Compliance Center

The Compliance Center runs 21+ AI experts against your codebase — Security, Privacy, Accessibility, Performance, Code Quality, Architecture, Licensing, Ethical AI, and more. Each expert is a focused single-turn assessor that produces findings.

The Dashboard has two jobs

Run scans — pick experts (single, group, all) and click Run selected.
Summarise scan state — three tiles show Last scan, Needs review, Open findings. Click any tile to jump to the corresponding surface.

It is not a scoreboard. There are no per-expert scores or critical / high count badges on the rows. That noise lives on the audit cards themselves.

Status badges

Each expert row carries a badge that mirrors the expert's Mission Board audit card lifecycle:

○ Not run — never scanned
⟳ Running… — audit in progress
✓ Clean — audit completed, no findings, card auto-closed to Done
● Review → — audit has findings, card in Review (click to jump)
⚠ Failed → — audit errored, card in Review

Where the actionable work lives

Audit cards on the Mission Board carry severity chips (2 critical · 3 high) and a triage-progress chip (3 / 4 triaged).
Findings tab lets you browse and filter all findings across audits.
Settings tab lets you enable / disable individual experts; only enabled experts appear on the Dashboard.

Triage = Brainstorm

You don't act on findings via direct UI buttons. Click Open in Orbit (per-finding or per-card) and Orbit opens a Brainstorm session preloaded with the finding(s) plus a mandatory staleness check. You discuss; agreed actions (promote to fix card, dismiss as stale, accept risk) execute as Brainstorm side-effects. See Brainstorm-First.

Settings is the source of truth for which experts appear

Disabling an expert in Settings removes it from the Dashboard entirely (and from the Select all count). Categories with no enabled experts collapse out.