FI EN Get in touch

Product

“Your AI agent is running. Do you know what it’s doing?”

One platform. Every agent behavior.

AgentPentest is a self-serve software platform for teams shipping autonomous agents. Run standardized evaluations, harden the model, and export the evidence and training data, all from one workspace.

Get started See pricing

AgentPentest is owned and operated by RNDLabs Oy.

Modules

Composable by default

Each module is a standardized part of the platform. Enable what you need. There is no custom scoping call.

Module 01

Behavior Testing

Connect an agent endpoint and run a standardized adversarial battery. The platform returns a signed behavior report with replayable evidence.

  • One-click runs against a fixed threat matrix
  • Signed report plus replayable evidence
  • Escalation, credential leakage, collusion, jailbreaks, scope
  • Standard benchmarks: AgentDojo, HarmBench, InjecAgent
  • EU AI Act Annex IV technical documentation
Explore

Module 02

Alignment & Hardening

Turn findings into fixed behavior with a standardized alignment pipeline that runs on your model or a managed one.

  • Reduce escalation, leakage, collusion and scope risk
  • Standardized fine-tuning and alignment pipeline
  • Reliability and refusal-consistency tuning
  • Before/after evidence on the same threat matrix
  • Documented, repeatable compliance posture
Explore

Module 03

Training Data

Export execution-verified, multi-agent trajectories straight from the platform: standardized, scored and training-ready.

  • Execution-verified, not scraped or synthetic
  • CoT-rich, multi-agent interaction traces
  • Standard, versioned data packs, not one-off builds
  • Training-ready exports for RL and fine-tuning pipelines
  • Multi-layer scoring and full provenance on every sample
Explore

Workflow

One pipeline, from risk to compliance

Every workspace runs the same four-step pipeline. Modules can be used on their own or end to end.

Scan

Point the platform at an agent endpoint and launch a standardized threat matrix.

Report

Get a signed behavior report with replayable evidence and prioritized findings.

Harden

Apply the alignment pipeline to fix the failures the scan surfaced.

Certify

Export EU AI Act Annex IV documentation and a signed misuse-risk verdict.

Output

Evidence you can use

Every run produces artifacts a security team, a procurement board or an auditor can actually work with.

Signed behavior report

A scored, cryptographically signed report across six behavioral dimensions.

Secure logs & evidence

The raw, replayable run records behind every finding: tamper-evident and auditable.

Actionable findings

A prioritised remediation list, not just a pass/fail score.

Hardened model

The standardized alignment pipeline applied to the failures the scan found.

EU AI Act Annex IV docs

Technical documentation structured for Annex IV review, generated from real evidence.

Risk certification

A signed jailbreak and misuse risk verdict for procurement and sign-off.

Pricing

Evaluate first. Scale when you ship.

Transparent, self-serve pricing: evaluate first, scale when you ship.

Enterprise

from €5,900 per month

Scale, governance and procurement.

Talk to us

Compare all plans All plans run on agentpentest.app/portal. Prices exclude VAT (intra-EU B2B is reverse-charged). Annual prepay available.

Why now

A self-reinforcing platform

Agent adoption is outpacing the tools to verify what agents actually do. Regulation is arriving. AgentPentest sits between the two.

  • Generate

    Swarm runs produce execution-verified trajectory data.

  • Train

    We use that data to train stronger, safer agent models.

  • Test

    Those models independently test other companies’ agents for behavioral risk.

  • Compound

    Every test run generates more trajectories. The loop tightens each time.

Built for regulation

EU AI Act Annex IV documentation and auditable evidence are platform outputs, so compliance becomes a reason customers buy.

EU-based and sovereign

A European product for teams that need control over their data, models and supply chain.

Running today

The platform is live and self-serve. Create a workspace and run your first evaluation in minutes.

Get started

Run your first evaluation.

Create a workspace, connect one agent endpoint and run a standardized scan. Upgrade when you need more runs, reports or exports.

Get started See how it works Talk to RNDLabs

AgentPentest is owned and operated by RNDLabs Oy.