A comprehensive security and monitoring solution for AI agents with OpenTelemetry-compliant telemetry. Cylestio Monitor provides lightweight, drop-in security monitoring for LLM clients and frameworks with just two lines of code.
Cylestio Monitor is a Python SDK that provides security and monitoring capabilities for AI agents with OpenTelemetry-compliant telemetry. While it works as a standalone solution, it integrates seamlessly with the Cylestio UI and smart dashboards for enhanced visibility and security monitoring across your entire agentic workforce.
pip install cylestio-monitor
import cylestio_monitor
from anthropic import Anthropic
# Start monitoring with minimal configuration
cylestio_monitor.start_monitoring(agent_id="my-agent")
# Create your LLM client - it will be automatically patched
client = Anthropic()
# Use your client as normal - all calls are automatically monitored
response = client.messages.create(
model="claude-3-haiku-20240307",
max_tokens=1000,
messages=[{"role": "user", "content": "Hello, Claude!"}]
)
# When finished, stop monitoring
cylestio_monitor.stop_monitoring()
# All events will use ISO8601 format with UTC timezone and Z suffix
# e.g., "2024-03-27T15:31:40.622Z"
Cylestio Monitor works by automatically patching supported LLM clients and frameworks. No additional configuration is required - the SDK detects which libraries are available and applies the appropriate monitoring.
The start_monitoring function accepts these configuration options:
cylestio_monitor.start_monitoring(
agent_id="my-agent", # Required: unique identifier for your agent
config={ # Optional: configuration dictionary
"debug_level": "INFO", # Logging level (DEBUG, INFO, WARNING, ERROR)
"log_file": "output/my_logs.json", # Path for local JSON logs
"telemetry_endpoint": "http://custom.telemetry.server:9000", # Custom telemetry host/port
"development_mode": False # Enable additional development features
}
)
By default, Cylestio Monitor logs events to a file when the log_file option is provided:
cylestio_monitor.start_monitoring(
agent_id="weather-agent",
config={
"log_file": "output/weather_monitoring.json"
}
)
All events follow OpenTelemetry standards with trace context:
{
"timestamp": "2024-03-27T15:31:40.622Z",
"trace_id": "2a8ec755032d4e2ab0db888ab84ef595",
"span_id": "96d8c2be667e4c78",
"parent_span_id": "f1490a668d69d1dc",
"name": "llm.call.start",
"attributes": {
"method": "messages.create",
"model": "claude-3-haiku-20240307"
},
"agent_id": "weather-agent"
}
Cylestio Monitor maintains security through a comprehensive security pipeline that includes:
This security pipeline is designed to help organizations meet regulatory and compliance requirements including SOC2, GDPR, HIPAA, and industry best practices for machine learning systems. The pipeline is continuously monitored and updated to address emerging security threats specific to AI and LLM systems.
For more details on our security approach and best practices, see our security documentation.
This repository uses pre-commit hooks to ensure code quality and security:
# Install pre-commit
pip install pre-commit
# Set up the git hooks
pre-commit install
After installation, the hooks will run automatically on git commit.
Cylestio Monitor enables organizations to maintain regulatory compliance as they transition to agentic workforces. The system:
The monitoring system itself is designed with compliance in mind, ensuring all sensitive data is properly masked in logs and events. Security patterns and detection rules are fully configurable to match your organization's specific compliance requirements.
For practical implementations of Cylestio Monitor in various agent architectures, check out the Cylestio Use Cases repository, which includes working examples such as:
Each example demonstrates how to integrate Cylestio monitoring with just a few lines of code:
import cylestio_monitor
cylestio_monitor.start_monitoring(
agent_id="agent-name",
config={
"log_file": "output/monitoring.json",
"debug_level": "DEBUG"
}
)
# Your agent code here...
cylestio_monitor.stop_monitoring()
Latest release: v0.1.12 (May 13, 2025)
Highlights:
Previous releases:
See CHANGELOG.md for the complete version history.
We welcome contributions! Please see our Contributing Guide for details on how to get started.
Apache License 2.0
This project is licensed under the Apache License, Version 2.0. See the LICENSE file for details.
FAQs
A monitoring tool for LLM API calls
We found that cylestio-monitor demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 0 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Research
Security News
A look at the top trends in how threat actors are weaponizing open source packages to deliver malware and persist across the software supply chain.
Security News
ESLint now supports HTML linting with 48 new rules, expanding its language plugin system to cover more of the modern web development stack.
Security News
CISA is discontinuing official RSS support for KEV and cybersecurity alerts, shifting updates to email and social media, disrupting automation workflows.