🔐 Security Monitor

Aggregate security telemetry across AI Copilots · LLM Gateway · Studio Agents · MCP Hub

Live

Overall Security Score

—

/ 100

Computing…

Waiting for telemetry data

Threat Rate

—

Block Rate

—

🚨

Threats Detected

—

in selected range

⚠️

High Risk Events

—

severity ≥ high

🛡️

Blocked Requests

—

auto-blocked by gateway

🤖

Agents Monitored

—

across all types

📜

Compliance Gaps

—

policy violations found

✅

Approved Connections

—

of total connections

Threat Type Breakdown

💉

Prompt Injection

—

Malicious instructions embedded in user input to override system behaviour.

—

📤

Data Exfiltration

—

Attempts to extract sensitive data through LLM responses or tool outputs.

—

🧠

IP / Confidential Leak

—

Proprietary or confidential data detected in model prompts or completions.

—

🔓

Jailbreak Attempts

—

Attempts to bypass safety guardrails or model alignment restrictions.

—

🌀

Hallucination Risk

—

High-confidence fabrications flagged by anomaly detection on token patterns.

—

📋

Compliance Violations

—

Requests or responses that breach defined policy or regulatory controls.

—

Event Trend & Source Breakdown

📊 Security Event Trend Last 14 days · stacked by threat type

🗂️ Events by Source All monitored agent types

🗂️

Loading sources…

Fetching event distribution across AI Copilots, LLM Gateway, Studio Agents and MCP Hub.

Risk Distribution & Threat Radar

🍩 Risk Level Distribution By threat category

🍩

No risk data

Risk distribution will populate from telemetry events.

🕸️ Threat Category Radar Exposure across all threat categories

Recent Security Events

🕒 Security Event Log

Time	Source	Agent Type	Threat	Severity	Action	User	Details
🔍 Loading security events… Fetching recent telemetry from the LLM Gateway and MCP Hub.

Agent Risk Registry

🤖 Monitored Agent Risk Profiles

Agent Name	Type	Risk Score	Risk Level	Top Risk Factor	Status	Last Reviewed
🤖 Loading agent registry… Agent risk profiles are compiled from connections, telemetry and governance metadata.