Trust & Security

This page is maintained by the RightRole team to answer common security and privacy questions about RightRole. It describes controls that are in place today and items on our roadmap. It is not an independent certification.

Access Control

• Role-based access model: internal roles — owner, admin, support, developer, customer_user, and champion_guest — gate access to administrative surfaces and operational tooling.
• Row-Level Security (RLS): user data is protected by RLS policies on the database, scoped to the authenticated user. Privileged server operations use security-definer functions for role checks.
• Admin allowlist fallback: a small hardcoded allowlist is retained as a short-term safeguard alongside database-stored roles.

Application Security

• CI quality gates: every pull request runs dependency install, lint, typecheck, and build. Failing checks block merges.
• Dependency monitoring: weekly automated dependency and GitHub Actions updates via Dependabot.
• Security scanning: automated scans flag risky patterns and configuration drift.
• Input validation: server endpoints validate inputs and authenticate callers; public webhook endpoints verify signatures before processing.

Observability

• Structured logging with redaction: server logs are emitted as structured JSON. Sensitive content (resumes, job descriptions, cover letters, prompts, tokens) is automatically redacted.
• Correlation IDs: request identifiers thread through logs and error reports for safer troubleshooting.
• Error monitoring: Sentry-ready client and server-side error capture, including React error boundaries and unhandled promise rejections.
• Health check: a public /api/public/healthz endpoint reports application and database connectivity for uptime monitoring.
• Privacy-safe product analytics: when configured, RightRole sends a small, allowlisted set of explicit product events to PostHog (no autocapture, no session recording). Resumes, job descriptions, prompts, AI completions, cover letters, and champion/referral messages are never sent. Role-fit scores are bucketed to low / medium / high before transmission. When the analytics key is not configured the wrapper is a no-op.

Operational Readiness

• Incident response runbook: standardized "Detect → Investigate → Mitigate" playbooks for application, database, payment, email, and AI provider outages.
• Production checklist: pre-deploy validation, smoke testing of core flows (scoring, save, champion links, admin access), and a documented rollback procedure.

Vulnerability Disclosure

Found a security issue? Please report it privately. Our policy is published in SECURITY.md and reports can be sent to security@rightrole.ai. We support coordinated disclosure and credit verified reporters.

What We Do Not Claim

RightRole is an early-stage commercial SaaS. We do not currently claim SOC 2, ISO 27001, HIPAA, or GDPR certification, and we do not offer a formal Data Processing Agreement (DPA) at this stage. If you need any of these for a procurement process, contact us so we can discuss timelines.

Your Data Rights

• Self-service data export: signed-in users can download a JSON copy of their profile, pursuits, champions, generated artifacts, and activity from Settings → Data & Account.
• Account deletion request: signed-in users can submit a deletion request from the same panel. Requests are reviewed and processed manually by our team; data is not hard-deleted on submission. Status moves through requested → processing → completed (or rejected with a note).

Usage during beta

• Free during beta: RightRole is free for everyone while we're in beta. There are no subscriptions and no monthly limits. Optional one-time contributions are available on the Support RightRole Beta page.
• Usage tracking: we count how often you use features like role scoring or champion invites so we can improve the product. We log feature keys, counts, and ids only — never CV text, job descriptions, prompts, AI completions, cover letters, or message bodies.
• Atomic recording: AI-heavy workflows (resume personalization, cover letters, champion intro emails, champion outreach drafts, company research, champion invites, and new pursuit saves) record usage in a single atomic database transaction, under a per-user / per-feature lock. Direct writes to the usage ledger are locked down at the database level; only the server can record usage.
• Payments: live subscription checkout is not enabled yet. Voluntary supporter packs are processed through Stripe; we do not store card details.

Roadmap

• Formal Data Processing Agreement (DPA) for enterprise customers.
• Automated, fully self-service account deletion (no manual review).
• Backup and restore runbook with periodic restore drills.
• External uptime monitoring with public status reporting.
• SOC 2 readiness work as the company matures.

Contact

Security: security@rightrole.ai
Privacy: privacy@rightrole.ai