4 projects
open-redactor
Open-source CLI to make video share-safe with SAM 3.1 prompt-driven redaction
reward-contract-fuzzer
Find the policy your reward function actually incentivizes, before an agent does. Optimizer-proved reward contract testing for RL environments.
compaction-conformance-kit
Find out what your AI agent forgets when context compaction runs: typed canaries, per-type survival curves, and cliff detection for safety rules, facts, and task state.
evalwarden
A linter for agent evaluations: audit the measurement system around a score.
