Loading the collection…
High-quality research data and security task environments for agent training and evaluation. Explore the tasks, inspect the environment, and understand how results are scored.
Explore the catalogClear objective · task context
Defined workspace · tools · inputs
Submission criteria · reward contract
Research tasks, environments, verifiers and trajectories.
Security tasks, Docker sandboxes and verifiers.
Estimated delivery capacity by product type. Scope, task mix and delivery schedules are agreed with your team.
High-quality RL task packages for scientific reasoning and computation. Real research problems with Docker environments, independent verifiers and complete agent trajectories.
Security tasks, Docker-based sandbox environments, and inspectable verifiers for agent training and evaluation. Build your own agent experience in synthetic business workflows.
Need a different domain, task mix or delivery format?
Discuss a tailored collectionEvaluate what each task tests, how its environment supports the work, and what its verifier actually measures.
Scientific research and synthetic security workflows, with defined objectives, context and expected outputs.
Inspect the task context, Docker setup and declared resources. Understand the conditions in which your agent works.
Read the actual verifier and its scoring contract. Distinguish the checks it defines from the results recorded in a particular run.
Scientific packages include archived trajectories and outputs. Security examples show available run summaries and clearly identify missing scores.
Start a conversation with our team. Review sample materials by invitation, agree on what works, and build a data supply around your training and evaluation needs.
Tell us your domain, training objectives and the tasks or data you need.
Receive an invitation code to inspect the available sample materials.
Review task quality, environments, verifier behavior and available evidence. Align on suitability and acceptance criteria.
Confirm volume, format, usage rights, pricing and delivery milestones, then sign the agreement.
Receive data in agreed batches, complete delivery acceptance and extend the collection as your needs grow.
Discuss available tasks, a tailored collection or ongoing production. We’ll help you choose samples to evaluate first.