falsify
Fix the pass bar before the test runs. PRML is an open spec and CLI that hashes and timestamps evaluation criteria (SHA-256, RFC 3161) so anyone can check them later. Python, JS, Go, Rust.
- 8
- Python
- MIT
- 6 h ago
falsify-dev
Fix the pass bar before the test runs. Open spec, four reference implementations, public registry.
Built from falsify-dev's public GitHub profile. Not on Take My Repo yet, so there are no posts or updates here.
Fix the pass bar before the test runs. PRML is an open spec and CLI that hashes and timestamps evaluation criteria (SHA-256, RFC 3161) so anyone can check them later. Python, JS, Go, Rust.
Field manual for the PRML v0.1 specification. 13 patterns, 4 anti-patterns, 4 working examples. CC0.
Do AI agents ask for the evidence that would change their verdict? Synthetic evaluation-claim cases, runner, raw responses and report (Apart AI x Epistemics Sprint, Nov 2026)
Verify PRML (Pre-Registered ML Manifest) commitments in CI. Block merges on tampered or regressed eval claims.
MLflow plugin: automatic PRML manifest hash tagging for runs. Pre-Registered ML Manifest evidence layer.
PRML pre-registration for Giskard scenario results: commit an eval claim to a SHA-256 before the run, then verify the result against it.
PRML adapter for Inspect AI: record an eval's pass criteria in a hashed manifest and check eval logs against them. Listed in the Inspect extensions catalog. MIT.
JavaScript reference implementation of the PRML (Pre-Registered ML Manifest) v0.1 specification. Byte-equivalent to the Python reference across 21 conformance vectors. Zero runtime dependencies.
Public scorecard of how 27 ML eval claims meet 9 PRML falsifiability criteria. CC0 data; MIT tooling.