Product surfaces
The dashboard is split into a small number of named surfaces. Most operator tasks happen in one of these.
Reviews
The default landing surface at /. Lists recent reviews and opens detail for the selected one — story insights, readiness, and a rerun affordance when readiness allows it. Active runs poll for status every few seconds while you watch.
Findings
At /findings. The cross-review backlog. Filter by severity, source, exploitability, delta state (new, recurring, regressed, fixed, suppressed), category, suppression scope, and free-text search. Selecting a finding opens identity, location, evidence, replay, and bundle access.
Security
At /security. Per-account scanner and runtime-agent configuration: enabled scanners, custom rule paths, severity threshold, exclude paths, exclude rules, runtime agent set, max plans, wall-clock and per-agent timeouts, and evidence detail level. Saved values become dashboard defaults at scan launch — see Configuration.
Settings
At /settings. Connection state and automation preferences:
- GitHub auto-review repository list (empty list = run on any installed repo).
- Own-PR-only filter — restricts auto-review to PRs whose author matches your connected GitHub username.
- Push debounce window — collapses bursts of
synchronizeevents. Range 1–120 minutes. - Encrypted repo secrets injected into the sandbox at scan time.
- Completion email preferences.
Scan history
At /scan-history. Run timeline and quality view: phase durations, scanner coverage, compare counts, trust score, gaps, and recommendations. Useful when triaging a degraded scan or deciding whether to rerun.
Scan run detail
At /runs/{id}. Status badge, branch badge, started and completed timestamps, a degradation notice when applicable, the findings list, live progress while running, and a CSV download once the run reaches a terminal status.
Command palette
Press Cmd K (or Ctrl K) to navigate across reviews, findings, settings, security configuration, scan history, and finding detail.