engine.make_autonomous_decision(context)
+        
+        # Evaluate against goals
+        compliance = await self.policy_engine.evaluate_policy_compliance(decision)
+        
+        # Verify goal alignment
+        assert decision['action'] in ['maintain_status', 'allocate_more_agents']
+        assert compliance['approved'] is True  # Should be policy compliant
+
+# Integration tests
+class TestAutonomousIntegration:
+    """Integration tests for autonomous systems"""
+    
+    @pytest.mark.asyncio
+    async def test_full_autonomous_cycle(self):
+        """Test complete autonomous decision cycle"""
+        autonomous_engine = MockAutonomousEngine()
+        learning_system = MockLearningSystem()
+        policy_engine = MockPolicyEngine()
+        
+        # Step 1: Make autonomous decision
+        context = {
+            'system_load': 0.85,
+            'error_rate': 0.08,
+            'task_queue_size': 120
+        }
+        
+        decision = await autonomous_engine.make_autonomous_decision(context)
+        
+        # Step 2: Evaluate policy compliance
+        compliance = await policy_engine.evaluate_policy_compliance(decision)
+        
+        # Step 3: Execute and learn from result
+        execution_result = {
+            'action': decision['action'],
+            'success': compliance['approved'],
+            'performance_gain': 0.1 if compliance['approved'] else -0.05
+        }
+        
+        learning_result = await learning_system.learn_from_experience(execution_result)
+        
+        # Step 4: Adapt if needed
+        if not compliance['approved']:
+            adaptation = await learning_system.adapt_behavior({
+                'type': 'policy_compliance',
+                'changes': {'more_conservative_thresholds': True}
+            })
+        
+        # Verify complete cycle
+        assert decision['decision_id'].startswith('auto_decision_')
+        assert 'compliance_score' in compliance
+        assert learning_result['experience_id'].startswith('exp_')
+    
+    @pytest.mark.asyncio
+    async def test_multi_goal_optimization(self):
+        """Test optimization across multiple goals"""
+        goals = {
+            'stability': {'weight': 0.4, 'target': 0.95},
+            'performance': {'weight': 0.3, 'target': 0.8},
+            'efficiency': {'weight': 0.3, 'target': 0.75}
+        }
+        
+        contexts = [
+            {'system_load': 0.7, 'error_rate': 0.05, 'goals': goals},
+            {'system_load': 0.8, 'error_rate': 0.06, 'goals': goals},
+            {'system_load': 0.6, 'error_rate': 0.04, 'goals': goals}
+        ]
+        
+        autonomous_engine = MockAutonomousEngine()
+        decisions = []
+        
+        for context in contexts:
+            decision = await autonomous_engine.make_autonomous_decision(context)
+            decisions.append(decision)
+        
+        # Verify multi-goal consideration
+        assert len(decisions) == 3
+        for decision in decisions:
+            assert 'action' in decision
+            assert 'confidence' in decision
+
+if __name__ == '__main__':
+    pytest.main([__file__])
diff --git a/tests/phase5/test_vision_integration.py b/tests/phase5/test_vision_integration.py
new file mode 100644
index 00000000..7e95e52c
--- /dev/null
+++ b/tests/phase5/test_vision_integration.py
@@ -0,0 +1,641 @@
+"""
+Phase 5: Computer Vision Integration Tests
+Tests for visual intelligence, image processing, and multi-modal integration
+"""
+
+import pytest
+import asyncio
+import json
+import base64
+from datetime import datetime, timedelta
+from unittest.mock import Mock, AsyncMock
+from typing import Dict, List, Any, Optional, Tuple
+
+# Mock imports for testing
+class MockVisionProcessor:
+    def __init__(self):
+        self.processed_images = {}
+        self.detection_results = {}
+        self.analysis_results = {}
+        
+    async def process_image(self, image_data: bytes, processing_type: str = 'general') -> Dict[str, Any]:
+        """Process image data"""
+        image_id = f"img_{len(self.processed_images)}"
+        result = {
+            'image_id': image_id,
+            'processing_type': processing_type,
+            'size': len(image_data),
+            'format': 'processed',
+            'timestamp': datetime.utcnow().isoformat(),
+            'analysis': await self._analyze_image(image_data, processing_type)
+        }
+        self.processed_images[image_id] = result
+        return result
+    
+    async def _analyze_image(self, image_data: bytes, processing_type: str) -> Dict[str, Any]:
+        """Analyze image based on processing type"""
+        if processing_type == 'object_detection':
+            return await self._detect_objects(image_data)
+        elif processing_type == 'scene_analysis':
+            return await self._analyze_scene(image_data)
+        elif processing_type == 'text_extraction':
+            return await self._extract_text(image_data)
+        else:
+            return await self._general_analysis(image_data)
+    
+    async def _detect_objects(self, image_data: bytes) -> Dict[str, Any]:
+        """Detect objects in image"""
+        # Mock object detection
+        objects = [
+            {'class': 'person', 'confidence': 0.92, 'bbox': [100, 150, 200, 300]},
+            {'class': 'car', 'confidence': 0.87, 'bbox': [300, 200, 500, 350]},
+            {'class': 'building', 'confidence': 0.95, 'bbox': [0, 0, 600, 400]}
+        ]
+        
+        self.detection_results[f"detection_{len(self.detection_results)}"] = objects
+        
+        return {
+            'objects_detected': len(objects),
+            'objects': objects,
+            'detection_confidence': sum(obj['confidence'] for obj in objects) / len(objects)
+        }
+    
+    async def _analyze_scene(self, image_data: bytes) -> Dict[str, Any]:
+        """Analyze scene context"""
+        # Mock scene analysis
+        scene_info = {
+            'scene_type': 'urban_street',
+            'confidence': 0.88,
+            'elements': ['vehicles', 'pedestrians', 'buildings'],
+            'weather': 'clear',
+            'time_of_day': 'daytime',
+            'complexity': 'medium'
+        }
+        
+        return scene_info
+    
+    async def _extract_text(self, image_data: bytes) -> Dict[str, Any]:
+        """Extract text from image"""
+        # Mock OCR
+        text_data = {
+            'text_found': True,
+            'extracted_text': ['STOP', 'MAIN ST', 'NO PARKING'],
+            'confidence': 0.91,
+            'language': 'en',
+            'text_regions': [
+                {'text': 'STOP', 'bbox': [50, 100, 150, 150]},
+                {'text': 'MAIN ST', 'bbox': [200, 100, 350, 150]}
+            ]
+        }
+        
+        return text_data
+    
+    async def _general_analysis(self, image_data: bytes) -> Dict[str, Any]:
+        """General image analysis"""
+        return {
+            'brightness': 0.7,
+            'contrast': 0.8,
+            'sharpness': 0.75,
+            'color_distribution': {'red': 0.3, 'green': 0.4, 'blue': 0.3},
+            'dominant_colors': ['blue', 'green', 'white'],
+            'image_quality': 'good'
+        }
+
+class MockMultiModalAgent:
+    def __init__(self):
+        self.vision_processor = MockVisionProcessor()
+        self.integrated_results = {}
+        
+    async def process_multi_modal(self, inputs: Dict[str, Any]) -> Dict[str, Any]:
+        """Process multi-modal inputs"""
+        result_id = f"multi_{len(self.integrated_results)}"
+        
+        # Process different modalities
+        results = {}
+        
+        if 'image' in inputs:
+            results['vision'] = await self.vision_processor.process_image(
+                inputs['image'], 
+                inputs.get('vision_processing_type', 'general')
+            )
+        
+        if 'text' in inputs:
+            results['text'] = await self._process_text(inputs['text'])
+        
+        if 'sensor_data' in inputs:
+            results['sensor'] = await self._process_sensor_data(inputs['sensor_data'])
+        
+        # Integrate results
+        integrated_result = {
+            'result_id': result_id,
+            'modalities_processed': list(results.keys()),
+            'integration': await self._integrate_modalities(results),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+        
+        self.integrated_results[result_id] = integrated_result
+        return integrated_result
+    
+    async def _process_text(self, text: str) -> Dict[str, Any]:
+        """Process text input"""
+        return {
+            'text_length': len(text),
+            'language': 'en',
+            'sentiment': 'neutral',
+            'entities': [],
+            'keywords': text.split()[:5]
+        }
+    
+    async def _process_sensor_data(self, sensor_data: Dict[str, Any]) -> Dict[str, Any]:
+        """Process sensor data"""
+        return {
+            'sensor_type': sensor_data.get('type', 'unknown'),
+            'readings': sensor_data.get('readings', {}),
+            'timestamp': sensor_data.get('timestamp', datetime.utcnow().isoformat()),
+            'quality': 'good'
+        }
+    
+    async def _integrate_modalities(self, results: Dict[str, Any]) -> Dict[str, Any]:
+        """Integrate results from different modalities"""
+        integration = {
+            'confidence': 0.85,
+            'completeness': len(results) / 3.0,  # Assuming 3 modalities max
+            'cross_modal_insights': []
+        }
+        
+        # Add cross-modal insights
+        if 'vision' in results and 'text' in results:
+            if 'objects' in results['vision'].get('analysis', {}):
+                integration['cross_modal_insights'].append(
+                    f"Visual context: {len(results['vision']['analysis']['objects'])} objects detected"
+                )
+        
+        return integration
+
+class MockContextIntegration:
+    def __init__(self):
+        self.context_history = []
+        self.context_models = {}
+        
+    async def integrate_context(self, vision_result: Dict[str, Any], context_data: Dict[str, Any]) -> Dict[str, Any]:
+        """Integrate vision results with context"""
+        context_id = f"ctx_{len(self.context_history)}"
+        
+        integration = {
+            'context_id': context_id,
+            'vision_result': vision_result,
+            'context_data': context_data,
+            'enhanced_understanding': await self._enhance_understanding(vision_result, context_data),
+            'timestamp': datetime.utcnow().isoformat()
+        }
+        
+        self.context_history.append(integration)
+        return integration
+    
+    async def _enhance_understanding(self, vision_result: Dict[str, Any], context_data: Dict[str, Any]) -> Dict[str, Any]:
+        """Enhance understanding with context"""
+        enhanced = {
+            'scene_understanding': vision_result.get('analysis', {}),
+            'contextual_insights': [],
+            'confidence_boost': 0.0
+        }
+        
+        # Add contextual insights
+        if context_data.get('location') == 'intersection':
+            enhanced['contextual_insights'].append("Traffic monitoring context")
+            enhanced['confidence_boost'] += 0.1
+        
+        if context_data.get('time_of_day') == 'night':
+            enhanced['contextual_insights'].append("Low light conditions detected")
+            enhanced['confidence_boost'] -= 0.05
+        
+        return enhanced
+
+class TestVisionProcessor:
+    """Test vision processing functionality"""
+    
+    def setup_method(self):
+        self.vision_processor = MockVisionProcessor()
+        self.sample_image = b'sample_image_data_for_testing'
+    
+    @pytest.mark.asyncio
+    async def test_image_processing(self):
+        """Test basic image processing"""
+        result = await self.vision_processor.process_image(self.sample_image)
+        
+        assert result['image_id'].startswith('img_')
+        assert result['size'] == len(self.sample_image)
+        assert result['format'] == 'processed'
+        assert 'analysis' in result
+        assert 'timestamp' in result
+    
+    @pytest.mark.asyncio
+    async def test_object_detection(self):
+        """Test object detection functionality"""
+        result = await self.vision_processor.process_image(self.sample_image, 'object_detection')
+        
+        assert 'analysis' in result
+        assert 'objects_detected' in result['analysis']
+        assert 'objects' in result['analysis']
+        assert len(result['analysis']['objects']) > 0
+        
+        # Check object structure
+        for obj in result['analysis']['objects']:
+            assert 'class' in obj
+            assert 'confidence' in obj
+            assert 'bbox' in obj
+            assert 0 <= obj['confidence'] <= 1
+    
+    @pytest.mark.asyncio
+    async def test_scene_analysis(self):
+        """Test scene analysis functionality"""
+        result = await self.vision_processor.process_image(self.sample_image, 'scene_analysis')
+        
+        assert 'analysis' in result
+        assert 'scene_type' in result['analysis']
+        assert 'confidence' in result['analysis']
+        assert 'elements' in result['analysis']
+        
+        assert result['analysis']['scene_type'] == 'urban_street'
+        assert 0 <= result['analysis']['confidence'] <= 1
+    
+    @pytest.mark.asyncio
+    async def test_text_extraction(self):
+        """Test text extraction (OCR) functionality"""
+        result = await self.vision_processor.process_image(self.sample_image, 'text_extraction')
+        
+        assert 'analysis' in result
+        assert 'text_found' in result['analysis']
+        assert 'extracted_text' in result['analysis']
+        
+        if result['analysis']['text_found']:
+            assert len(result['analysis']['extracted_text']) > 0
+            assert 'confidence' in result['analysis']
+    
+    @pytest.mark.asyncio
+    async def test_general_analysis(self):
+        """Test general image analysis"""
+        result = await self.vision_processor.process_image(self.sample_image, 'general')
+        
+        assert 'analysis' in result
+        assert 'brightness' in result['analysis']
+        assert 'contrast' in result['analysis']
+        assert 'sharpness' in result['analysis']
+        assert 'color_distribution' in result['analysis']
+        
+        # Check value ranges
+        assert 0 <= result['analysis']['brightness'] <= 1
+        assert 0 <= result['analysis']['contrast'] <= 1
+        assert 0 <= result['analysis']['sharpness'] <= 1
+
+class TestMultiModalIntegration:
+    """Test multi-modal integration"""
+    
+    def setup_method(self):
+        self.multi_modal_agent = MockMultiModalAgent()
+        self.sample_image = b'sample_image_data'
+        self.sample_text = "This is a sample text for testing"
+        self.sample_sensor_data = {
+            'type': 'temperature',
+            'readings': {'value': 25.5, 'unit': 'celsius'},
+            'timestamp': datetime.utcnow().isoformat()
+        }
+    
+    @pytest.mark.asyncio
+    async def test_vision_only_processing(self):
+        """Test processing with only vision input"""
+        inputs = {'image': self.sample_image}
+        
+        result = await self.multi_modal_agent.process_multi_modal(inputs)
+        
+        assert result['result_id'].startswith('multi_')
+        assert 'vision' in result['modalities_processed']
+        assert 'integration' in result
+        assert 'confidence' in result['integration']
+    
+    @pytest.mark.asyncio
+    async def test_text_only_processing(self):
+        """Test processing with only text input"""
+        inputs = {'text': self.sample_text}
+        
+        result = await self.multi_modal_agent.process_multi_modal(inputs)
+        
+        assert result['result_id'].startswith('multi_')
+        assert 'text' in result['modalities_processed']
+        assert 'integration' in result
+    
+    @pytest.mark.asyncio
+    async def test_sensor_only_processing(self):
+        """Test processing with only sensor input"""
+        inputs = {'sensor_data': self.sample_sensor_data}
+        
+        result = await self.multi_modal_agent.process_multi_modal(inputs)
+        
+        assert result['result_id'].startswith('multi_')
+        assert 'sensor' in result['modalities_processed']
+        assert 'integration' in result
+    
+    @pytest.mark.asyncio
+    async def test_full_multi_modal_processing(self):
+        """Test processing with all modalities"""
+        inputs = {
+            'image': self.sample_image,
+            'text': self.sample_text,
+            'sensor_data': self.sample_sensor_data
+        }
+        
+        result = await self.multi_modal_agent.process_multi_modal(inputs)
+        
+        assert result['result_id'].startswith('multi_')
+        assert len(result['modalities_processed']) == 3
+        assert 'vision' in result['modalities_processed']
+        assert 'text' in result['modalities_processed']
+        assert 'sensor' in result['modalities_processed']
+        assert 'integration' in result
+        assert 'cross_modal_insights' in result['integration']
+    
+    @pytest.mark.asyncio
+    async def test_cross_modal_insights(self):
+        """Test cross-modal insight generation"""
+        inputs = {
+            'image': self.sample_image,
+            'text': self.sample_text,
+            'vision_processing_type': 'object_detection'
+        }
+        
+        result = await self.multi_modal_agent.process_multi_modal(inputs)
+        
+        assert 'cross_modal_insights' in result['integration']
+        assert len(result['integration']['cross_modal_insights']) > 0
+
+class TestContextIntegration:
+    """Test context integration with vision"""
+    
+    def setup_method(self):
+        self.context_integration = MockContextIntegration()
+        self.vision_processor = MockVisionProcessor()
+        self.sample_image = b'sample_image_data'
+    
+    @pytest.mark.asyncio
+    async def test_basic_context_integration(self):
+        """Test basic context integration"""
+        vision_result = await self.vision_processor.process_image(self.sample_image)
+        context_data = {
+            'location': 'intersection',
+            'time_of_day': 'daytime',
+            'weather': 'clear'
+        }
+        
+        result = await self.context_integration.integrate_context(vision_result, context_data)
+        
+        assert result['context_id'].startswith('ctx_')
+        assert 'vision_result' in result
+        assert 'context_data' in result
+        assert 'enhanced_understanding' in result
+    
+    @pytest.mark.asyncio
+    async def test_location_context(self):
+        """Test location-based context integration"""
+        vision_result = await self.vision_processor.process_image(self.sample_image, 'object_detection')
+        context_data = {
+            'location': 'intersection',
+            'traffic_flow': 'moderate'
+        }
+        
+        result = await self.context_integration.integrate_context(vision_result, context_data)
+        
+        assert 'enhanced_understanding' in result
+        assert 'contextual_insights' in result['enhanced_understanding']
+        assert any('traffic' in insight for insight in result['enhanced_understanding']['contextual_insights'])
+    
+    @pytest.mark.asyncio
+    async def test_time_context(self):
+        """Test time-based context integration"""
+        vision_result = await self.vision_processor.process_image(self.sample_image)
+        context_data = {
+            'time_of_day': 'night',
+            'lighting_conditions': 'low'
+        }
+        
+        result = await self.context_integration.integrate_context(vision_result, context_data)
+        
+        assert 'enhanced_understanding' in result
+        assert 'confidence_boost' in result['enhanced_understanding']
+        assert result['enhanced_understanding']['confidence_boost'] < 0  # Night time penalty
+    
+    @pytest.mark.asyncio
+    async def test_context_history_tracking(self):
+        """Test context history tracking"""
+        for i in range(3):
+            vision_result = await self.vision_processor.process_image(self.sample_image)
+            context_data = {
+                'location': f'location_{i}',
+                'timestamp': datetime.utcnow().isoformat()
+            }
+            await self.context_integration.integrate_context(vision_result, context_data)
+        
+        assert len(self.context_integration.context_history) == 3
+        for context in self.context_integration.context_history:
+            assert context['context_id'].startswith('ctx_')
+
+class TestVisualReasoning:
+    """Test visual reasoning capabilities"""
+    
+    def setup_method(self):
+        self.vision_processor = MockVisionProcessor()
+        self.multi_modal_agent = MockMultiModalAgent()
+        self.sample_image = b'sample_image_data'
+    
+    @pytest.mark.asyncio
+    async def test_visual_scene_understanding(self):
+        """Test visual scene understanding"""
+        result = await self.vision_processor.process_image(self.sample_image, 'scene_analysis')
+        
+        assert 'analysis' in result
+        assert 'scene_type' in result['analysis']
+        assert 'elements' in result['analysis']
+        assert 'complexity' in result['analysis']
+        
+        # Verify scene understanding
+        scene = result['analysis']
+        assert len(scene['elements']) > 0
+        assert scene['complexity'] in ['low', 'medium', 'high']
+    
+    @pytest.mark.asyncio
+    async def test_object_relationships(self):
+        """Test understanding object relationships"""
+        result = await self.vision_processor.process_image(self.sample_image, 'object_detection')
+        
+        assert 'analysis' in result
+        assert 'objects' in result['analysis']
+        
+        objects = result['analysis']['objects']
+        if len(objects) > 1:
+            # Mock relationship analysis
+            relationships = []
+            for i, obj1 in enumerate(objects):
+                for obj2 in objects[i+1:]:
+                    if obj1['class'] == 'person' and obj2['class'] == 'car':
+                        relationships.append('person_near_car')
+            
+            assert len(relationships) >= 0
+    
+    @pytest.mark.asyncio
+    async def test_spatial_reasoning(self):
+        """Test spatial reasoning"""
+        result = await self.vision_processor.process_image(self.sample_image, 'object_detection')
+        
+        assert 'analysis' in result
+        assert 'objects' in result['analysis']
+        
+        objects = result['analysis']['objects']
+        for obj in objects:
+            assert 'bbox' in obj
+            assert len(obj['bbox']) == 4  # [x1, y1, x2, y2]
+            
+            # Verify bbox coordinates
+            x1, y1, x2, y2 = obj['bbox']
+            assert x2 > x1
+            assert y2 > y1
+    
+    @pytest.mark.asyncio
+    async def test_temporal_reasoning(self):
+        """Test temporal reasoning (changes over time)"""
+        # Simulate processing multiple images over time
+        results = []
+        for i in range(3):
+            result = await self.vision_processor.process_image(self.sample_image)
+            results.append(result)
+            await asyncio.sleep(0.01)  # Small delay
+        
+        # Analyze temporal changes
+        if len(results) > 1:
+            # Mock temporal analysis
+            changes = []
+            for i in range(1, len(results)):
+                if results[i]['analysis'] != results[i-1]['analysis']:
+                    changes.append(f"Change detected at step {i}")
+            
+            # Should have some analysis of changes
+            assert len(results) == 3
+
+class TestPerformanceMetrics:
+    """Test performance metrics for vision processing"""
+    
+    def setup_method(self):
+        self.vision_processor = MockVisionProcessor()
+        self.sample_image = b'sample_image_data'
+    
+    @pytest.mark.asyncio
+    async def test_processing_speed(self):
+        """Test image processing speed"""
+        start_time = datetime.utcnow()
+        
+        result = await self.vision_processor.process_image(self.sample_image)
+        
+        end_time = datetime.utcnow()
+        processing_time = (end_time - start_time).total_seconds()
+        
+        assert processing_time < 2.0  # Should process within 2 seconds
+        assert result['image_id'].startswith('img_')
+    
+    @pytest.mark.asyncio
+    async def test_batch_processing(self):
+        """Test batch image processing"""
+        images = [self.sample_image] * 5
+        
+        start_time = datetime.utcnow()
+        results = []
+        for image in images:
+            result = await self.vision_processor.process_image(image)
+            results.append(result)
+        end_time = datetime.utcnow()
+        
+        total_time = (end_time - start_time).total_seconds()
+        avg_time = total_time / len(images)
+        
+        assert len(results) == 5
+        assert avg_time < 1.0  # Average should be under 1 second per image
+    
+    @pytest.mark.asyncio
+    async def test_memory_usage(self):
+        """Test memory usage during processing"""
+        import psutil
+        import os
+        
+        process = psutil.Process(os.getpid())
+        memory_before = process.memory_info().rss
+        
+        # Process multiple images
+        for i in range(10):
+            await self.vision_processor.process_image(self.sample_image)
+        
+        memory_after = process.memory_info().rss
+        memory_increase = memory_after - memory_before
+        
+        # Memory increase should be reasonable (less than 100MB)
+        assert memory_increase < 100 * 1024 * 1024  # 100MB in bytes
+
+# Integration tests
+class TestVisionIntegration:
+    """Integration tests for vision system"""
+    
+    @pytest.mark.asyncio
+    async def test_end_to_end_vision_pipeline(self):
+        """Test complete vision processing pipeline"""
+        vision_processor = MockVisionProcessor()
+        multi_modal_agent = MockMultiModalAgent()
+        context_integration = MockContextIntegration()
+        
+        # Step 1: Process image with object detection
+        image_result = await vision_processor.process_image(b'test_image', 'object_detection')
+        
+        # Step 2: Integrate with context
+        context_data = {
+            'location': 'urban_intersection',
+            'time': 'daytime',
+            'purpose': 'traffic_monitoring'
+        }
+        
+        context_result = await context_integration.integrate_context(image_result, context_data)
+        
+        # Step 3: Multi-modal processing
+        multi_modal_inputs = {
+            'image': b'test_image',
+            'text': 'Traffic monitoring report',
+            'sensor_data': {'type': 'camera', 'status': 'active'}
+        }
+        
+        multi_modal_result = await multi_modal_agent.process_multi_modal(multi_modal_inputs)
+        
+        # Verify pipeline
+        assert image_result['image_id'].startswith('img_')
+        assert context_result['context_id'].startswith('ctx_')
+        assert multi_modal_result['result_id'].startswith('multi_')
+        assert 'objects' in image_result['analysis']
+        assert 'enhanced_understanding' in context_result
+        assert len(multi_modal_result['modalities_processed']) == 3
+    
+    @pytest.mark.asyncio
+    async def test_real_time_vision_processing(self):
+        """Test real-time vision processing capabilities"""
+        vision_processor = MockVisionProcessor()
+        
+        # Simulate real-time processing
+        processing_times = []
+        for i in range(10):
+            start_time = datetime.utcnow()
+            await vision_processor.process_image(f'frame_{i}'.encode())
+            end_time = datetime.utcnow()
+            processing_times.append((end_time - start_time).total_seconds())
+        
+        avg_time = sum(processing_times) / len(processing_times)
+        max_time = max(processing_times)
+        
+        # Real-time constraints
+        assert avg_time < 0.5  # Average under 500ms
+        assert max_time < 1.0  # Max under 1 second
+        assert len(processing_times) == 10
+
+if __name__ == '__main__':
+    pytest.main([__file__])
diff --git a/tests/run_all_phase_tests.py b/tests/run_all_phase_tests.py
new file mode 100644
index 00000000..5a8ca987
--- /dev/null
+++ b/tests/run_all_phase_tests.py
@@ -0,0 +1,100 @@
+#!/usr/bin/env python3
+"""
+Run all phase tests for agent systems implementation
+"""
+
+import subprocess
+import sys
+import os
+from pathlib import Path
+
+def run_phase_tests():
+    """Run tests for all phases"""
+    base_dir = Path(__file__).parent
+    phases = ['phase1', 'phase2', 'phase3', 'phase4', 'phase5']
+    
+    results = {}
+    
+    for phase in phases:
+        phase_dir = base_dir / phase
+        print(f"\n{'='*60}")
+        print(f"Running {phase.upper()} Tests")
+        print(f"{'='*60}")
+        
+        if not phase_dir.exists():
+            print(f"❌ {phase} directory not found")
+            results[phase] = {'status': 'skipped', 'reason': 'directory_not_found'}
+            continue
+        
+        # Find test files
+        test_files = list(phase_dir.glob('test_*.py'))
+        
+        if not test_files:
+            print(f"❌ No test files found in {phase}")
+            results[phase] = {'status': 'skipped', 'reason': 'no_test_files'}
+            continue
+        
+        # Run tests for this phase
+        phase_results = {}
+        for test_file in test_files:
+            print(f"\n🔹 Running {test_file.name}")
+            try:
+                result = subprocess.run([
+                    sys.executable, '-m', 'pytest', 
+                    str(test_file), 
+                    '-v', 
+                    '--tb=short'
+                ], capture_output=True, text=True, cwd=base_dir)
+                
+                phase_results[test_file.name] = {
+                    'returncode': result.returncode,
+                    'stdout': result.stdout,
+                    'stderr': result.stderr
+                }
+                
+                if result.returncode == 0:
+                    print(f"✅ {test_file.name} - PASSED")
+                else:
+                    print(f"❌ {test_file.name} - FAILED")
+                    print(f"Error: {result.stderr}")
+                    
+            except Exception as e:
+                print(f"❌ Error running {test_file.name}: {e}")
+                phase_results[test_file.name] = {
+                    'returncode': -1,
+                    'stdout': '',
+                    'stderr': str(e)
+                }
+        
+        results[phase] = {
+            'status': 'completed',
+            'tests': phase_results,
+            'total_tests': len(test_files)
+        }
+    
+    # Print summary
+    print(f"\n{'='*60}")
+    print("PHASE TEST SUMMARY")
+    print(f"{'='*60}")
+    
+    total_phases = len(phases)
+    completed_phases = sum(1 for phase in results.values() if phase['status'] == 'completed')
+    skipped_phases = sum(1 for phase in results.values() if phase['status'] == 'skipped')
+    
+    print(f"Total Phases: {total_phases}")
+    print(f"Completed: {completed_phases}")
+    print(f"Skipped: {skipped_phases}")
+    
+    for phase, result in results.items():
+        print(f"\n{phase.upper()}:")
+        if result['status'] == 'completed':
+            passed = sum(1 for test in result['tests'].values() if test['returncode'] == 0)
+            failed = sum(1 for test in result['tests'].values() if test['returncode'] != 0)
+            print(f"  Tests: {result['total_tests']} (✅ {passed}, ❌ {failed})")
+        else:
+            print(f"  Status: {result['status']} ({result.get('reason', 'unknown')})")
+    
+    return results
+
+if __name__ == '__main__':
+    run_phase_tests()