Compatibility Posture v0
Per-repo readiness derived from what Rosentic already sees. Detects exposure, validates it against stored evidence, and tells. Not a gate.
Enter an org to load posture.
Methodology — posture_formula_v0. Readiness starts at 100.
Penalties (each capped): blast-radius exposure (dependent call sites of open UNSAFE findings),
recurrence (findings recurring across scans in the trailing 90 days), and agent overlap
(distinct agent pairs active in the repo). A blocking gate policy earns headroom back.
Every score is stored with its formula version and exact inputs, so it is reproducible.
v0 is an early read model; weights will change and the version will change with them.