Pinned Loading

1. evalwarden evalwarden Public

   Integrity auditing for agent evals. Lints evals for leakage, grader defects, and cost-per-success instead of running them.

   Python 1

2. open-redactor open-redactor Public

   Open-source CLI to make photo, videos, and audios share-safe with prompt-driven redaction

   Python 1

3. reward-contract-fuzzer reward-contract-fuzzer Public

   Find the policy your reward function actually incentivizes, before an agent does

   Python 1

4. proof-of-steer proof-of-steer Public

   Local tool that turns an AI coding session or a document's git history into a steering receipt and a steerprint, a map of which parts the human shaped, so a reviewer can price their attention. Stee…

   Python