⚠Risk Analysis

Catch architectural drift before it merges.

Every push runs a full structural scan. A failing GitHub check, an email, and a ranked list of violations land in front of whoever needs to look first.

Join the waitlist →
FILERISK
auth/session.tsCRITICAL
api/payments.tsHIGH
utils/crypto.tsMEDIUM
components/FormLOW

fig 1. risk report — sorted by severity

How it works

From push to failing check, with the team notified.

▶GitHub App: anchoria
permissionscontents: read
permissionsmetadata: read
permissionschecks: write
eventspush, pull_request
Installed · webhook queue connected
×DriftGuard / Drift ScoreFailing
Drift score 0.72 (orange)
Threshold 0.60 · band: orange
Result: threshold breached. Review changed code before merge.
top violations
orphaned exportlib/payments/legacy.ts:42+0.28
missing call beforeroutes/charge.ts → rateLimiter+0.22
boundary importapps/web → packages/engine-internal+0.18

The check on every PR.

A real GitHub Check Run, with a score, a band, and the violations that produced it.

score + band0.00–1.00 score, mapped to green / yellow / orange / red. Threshold is per-repository.
ranked violationsEach violation contributes a weight. Highest weights show up first; the rest live in the dashboard.
email + slackWhen the band turns red or threshold breaks, the team gets a deduplicated alert in their channel.

Frequently asked
questions.

Code generation tools write code for you. Code review tools point out problems in your code. Anchoria teaches you what your AI-generated code actually does, encodes the architectural intent behind it, and blocks every future change that quietly breaks it. It's the difference between a spell checker and a grammar checker: one fixes typos, the other teaches you to write better sentences and prevents you from writing worse ones.