RAID, our Red Team AI Division
Attacker.
Intelligence.
RAID is the team that builds the agentic attack capabilities inside every Constant Cyber Attack™.
The result is a Swarm that thinks and strikes like your attacker: human-led and augmented by agentic tools. Included in every CovertSwarm subscription at no extra cost.
Attackers got faster.
Did you?
AI has closed the gap between attacker intent and how fast they can get it done.
More tools won’t answer that. You need to face a realistic attacker, armed the way real ones are. We’re the attacker you’re preparing for.
Power to your Swarm.
RAID’s attack agents work alongside the Swarm, each with its own job, taking on reconnaissance, enumeration and attack surface mapping at speed.
Our operators bring the intent: the lateral thinking and judgement that decide where the attack goes next. A threat that moves like the real one, aimed at what matters most to you.
Built by the Swarm.
RAID builds each agent for realistic attacker operations, designed around how the Swarm plans and runs its attacks.
You stay in control.
Agents work inside the limits you set, and operators approve each sensitive action before it runs.
Evidence for your board.
Show your board how you hold up against AI-enabled attacks. Our operators validate each finding before it reaches your Portal.
Agentic attack capabilities.
RAID gives your Swarm more intelligence to work with and more ways in. Our agents connect discoveries across your attack surface, giving our operators new paths to explore, weaknesses to exploit and defences to challenge.
Attack surface mapping.
Connects your assets, technologies and external relationships, so we know where to look next.
Research.
Digs into your technologies, public disclosures and potential weaknesses, then chases the leads that follow.
Web.
Investigates your applications in depth, with each finding reviewed for validity and business impact.
Network.
Maps your exposure, relationships and attack paths, with our operators directing each next move.
Social.
Researches context and prepares credible scenarios, and our operators choose which ones to run.
Operator AI.
Our operators use AI to analyse datasets, reverse engineer patches and adapt proof-of-concept code.
We know that real attack surfaces don’t come in neat categories. If yours doesn’t fit above, tell us what you’re working with.
The RAID Files
The latest research, findings and opinion from the RAID team.
Questions about RAID and AI red teaming.
It means two different things. One is attacking AI systems, for example probing large language models for prompt injection, data leakage or guardrail bypasses. The other is using AI in real attacks. RAID is built for the second: real attacks led by people and augmented by agentic tools.
No. RAID makes our people faster. It doesn’t replace them. Agents take on the groundwork: reconnaissance, enumeration and attack surface mapping. Our operators choose the target, chain the route and decide what happens next. Every engagement is human-led and augmented by agentic tools.
RAID isn’t a separate product or an add-on. It’s part of how we deliver Constant Cyber Attack, so there’s no extra cost and nothing extra to request or configure. For the background, read Luke Potter’s announcement of RAID.
In the CovertSwarm Portal. RAID findings are handled like any other discovery: prioritised, with context and remediation guidance, and visible as they’re validated.
We plan every action around your live operations. You set the rules of engagement before we attack, and limits apply before an action runs, not after.
RAID is model agnostic. We choose the model that suits each task. We’ll walk you through how your data is handled on your Attack Discovery Call.
Looking for something specific?
Talk to us. We’d love to hear from you.