Skip to content

Security: arturict/marktake

SECURITY.md

Security policy

Supported versions

Version Security fixes
0.1.x Supported
Earlier or unreleased snapshots Not supported

Reporting a vulnerability

Do not open a public issue for a suspected vulnerability.

Use GitHub's private vulnerability reporting for this repository:

https://github.com/arturict/marktake/security/advisories/new

Include:

  • affected version or commit;
  • deployment assumptions;
  • reproduction steps or a minimal proof of concept;
  • impact and the security boundary crossed;
  • any known workaround.

Please do not access other people's systems or media, run destructive tests against a public instance, exfiltrate data, or publish details before a fix is available. No response-time or bounty promise is made.

Operator responsibilities

Marktake is self-hosted software. The operator is responsible for:

  • TLS and reverse-proxy hardening;
  • a strong unique administrator password;
  • host and volume access controls;
  • firewalling the Node port;
  • backups and disk monitoring;
  • legal authority to store and share uploaded media;
  • incident response and guest-link revocation;
  • keeping the image and host patched.

Marktake v0.1 does not include malware scanning, encryption at rest, SSO, audit logs, watermarking, DRM, or a hosted trust and compliance program.

Development dependency note

The production dependency audit is required to be free of known high and critical advisories. Development tooling currently uses major-compatible patched brace-expansion releases, but the upstream advisory database recognizes only the breaking v5 line as patched. These packages process repository-owned test and lint globs, not request input, and are absent from the production image. This residual build-tool alert is tracked rather than hidden with a broad production override.

Design overview

See docs/architecture.md for the owner, guest, browser, media, and storage boundaries. Security-focused tests cover spoofed uploads, size caps, interrupted streams, path-like metadata, CSRF, link revocation, IDOR, invalid annotations, XSS strings, byte ranges, and mismatched frame and time pairs.

There aren't any published security advisories