Investors

The security layer for agents that act.

AgentHacker sits inside the agent loop and stops attacks before the agent acts on them. This page is the short version of why we think that is a company.

Talk to the team Read the benchmark →

What we have shown

Measured against the tools buyers can already download.

Two public benchmarks, one fixed victim model, competitors on their published configurations. Our failures count against us.

0 of 642

attacks succeeded against an agent running AgentHacker.

0 of 62

benign episodes with an intervention, so the agent keeps doing its job.

Read the full benchmark report →

Why now

Agents moved from answering to acting.

01

Tool output is the new attack surface

An agent that reads a web page, an email or an API response will follow instructions hidden in it. Prompt injection is the first item on OWASP's Top 10 for LLM applications.

02

Buyers and acquirers are moving

Large security vendors have been acquiring AI security startups since 2024. Forrester says acquirers are paying for talent, early Fortune 500 traction, and a story they were missing.

03

Today's guardrails force a trade

In our benchmark the open-source options either let attacks through or broke normal work: LlamaFirewall let 29.0% of attacks succeed and intervened on 38.7% of benign episodes; NeMo Guardrails intervened on 54.8%.

Where we sit

Protecting an agent can also break it.

A guardrail can stop attacks and still block the agent's normal work. Buyers need both numbers, so we publish both. Here is how the main approaches compare.

Category

Prompt and model guardrails

What it does

Screen prompts and replies for attacks and policy violations.

What it leaves open

Typically built around the prompt and the reply, not the agent's tool calls.

Category

Discovery, posture and governance

What it does

Find the agents and AI tools in use, then set and monitor policy.

What it leaves open

Shows where the risk is. Stopping it inside the agent loop is a separate job.

Category

Lifecycle platforms

What it does

Discover agents, test them before launch, monitor them in production.

What it leaves open

Broad scope. Buyers still need proof that normal work is left alone.

Category

Sandboxes and action checks

What it does

Limit what an agent can touch, or verify an action against policy before it runs.

What it leaves open

Narrow by design, so usually paired with detection.

Category

AgentHacker

How

Four checkpoints inside the agent loop: inputs, tool data, tool calls, outputs.

Evidence

Tested head to head against two named competitors on public datasets. Figures are in the benchmark report.

Categories overlap and the descriptions are general. Benchmark figures are ours; see the report for method and limits.

Why it compounds

Why this gets harder to copy.

Wrapping a model is not a moat. These are the three things we are building the company on. The tags show what is live today and what is still to come.

SHIPPING

Position in the loop

We sit between the model and its tools, so we see the whole trajectory, not only a prompt. That is what lets us stop an attack at the tool call.

SHIPPING / NEXT

Data that compounds

Every verdict on a detection is training signal. Detectors trained on design partner signal are the next build. They do not exist yet.

PUBLISHED

Evidence as distribution

Public datasets, competitor configurations and our own misses are all in the report. A buyer can check the claim instead of taking our word.

Where we are

What is done, what is open, what comes next.

DONE
Benchmark published against two public datasets and two named competitors.
LIVE
Design partner beta: 4 spots open.
PARTNERS
Design partner beta open to teams running AI agents.
NEXT
Detectors trained on partner signal.
ROUND
Open to conversations with investors who back AI security.

Talk to us

Start with the benchmark, then ask us anything.

Write to Haider Shareef, CEO. hshareef@agenthacker.ai

Book 20 min with Haider