Skip to content

Security: SafetyMP/Autonomous-EHS-Management

SECURITY.md

Security policy

The deployable application includes authentication, tenant and role boundaries, scheduled jobs, and persistent records. The separate GitHub Pages walkthrough uses synthetic data and has no application backend or sign-in. Neither the walkthrough nor its accessibility checks validate a deployed application’s security.

Supported versions

Area Supported
Latest main Yes
Prior release tags Org policy

Reporting a vulnerability

Do not open a public GitHub issue for active security findings.

Use one of:

  • GitHub private vulnerability reporting for this repository (Report a vulnerability on the Security tab).
  • Your contractual / security channel with SafetyMP (do not disclose active exploits on a public issue).

Include: affected routes or components, reproduction steps, impact, and suggested fix if any.

We aim to acknowledge within 7 business days; timelines for fix and disclosure depend on severity and coordinated disclosure with your security team.

Scope (application)

Out of scope unless agreed in writing: third-party SaaS-only issues, generic dependency advisories already tracked by Dependabot (see PRs / npm audit policy in CI), social engineering.

Credentials

Never commit secrets. Use .env* examples only; rotate any credential that was ever pasted into an issue or PR.

The committed .env.ci file contains non-secret placeholders that satisfy runtime env schema in CI only. Do not point it at production databases or reuse its BETTER_AUTH_SECRET outside disposable CI.

HTTP security headers

next.config.ts sets a baseline Content-Security-Policy (including frame-ancestors 'none') plus standard headers. If you enable browser-visible OIDC flows or third-party script/embeds, review and extend CSP connect-src / script-src with counsel and security review.

Optional CRON_FAILURE_WEBHOOK_URL: HTTPS webhook (e.g. Slack incoming) notified when a cron route throws; see src/server/cron/notifyOnFailure.ts.

There aren't any published security advisories