Coding Agents Overview

Code Guardian is a lightweight endpoint security component that connects AI coding assistants to AppViewX Agent Identity Security, providing automatic discovery, identity binding, centralized visibility, and real-time policy enforcement.

What is Code Guardian?

AppViewX Agent Identity Security uses Code Guardian to monitor AI coding tools on developer workstations, such as Claude Code, Cursor, and GitHub Copilot CLI. It runs on the developer's machine and handles three things: discovering which coding agents are installed, collecting available agent and endpoint context, and enforcing supported runtime policies where configured.
Key capabilities:
  • Automatic discovery: Detects coding assistant sessions as soon as they start on an endpoint and registers them in the central Agent Identity Security Agent Registry / AIBOM.
  • Identity and context attribution: Associates coding agent activity with the local user account and endpoint device.
  • Runtime policy enforcement: Evaluates tool execution, file access, and terminal commands in real time against configured enterprise policies.
  • Comprehensive audit logging: Streams session events and execution telemetry to Agent Identity Security Event Logs for compliance and security reviews.

How Code Guardian works

Code Guardian operates locally alongside your coding tools to inspect activity, check runtime policies, and report telemetry to the Agent Identity Security platform.
Table 1. The following sequence explains how Code Guardian integrates into developer workflows:
Component Action
Endpoint machine A supported coding agent is installed and running on the endpoint.
Code Guardian hooks Code Guardian intercepts tool execution, collects session metadata, and evaluates local security policies.
Agent Identity Security Platform synchronization Code Guardian securely transmits agent identity, session state, and audit telemetry to the Agent Identity Security platform over outbound HTTPS.
Governance and registry The Agent Identity Security platform updates the Agent Registry / AIBOM, recalculates risk posture, and logs events for continuous governance and configured Agent Policies are applied where supported.

Your deployment options

Choose the deployment mode manual installation for individual machines or Mobile Device Management (MDM) deployment for fleet-wide rollouts that matches your scenario.

Table 2. Select the deployment method that fits your operating model:
Deployment mode Authentication method Recommended use
Manual Installation Agent Identity Security API key Individual machines, pilots, controlled onboarding, and validation.
MDM Deployment Bootstrap token Silent, automated deployment across enterprise fleets using Microsoft Intune (Windows) or Jamf Pro (macOS). Enterprise-scale deployment through supported device management platforms.

Supported coding agents

Code Guardian currently supports the following coding agents: Claude Code, Cursor, and GitHub Copilot. Check the current supported-agent matrix in the Agent Identity Security Platform for the latest list as support evolves.

What becomes available after discovery

  • Agent Registry / AIBOM: Centralized inventory and relationship context for each discovered agent.
  • Agent Posture & Risk: Evaluates supported security, configuration, identity, and access conditions.
  • Event Logs: Provides activity and security context for investigation and audit.
  • Workflows: Supports governance actions in response to agent, identity, posture, or lifecycle events.
  • Agent Policies: Applies supported runtime controls to agent actions.