Skip to content

Security: QuantStrategyLab/QmtPlatform

Security

SECURITY.md

Security Policy

Reporting a Vulnerability

If you find a security issue in QmtPlatform, please report it privately to @Pigbibi instead of opening a public issue. Include what you found, the affected file or endpoint, and steps to reproduce if possible.

Secret and Credential Exposure

This repository must never contain real broker credentials, account identifiers, API tokens, or other secrets in source, configuration samples, test fixtures, or workflow logs. .env.example and the fixtures under data/ hold placeholder values only.

If you discover a committed secret or a real account identifier, report it the same way as a vulnerability (do not open a public issue describing it) so it can be rotated and removed.

Scope Notes

  • Current scope is offline dry-run only. QmtPlatform evaluates strategy targets and previews orders; it does not hold a live or paper broker connection, and any non-dry-run order request is rejected as blocked, never reported as submitted.
  • The offline --paper-admission gate (see README.md) is a local configuration check, not a miniQMT/QMT paper account connection — it does not call a QMT SDK/provider or read credentials.
  • Despite the dry-run-only scope, the code paths here still model order construction, strategy profile selection, and QMT/miniQMT integration points, so changes to those paths are treated as security-relevant and should be reviewed with that in mind.
  • The scheduled Runtime Target Lifecycle workflow runs against a disabled target with no broker runtime configured; it does not expose or require live credentials.

There aren't any published security advisories