Spaces:
Running
Running
File size: 16,128 Bytes
7b7b417 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 | """Reflexive Core Engine - Main runtime for self-monitoring and corrective actions."""
import asyncio
import json
import hashlib
from datetime import datetime
from enum import Enum
from typing import Any, Dict, List, Optional, Callable, Union
from uuid import UUID, uuid4
from pydantic import BaseModel, Field
from fastmcp.utilities.logging import get_logger
logger = get_logger(__name__)
class DecisionType(str, Enum):
"""Types of reflexive decisions."""
HALT = "halt"
ESCALATE = "escalate"
MONITOR = "monitor"
ALLOW = "allow"
class RiskLevel(str, Enum):
"""Risk levels for reflexive decisions."""
LOW = "low"
MEDIUM = "medium"
HIGH = "high"
CRITICAL = "critical"
class ActionContext(BaseModel):
"""Context for an action being evaluated by the reflexive core."""
action_id: str = Field(..., description="Unique identifier for the action")
actor_id: str = Field(..., description="ID of the entity performing the action")
action_type: str = Field(..., description="Type of action being performed")
resource_id: Optional[str] = Field(default=None, description="ID of the resource being accessed")
metadata: Dict[str, Any] = Field(default_factory=dict, description="Additional action metadata")
timestamp: datetime = Field(default_factory=datetime.utcnow, description="When the action occurred")
session_id: Optional[str] = Field(default=None, description="Session identifier")
request_id: Optional[str] = Field(default=None, description="Request identifier")
def get_context_hash(self) -> str:
"""Get SHA-256 hash of action context for integrity verification."""
content = {
"action_id": self.action_id,
"actor_id": self.actor_id,
"action_type": self.action_type,
"resource_id": self.resource_id,
"metadata": self.metadata,
"session_id": self.session_id,
"request_id": self.request_id
}
content_str = json.dumps(content, sort_keys=True, default=str)
return hashlib.sha256(content_str.encode()).hexdigest()
class ReflexiveDecision(BaseModel):
"""A decision made by the reflexive core."""
decision_id: UUID = Field(default_factory=uuid4, description="Unique decision identifier")
decision_type: DecisionType = Field(..., description="Type of decision made")
risk_level: RiskLevel = Field(..., description="Risk level of the situation")
action_context: ActionContext = Field(..., description="Context of the action being evaluated")
reason: str = Field(..., description="Reason for the decision")
evidence: Dict[str, Any] = Field(default_factory=dict, description="Evidence supporting the decision")
timestamp: datetime = Field(default_factory=datetime.utcnow, description="When the decision was made")
escalated_to: Optional[str] = Field(default=None, description="Role/entity escalated to")
proof_hash: Optional[str] = Field(default=None, description="Hash of decision proof")
model_config = {"use_enum_values": True}
def get_decision_hash(self) -> str:
"""Get SHA-256 hash of decision for integrity verification."""
content = {
"decision_id": str(self.decision_id),
"decision_type": self.decision_type,
"risk_level": self.risk_level,
"action_context": self.action_context.model_dump(),
"reason": self.reason,
"evidence": self.evidence,
"escalated_to": self.escalated_to
}
content_str = json.dumps(content, sort_keys=True, default=str)
return hashlib.sha256(content_str.encode()).hexdigest()
class ReflexiveEngine:
"""Main reflexive core engine for self-monitoring and corrective actions."""
def __init__(self, policy_engine=None, ledger=None):
"""Initialize the reflexive engine.
Args:
policy_engine: Policy engine instance for policy monitoring
ledger: Provenance ledger instance for audit logging
"""
self.policy_engine = policy_engine
self.ledger = ledger
self.monitors: List[Callable] = []
self.decision_handlers: Dict[DecisionType, Callable] = {}
self.is_running = False
self.event_queue = asyncio.Queue()
# Register default decision handlers
self._register_default_handlers()
logger.info("Reflexive engine initialized")
def _register_default_handlers(self):
"""Register default decision handlers."""
self.decision_handlers[DecisionType.HALT] = self._handle_halt
self.decision_handlers[DecisionType.ESCALATE] = self._handle_escalate
self.decision_handlers[DecisionType.MONITOR] = self._handle_monitor
self.decision_handlers[DecisionType.ALLOW] = self._handle_allow
async def start(self):
"""Start the reflexive engine."""
if self.is_running:
logger.warning("Reflexive engine is already running")
return
self.is_running = True
logger.info("Reflexive engine started")
# Start the main event processing loop
asyncio.create_task(self._process_events())
async def stop(self):
"""Stop the reflexive engine."""
self.is_running = False
logger.info("Reflexive engine stopped")
async def _process_events(self):
"""Main event processing loop."""
while self.is_running:
try:
# Wait for events with timeout
event = await asyncio.wait_for(self.event_queue.get(), timeout=1.0)
await self._handle_event(event)
except asyncio.TimeoutError:
# No events, continue
continue
except Exception as e:
logger.error(f"Error processing reflexive event: {e}")
async def _handle_event(self, event: Dict[str, Any]):
"""Handle a reflexive event."""
try:
# Create action context from event
action_context = ActionContext(**event.get("action_context", {}))
# Evaluate the action
decision = await self._evaluate_action(action_context)
# Execute the decision
await self._execute_decision(decision)
# Log the decision to audit trail
await self._log_decision(decision)
except Exception as e:
logger.error(f"Error handling reflexive event: {e}")
async def _evaluate_action(self, action_context: ActionContext) -> ReflexiveDecision:
"""Evaluate an action and make a reflexive decision."""
# Run all monitors
violations = []
anomalies = []
for monitor in self.monitors:
try:
# Check if monitor is async
if asyncio.iscoroutinefunction(monitor):
result = await monitor(action_context)
else:
result = monitor(action_context)
if result:
if result.get("type") == "violation":
violations.append(result)
elif result.get("type") == "anomaly":
anomalies.append(result)
except Exception as e:
logger.error(f"Monitor error: {e}")
# Make decision based on findings
if violations or anomalies:
# Determine risk level
risk_level = self._assess_risk_level(violations, anomalies)
# Make decision based on risk level
if risk_level == RiskLevel.CRITICAL:
decision_type = DecisionType.HALT
reason = f"Critical risk detected: {len(violations)} violations, {len(anomalies)} anomalies"
elif risk_level == RiskLevel.HIGH:
decision_type = DecisionType.HALT
reason = f"High risk detected: {len(violations)} violations, {len(anomalies)} anomalies"
elif risk_level == RiskLevel.MEDIUM:
decision_type = DecisionType.ESCALATE
reason = f"Medium risk detected: {len(violations)} violations, {len(anomalies)} anomalies"
else:
decision_type = DecisionType.MONITOR
reason = f"Low risk detected: {len(violations)} violations, {len(anomalies)} anomalies"
else:
decision_type = DecisionType.ALLOW
reason = "No violations or anomalies detected"
risk_level = RiskLevel.LOW
# Create decision
decision = ReflexiveDecision(
decision_type=decision_type,
risk_level=risk_level,
action_context=action_context,
reason=reason,
evidence={
"violations": violations,
"anomalies": anomalies
}
)
# Set proof hash
decision.proof_hash = decision.get_decision_hash()
return decision
def _assess_risk_level(self, violations: List[Dict], anomalies: List[Dict]) -> RiskLevel:
"""Assess the overall risk level based on violations and anomalies."""
total_issues = len(violations) + len(anomalies)
# Check for critical violations or anomalies
critical_violations = [v for v in violations if v.get("severity") == "critical"]
critical_anomalies = [a for a in anomalies if a.get("severity") == "critical"]
if critical_violations or critical_anomalies:
return RiskLevel.CRITICAL
# Check for high severity issues
high_violations = [v for v in violations if v.get("severity") == "high"]
high_anomalies = [a for a in anomalies if a.get("severity") == "high"]
if high_violations or high_anomalies or total_issues >= 5:
return RiskLevel.HIGH
# Check for medium severity issues
medium_violations = [v for v in violations if v.get("severity") == "medium"]
medium_anomalies = [a for a in anomalies if a.get("severity") == "medium"]
if medium_violations or medium_anomalies or total_issues >= 2:
return RiskLevel.MEDIUM
return RiskLevel.LOW
async def _execute_decision(self, decision: ReflexiveDecision):
"""Execute a reflexive decision."""
handler = self.decision_handlers.get(decision.decision_type)
if handler:
try:
await handler(decision)
except Exception as e:
logger.error(f"Error executing decision {decision.decision_type}: {e}")
else:
logger.warning(f"No handler for decision type: {decision.decision_type}")
async def _handle_halt(self, decision: ReflexiveDecision):
"""Handle a halt decision."""
logger.critical(f"HALTING ACTION: {decision.action_context.action_id} - {decision.reason}")
# In a real implementation, this would stop the action execution
# For now, we just log the halt decision
async def _handle_escalate(self, decision: ReflexiveDecision):
"""Handle an escalate decision."""
# Determine escalation target
escalation_target = self._determine_escalation_target(decision)
decision.escalated_to = escalation_target
logger.warning(f"ESCALATING TO {escalation_target}: {decision.action_context.action_id} - {decision.reason}")
# In a real implementation, this would notify the escalation target
async def _handle_monitor(self, decision: ReflexiveDecision):
"""Handle a monitor decision."""
logger.info(f"MONITORING ACTION: {decision.action_context.action_id} - {decision.reason}")
# In a real implementation, this would increase monitoring for this action
async def _handle_allow(self, decision: ReflexiveDecision):
"""Handle an allow decision."""
logger.debug(f"ALLOWING ACTION: {decision.action_context.action_id} - {decision.reason}")
# Action is allowed to proceed
def _determine_escalation_target(self, decision: ReflexiveDecision) -> str:
"""Determine the appropriate escalation target based on the decision."""
if decision.risk_level == RiskLevel.CRITICAL:
return "security_admin"
elif decision.risk_level == RiskLevel.HIGH:
return "system_admin"
else:
return "monitoring_team"
async def _log_decision(self, decision: ReflexiveDecision):
"""Log the reflexive decision to the audit trail."""
if self.ledger:
try:
from fastmcp.ledger import LedgerEvent, EventType
event = LedgerEvent(
event_type=EventType.REFLEXIVE_DECISION,
actor_id="reflexive_core",
resource_id=decision.action_context.action_id,
action=f"reflexive_{decision.decision_type}",
metadata={
"decision_id": str(decision.decision_id),
"risk_level": decision.risk_level,
"reason": decision.reason,
"proof_hash": decision.proof_hash,
"escalated_to": decision.escalated_to
}
)
self.ledger.append_event(event)
logger.debug(f"Logged reflexive decision {decision.decision_id} to audit trail")
except Exception as e:
logger.error(f"Failed to log reflexive decision: {e}")
else:
# Log to standard logger if no ledger is available
logger.info(f"Reflexive decision: {decision.decision_type} - {decision.reason} (Decision ID: {decision.decision_id})")
def add_monitor(self, monitor: Callable):
"""Add a monitor function to the reflexive engine."""
self.monitors.append(monitor)
logger.info(f"Added monitor: {monitor.__name__}")
def remove_monitor(self, monitor: Callable):
"""Remove a monitor function from the reflexive engine."""
if monitor in self.monitors:
self.monitors.remove(monitor)
logger.info(f"Removed monitor: {monitor.__name__}")
async def submit_action(self, action_context: ActionContext):
"""Submit an action for reflexive evaluation."""
event = {
"action_context": action_context.model_dump(),
"timestamp": datetime.utcnow().isoformat()
}
await self.event_queue.put(event)
logger.debug(f"Submitted action {action_context.action_id} for reflexive evaluation")
async def simulate_risk(self, risk_scenario: Dict[str, Any]) -> ReflexiveDecision:
"""Simulate a risk scenario and return the reflexive decision."""
# Create action context from scenario
action_context = ActionContext(**risk_scenario.get("action_context", {}))
# Override monitors temporarily for simulation
original_monitors = self.monitors.copy()
# Add simulation monitors
simulation_monitors = risk_scenario.get("monitors", [])
for monitor_func in simulation_monitors:
self.monitors.append(monitor_func)
try:
# Evaluate the action
decision = await self._evaluate_action(action_context)
return decision
finally:
# Restore original monitors
self.monitors = original_monitors
def get_engine_status(self) -> Dict[str, Any]:
"""Get the current status of the reflexive engine."""
return {
"is_running": self.is_running,
"monitor_count": len(self.monitors),
"queue_size": self.event_queue.qsize(),
"decision_handlers": list(self.decision_handlers.keys())
}
|