Agent skill
error-sanitization
Production-safe error handling that logs full details server-side while exposing only generic, safe messages to users. Prevents information leakage of database strings, file paths, stack traces, and API keys.
Install this agent skill to your Project
npx add-skill https://github.com/dadbodgeoff/drift/tree/main/drift v1 depreciated/skills/error-sanitization
Metadata
Additional technical details for this skill
- time
- 2h
- source
- drift-masterguide
- category
- security
SKILL.md
Error Sanitization
Production-safe error handling: log everything server-side, expose nothing sensitive to users.
When to Use This Skill
- Building APIs that return error messages to clients
- Handling exceptions in production environments
- Processing batch operations with partial failures
- Any system where error messages could leak sensitive information
Core Concepts
Error messages can leak sensitive information including database connection strings, internal file paths, stack traces, API keys, and business logic details. The solution is to always log full error details server-side for debugging while returning only generic, safe messages to users.
The flow is:
- Exception occurs
- Log FULL error server-side with context
- Classify error type
- Return GENERIC message to user
- Only expose safe, actionable errors (like validation)
Implementation
Python
import os
import logging
from typing import Optional
from fastapi import HTTPException
from pydantic import ValidationError
logger = logging.getLogger(__name__)
class ErrorSanitizer:
"""
Sanitizes error messages to prevent information leakage.
"""
SENSITIVE_PATTERNS = [
"password", "secret", "key", "token", "credential",
"postgresql://", "mysql://", "mongodb://", "redis://",
"localhost", "127.0.0.1", "internal", "0.0.0.0",
"traceback", "exception", "error at", "line ",
"/home/", "/var/", "/etc/", "C:\\",
"SUPABASE", "AWS", "STRIPE", "SENDGRID",
]
@staticmethod
def is_production() -> bool:
return os.getenv("ENVIRONMENT", "development").lower() == "production"
@staticmethod
def sanitize_error(
e: Exception,
user_message: str = "Operation failed",
log_context: Optional[dict] = None
) -> str:
"""Sanitize error message for user display."""
# ALWAYS log full error server-side
logger.error(
f"Error occurred: {type(e).__name__}: {str(e)}",
exc_info=True,
extra=log_context or {}
)
# In development, show more details
if not ErrorSanitizer.is_production():
return f"{user_message}: {str(e)}"
# Validation errors are safe (user input issues)
if isinstance(e, ValidationError):
return f"Validation error: {str(e)}"
# HTTPException with client error (4xx) is safe
if isinstance(e, HTTPException):
if 400 <= e.status_code < 500:
return e.detail
return user_message
# Check for sensitive patterns
error_str = str(e).lower()
for pattern in ErrorSanitizer.SENSITIVE_PATTERNS:
if pattern in error_str:
return user_message
# Short, simple errors without sensitive patterns might be safe
if len(str(e)) < 100 and not any(c in str(e) for c in ['/', '\\', '@', ':']):
return str(e)
return user_message
@staticmethod
def create_http_exception(
e: Exception,
status_code: int = 500,
user_message: str = "Operation failed",
log_context: Optional[dict] = None
) -> HTTPException:
"""Create HTTPException with sanitized error message."""
safe_message = ErrorSanitizer.sanitize_error(e, user_message, log_context)
return HTTPException(status_code=status_code, detail=safe_message)
TypeScript
import { Logger } from './logger';
const SENSITIVE_PATTERNS = [
'password', 'secret', 'key', 'token', 'credential',
'postgresql://', 'mysql://', 'mongodb://', 'redis://',
'localhost', '127.0.0.1', 'internal', '0.0.0.0',
'/home/', '/var/', '/etc/', 'C:\\',
];
interface SanitizeOptions {
userMessage?: string;
logContext?: Record<string, unknown>;
}
export class ErrorSanitizer {
private static isProduction(): boolean {
return process.env.NODE_ENV === 'production';
}
static sanitize(
error: Error,
options: SanitizeOptions = {}
): string {
const { userMessage = 'Operation failed', logContext = {} } = options;
// Always log full error server-side
Logger.error('Error occurred', {
name: error.name,
message: error.message,
stack: error.stack,
...logContext,
});
// In development, show more details
if (!this.isProduction()) {
return `${userMessage}: ${error.message}`;
}
// Check for sensitive patterns
const errorStr = error.message.toLowerCase();
for (const pattern of SENSITIVE_PATTERNS) {
if (errorStr.includes(pattern)) {
return userMessage;
}
}
// Short, simple errors might be safe
if (error.message.length < 100 && !/[\/\\@:]/.test(error.message)) {
return error.message;
}
return userMessage;
}
static createHttpError(
error: Error,
statusCode: number = 500,
options: SanitizeOptions = {}
): { statusCode: number; message: string } {
return {
statusCode,
message: this.sanitize(error, options),
};
}
}
Usage Examples
Route Handler
@router.post("/process")
async def process_invoice(invoice_id: str):
try:
result = await processor.process(invoice_id)
return result
except ValidationError as e:
# Validation errors are safe to expose
raise HTTPException(status_code=400, detail=str(e))
except HTTPException:
raise # Re-raise as-is
except Exception as e:
# All other errors get sanitized
raise ErrorSanitizer.create_http_exception(
e,
status_code=500,
user_message="Failed to process invoice",
log_context={"invoice_id": invoice_id}
)
Domain-Specific Sanitizers
def sanitize_database_error(e: Exception) -> str:
return ErrorSanitizer.sanitize_error(
e,
user_message="Database operation failed. Please try again.",
log_context={"error_type": "database"}
)
def sanitize_api_error(e: Exception, service_name: str = "external service") -> str:
return ErrorSanitizer.sanitize_error(
e,
user_message=f"Failed to communicate with {service_name}.",
log_context={"error_type": "external_api", "service": service_name}
)
Batch Operations with Partial Failures
def process_items(items: list[dict]) -> dict:
failed_items = []
for idx, item in enumerate(items):
try:
process_item(item)
except Exception as e:
error_type = classify_error(e)
failed_items.append({
"line": idx,
"description": item['description'][:50], # Truncate
"error_type": error_type,
"message": get_user_friendly_message(error_type)
# NOTE: Don't include str(e) - might be sensitive
})
return {
"status": "partial_success" if failed_items else "success",
"failed_items": failed_items
}
Best Practices
- Always log full error details server-side with
exc_info=True - Include correlation IDs (request_id, session_id) in logs for tracing
- Truncate user input in error messages to prevent log injection
- Test error handling in production mode - errors look different in dev vs prod
- Monitor "unknown" error rates - high rates indicate missing classification
Common Mistakes
- Exposing raw exception messages to users in production
- Forgetting to log the full error before sanitizing
- Including sensitive data in log messages (passwords, connection strings)
- Not testing error responses in production mode
- Trusting that "safe" errors don't contain injected content
Related Patterns
- exception-taxonomy - Hierarchical exception system with error codes
- circuit-breaker - Prevent cascading failures
- error-handling - General error handling patterns
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
worker-orchestration
Manage concurrent background workers with scheduling, dependencies, health monitoring, and automatic disabling of failing workers.
leader-election
Elect a single leader among multiple instances. Only one instance runs cron jobs or processes queues. Automatic failover when leader dies.
fuzzy-matching
Multi-stage fuzzy matching pipeline for entity reconciliation. PostgreSQL trigram pre-filter, salient overlap check, and multi-factor similarity scoring.
stripe-integration
Complete Stripe payments integration with subscriptions, webhooks, and customer portal. Use when adding billing to a SaaS application with subscription tiers, usage-based pricing, or one-time payments.
prompt-engine
Template-based AI prompt engine with YAML templates, brand kit injection, input sanitization for security, and token-efficient context blocks.
scoring-engine
Statistical scoring with z-scores, percentiles, freshness decay, and cross-category normalization. Rank and compare items with confidence scoring.
Didn't find tool you were looking for?