# Audit card

## Intended use

Reproduce released derived parameters, quantify one-dimensional prior-to-posterior information gain,
and measure pairwise marginal divergence among the three released standard-PE waveform analyses for
GW241011 and GW241110.

## Not intended for

- discovery or event validation;
- hierarchical-merger or compact-object classification;
- population, formation-channel, ancestral, or eccentricity inference;
- replacement of multidimensional posterior review;
- qualitative strong/moderate/weak grading of KL or JS values.

## Evidence boundary

Primary data are the pinned paper-release v1 standard-PE samples. The pinned paper defines quoted
values and the HPD-in-cosine tilt convention. GWOSC O4 Discovery Papers v1 and PESummary parameter
documentation are secondary checks. GWTC-5 posterior samples are excluded from the primary analysis.

## Known limitations

Histogram estimates depend on binning and finite sample size, so the ledger reports resolution
sensitivity and bootstrap Monte Carlo variation. Marginal KL is prior dependent and non-additive.
Marginal JS cannot reveal changes in joint correlations. Prior differences can contribute to
posterior-analysis divergence. Artifact hashes establish integrity, not scientific equivalence
across changed environments.

## Current pinned-release block

All six individual-analysis `priors/samples` groups in the two released standard-PE HDF5 files are
empty, and the accompanying `config_file` groups are empty. The release therefore does not satisfy
the project's matching-prior contract, so KL and prior-vs-prior JS checks remain explicitly blocked.