# Faultprint Core Baseline v7 Summary

## Status

v7 is the current working core baseline.

It is not the final commercial scoring engine, but it is the first baseline to clear the rule-based calibration audit sample after targeted fixes.

## Counts

- Operating rows: 680
- Crosswalk rows: 680
- Eligibility rows: 680
- Calibration audit sample: 65 rows
- Audit result: 65 pass, 0 revise, 0 hold

## Primary Pathology Distribution

- Contestability Lag: 249
- AI Authority Acceleration: 211
- Record-Authority Capture: 71
- Access Interruption Before Review: 43
- Classification Gatekeeping: 28
- Documentation Burden Transfer: 26
- Provisional Relief Reversal: 24
- Evidence-Container Illusion: 10
- Deadline Compression: 9
- Multi-Actor Handoff Diffusion: 9

## Resemblance Distribution

- moderate_resemblance: 384
- elevated_resemblance: 235
- low_resemblance: 59
- high_resemblance: 2

## Client Output Distribution

- client_ready: 334
- client_ready_with_caveat: 293
- revise_internal: 29
- hold_internal: 24

## What v7 Means

v7 means the engine now has a working baseline for:

- pathology assignment
- postmortem resemblance banding
- client-output eligibility gating
- caveat/hold handling where support is thin

## What v7 Does Not Mean

v7 does not mean the product is fully validated.

The next defensibility step is human review of the v7 audit sample and then expansion of postmortem support for weak categories, especially Provisional Relief Reversal, Classification Gatekeeping, Multi-Actor Handoff Diffusion, and Deadline Compression.