Author checks once. Then run, monitor, and guard every repository from the same Scanner workspace.
Scanner turns migration sweeps, policy drift checks, dependency hygiene, and PR enforcement into one connected workflow with real repository state, live findings, and enforcement that stays attached to the same workspace.
Workspace lane
Run → Monitor → Guard
name: "Legacy button imports"
engine: imports
targets:
include:
- "src/**/*.{ts,tsx}"
detect:
imports:
- from: "@legacy/ui/Button"
threshold:
fail_if_matches_gt: 0Scanner keeps the authoring surface close to the evidence surface, so teams can move from idea to enforcement without rebuilding the rule elsewhere.
Repository sync
GitHub App installation, repo indexing, branch snapshots, and PR context.
Authoring
Template launchers, YAML editing, AI config generation, and validation.
Execution
Manual runs, scheduled monitors, findings timelines, and alert delivery.
Enforcement
PR comments, soft-fail, and hard-fail guardrails backed by evidence.
Authoring surfaces
Templates
Start from migration, dependency, structure, or risky-pattern presets.
YAML editor
Work directly in the Scanner config shape and validate estimated scope before running.
Prompt to config
Describe the rule in plain language and let Scanner draft the initial config.
What the MVP closes
Live workspace posture
Workspace access, onboarding, and the app shell all rely on the same authenticated Scanner APIs.
Real product surface
Public pages explain the same repository, findings, and workflow model that powers the app shell.
One product language
Public pages and the authenticated shell share the same premium Scanner brand treatment.
Operational surface
One workspace for search, evidence, alerting, and PR enforcement.
Evidence-first findings
Every finding keeps repo, path, line, snippet, message, severity, and lifecycle status in one place.
Monitors that stay useful
Scheduled checks can alert only on net-new findings and route to Slack, email, or webhooks.
Guardrails teams can trust
Promote stable checks into comment, soft-fail, or hard-fail enforcement once a rule is trusted.
Move from one-off sweeps into a repeatable engineering control plane.
Authenticate with GitHub, create the workspace, connect repositories, and start enforcing with real data.