Skip to content
@ocarinalabs

Ocarina Labs

independent safety testing for AI agents

Popular repositories Loading

  1. residency residency Public

    TypeScript 3

  2. aisea-www aisea-www Public

    TypeScript 1 1

  3. www www Public

    TypeScript

  4. aisafety.my aisafety.my Public

    TypeScript

  5. harbor harbor Public

    Forked from harbor-framework/harbor

    Harbor is a framework for running agent evaluations and creating and using RL environments.

    Python

  6. harbor-ext harbor-ext Public

    Adversarial probes for Harbor benchmark integrity. Ocarina Labs' Harbor extension — null-agent, output-echo, judge-injection, and verifier-tamper probes that flag broken benchmarks before they ship.

    Python

Repositories

Showing 6 of 6 repositories

Top languages

Loading…

Most used topics

Loading…