3LS research desk

Runtime AI governance intelligence

Incidents, control models, and operating notes for security teams governing prompts, files, memory, OAuth grants, and agent tool actions.

Read the playbooks Connect research to controls

Research board with transcript snippets, warning decision, endpoint boundary diagram, source citation card, and governance console tile

Mechanism:

On-device agent intercepts AI activity before it leaves the endpoint.

Privacy:

Detections and policy decisions stay within your organization.

Incidents

Provider leaks, exposed chats, agent failures, and operational lessons.

Controls

Policy, observability, and runtime decisions before data or authority leaves.

Tool surfaces

Browsers, MCP, extensions, coding agents, and delegated workflows.

Data boundaries

Where prompts, files, memory, and archives become enterprise exposure paths.

Featured Articles

Current briefs for AI security teams tracking data movement, tool authority, and runtime controls.

Printed compliance chart beside a laptop, representing APRA AI governance evidence and assurance

Compliance 10 min read

APRA's AI Letter Turns Governance Into an Operating Test

APRA's April 2026 letter makes AI governance practical: boards need AI literacy, executives need lifecycle accountability, and regulated entities need controls that prove what happened before AI changes data, resilience, suppliers, or authority.

Runtime AI governance intelligence

Incidents

Controls

Tool surfaces

Data boundaries

Featured Articles

APRA's AI Letter Turns Governance Into an Operating Test

Bring Your Agent to Teams Means Bring Your Controls Too

Workspace Features Create New Enterprise Data Paths

Workspace Agents Move ChatGPT Into the Operating Surface

Microsoft Native AI Controls Are a Starting Point, Not the Finish Line

Your Old Slack Archive Is Now AI Training Data

The Vercel Incident Shows AI OAuth Is Supply Chain Access

Anthropic Mythos Is a Provider-Trust Warning

AI Governance Is a Runtime Control System

AI Chat Security Risk and the Shadow Database Problem

Why AI Vendors Cannot Secure Your Enterprise Context

The Enterprise AI Visibility Gap Behind Shadow AI Risk

OpenClaw Malicious Skills Turn Agent Convenience Into a Supply Chain Problem

If CISA Can Put FOUO in ChatGPT, Your Exception Process Is the Breach

The Enterprise Agent Control Plane from Toggles to Policy as Code

Approval Fatigue Is an Enterprise Risk in Agent Sandboxes

Agent Sandboxes Are the Containment Boundary

Agentic Browser Prompt Injection and the Lethal Trifecta

MCP Tool Poisoning Turns Tool Metadata Into Data Exfiltration

Prompt Poaching in the Chrome Web Store and AI Chat Exfiltration

Claude.ai Email Exfiltration Shows How Assistants Can Leak Inboxes

M365 Copilot and Claude Create a Compliance Bomb

The AWS VSCode Supply Chain Near Miss That Almost Reached Millions

Clicking Yes to AI Disaster and the Approval Fatigue Crisis

The MCP Server That Wiped Production and the AI Tooling Risk

Prompt Injection Vulnerabilities and Lessons from the 2024 Wave

The $4.88M Question on AI PII Exposure

Shadow AI in the Enterprise and the Surge in Uncontrolled Usage

All Articles

Weekly AI Threat Intelligence Briefings