Boomi Agentstudio
Agent builder whose Agent Control Tower registers and monitors Boomi agents and agents on Amazon Bedrock, Microsoft Copilot and other platforms, keeps audit logs and can switch agents off where the provider allows it.
Data Protection
AI application security
Prompt injection, data leaks through AI tools and unapproved AI use get caught.
Addresses 361 stories in our coverage in the last 90 days 5 products that provide it
What it is
Controls for how staff use AI tools and how your own AI applications behave: discovery of AI use, filtering of prompts and responses, protection against prompt injection, and testing of models and agents.
How to value it
AI tools have become a new channel for data leaks and a new attack surface: agents that can be tricked into sending data out, and integrations with broad access. Visibility and guardrails let the business use AI without each new deployment becoming a new risk.
Questions to ask a vendor
- 1Can it discover which AI tools staff already use?
- 2Does it inspect prompts and responses for sensitive data and injection attempts?
- 3Does it cover AI agents that can use tools, not only chat?
- 4How does it help with EU AI Act obligations?
For your risk register
- Risk
- Sensitive data leaks through AI tools, or AI applications are manipulated.
- Control
- AI use policy with discovery, prompt filtering and testing of AI applications.
- Owner
- CISO together with the data protection lead.
- Measure
- Share of AI tools in use that are approved.
A starting point: set the owner and measure to fit your organisation.
Products that provide it
Listed alphabetically by vendor. A listing is not a recommendation: use the questions above to compare them, or tick up to 3 to see them side by side.
Boomi AI Gateway
Gateway for AI model and MCP tool traffic, built on Lunar.dev technology. It limits which tools each agent can call. Its enterprise version adds an approval step for new MCP servers and logs each call with the user or agent behind it.
Cisco AI Defense
Discovery of AI use, testing of models and guardrails for AI applications.
HiddenLayer AI Security Platform
Model scanning, AI red teaming and runtime protection for AI applications.
Prisma AIRS
Security for AI applications, models and agents, including runtime protection against prompt injection.
Recent stories it addresses
740 stories in total- 13 Nov 2026How one AI agent exposed five critical enterprise security gaps
AI agent processing invoices exposed five critical enterprise security gaps.
- 13 Nov 2026When an AI Agent Goes Wrong, Can You Stop It?
OWASP GenAI Security Project warns AI agents can exceed intended scope; proper containment requires pausing workflows,
- 12 Nov 2026Trained to betray: How post-training supply chain attacks hijack coding agents
Post-training supply chain attacks can hijack AI coding agents via poisoned datasets or model checkpoints.
- 12 Nov 2026Attackers Use Autonomous AI to Target Semiconductor, Defense and Civil Society Sectors
Chinese-speaking attackers weaponized Cairn AI framework to autonomously breach semiconductor, defense, and civil
- 12 Nov 2026GitHub Agentic Autofix and the question of who owns a bad AI patch
GitHub discusses accountability for AI-generated code patches and the implications for security.
- 11 Nov 2026Hackers Could Abuse Microsoft Entra AI Agents to Hijack Enterprise Accounts Without Stealing Passwords
Hackers can hijack enterprise accounts by tricking users into approving fake AI agents.
- 11 Nov 2026Claude Code AI Discovers Technique to Turn Dormant Model Neurons Into Hidden Malware Triggers
Microsoft researchers discover Lazarus, a technique to embed hidden malware triggers in AI models.
- 11 Nov 2026Datatilsynet (Denmark) - 2026-423-0075
Denmark's DPA reprimands municipality for inadequate AI monitoring risk assessment in nursing home.
How we list vendors
- A product is listed because it provides the control. Nobody pays to be listed, and on story pages the examples rotate, so no vendor is always first.
- Vendor pages show both the vendor's own research and the incidents involving its products, from our own coverage. We do not leave either out.
- Our summaries are written by us, based on the vendors' own pages. Text in quotation marks or marked "in their own words" is the vendor's own, quoted from its website.
- Logos, product names and quoted text belong to the vendors and are shown to identify them.
- Something wrong? Use "Suggest a correction" on the vendor's page. A product we should look at? Email support@threatnoir.com.