Define stable Finding, Scan Result, and comparison identity

[BleedingDev]

Question

What canonical schema distinguishes raw observations, evidence, normalized Findings, scores, action classes, External References, analyzer coverage, the Improvement Backlog, and current-versus-previous comparison? Define stable finding identity across commits and the exact machine-readable Scan Result without collapsing metrics into Findings.

[BleedingDev]

Resolved with the user.

Preserve each raw analyzer observation and its provenance. Build normalized Finding identity from the analyzer rule family, normalized repository-relative location, and enclosing symbol when available. Related cross-tool Findings are grouped for display and prioritization but never destructively merged. The Scan Result contains immutable scan metadata, analyzer coverage, evidence, normalized Findings, separate scores, one ordered Improvement Backlog, and a comparison against the preceding scan.