openalice-tenants
Architecture rating
Manifest present, valid, and descriptive (owns 12 tables, exposes 8 endpoint groups, 9 runtime deps); README excerpt is clear and purposeful; roadmap score 80/100; however no evidence of @feature annotations, in-code architecture notes, or API documentation beyond manifest; docs score carried primarily by strong manifest discipline.
12 owned tables with clear domain boundaries (tenants, quotas, consent, audit_log, data_export_jobs); TENANTS_RLS_ENFORCE env var implies RLS exists but is toggle-gated rather than always-on; GDPR export/delete mentioned in manifest; no inspector run so FK/constraint quality, migration hygiene, and tenant isolation completeness are unverifiable; drift score 0/100 with 28 undeclared env vars is a red flag for schema-config coherence.
FGA_ENFORCE+FGA_SHADOW suggest authz layer exists; TENANTS_RLS_ENFORCE implies RLS toggle; however 28 undeclared env vars (incl. MOLLIE_WEBHOOK_SECRET, OA_SECRETS_MASTER_KEY, BREVO_API_KEY, MISTRAL_API_KEY) are used but not declared in manifest, signaling weak secret inventory discipline; no evidence of rate-limiting, SSRF guards, or security headers; AUTH_JWKS_URL present but no inspector data to confirm validation depth.
Manifest present and valid (100/100 manifest score); 47 features tracked; but code_health 30/100 and drift 0/100 with 28 undeclared env vars indicate poor adherence to org conventions; env drift is the worst in the quality breakdown, suggesting the codebase has diverged significantly from declared contract.
CI is failing (conclusion: failure, 2026-07-05); health probe cannot even find an HTTPS endpoint to check (up: false); no inspector data on graceful shutdown, timeouts, retries, or idempotency; code_health score 30/100 from Atlas suggests significant structural issues; failing CI on a multi-tenant control plane is a reliability concern.
Health probe misconfigured (no https endpoint in manifest exposes, up: false); uptime_24h_pct unknown; no evidence of structured logging, metrics export, or distributed tracing in available signals; no inspector data to confirm any instrumentation exists.
No inspector analysis to count tests or measure coverage; CI is failing so whatever tests exist are not passing; no evidence of test count, critical-path coverage, or regression tests; graded conservatively given zero positive signal.
source · auto-grader-2026-07-06
Telemetry ribbon · last 90 days
1 eventNo @feature annotations recorded in this repo yet.