Skip to content

Security: Universal-Standard/.github

SECURITY.md

Security Policy

This is the organization-wide default security policy for Universal-Standard repositories. It applies to any repository that does not define its own SECURITY.md.

Reporting a Vulnerability

Do not open a public GitHub issue for a security vulnerability.

Report it privately using one of the following, in order of preference:

  1. GitHub Private Vulnerability Reporting — on the affected repository, go to the Security tab → Report a vulnerability. This creates a private advisory visible only to maintainers and lets you collaborate on a fix before public disclosure.
  2. Email — if the repository does not have private reporting enabled, email philip.cotton@spurs.agency with:
    • A description of the vulnerability and its potential impact
    • Steps to reproduce (proof-of-concept code or a minimal repro is extremely helpful)
    • The affected repository, file(s), and commit/version
    • Whether you believe the issue is already publicly known or exploited

What to expect

Stage Target timeline
Acknowledgment of your report Within 3 business days
Initial assessment (severity, affected repos) Within 7 business days
Fix or mitigation plan communicated to reporter Within 30 days for high/critical severity
Public disclosure Coordinated with the reporter, typically after a fix is released

These are targets, not contractual SLAs — actual timelines depend on severity and complexity.

Supported versions

Unless a specific repository's own SECURITY.md states otherwise, only the latest release (or the main/default branch, for repositories without tagged releases) is supported with security fixes.

Scope

This policy covers vulnerabilities in code, configuration, and infrastructure-as-code maintained in Universal-Standard repositories. It does not cover:

  • Vulnerabilities in third-party dependencies (report those upstream; we will still take a dependency-bump fix as a contribution)
  • Social engineering or physical security issues
  • Denial-of-service reports based purely on load/rate, without a specific exploitable flaw

Recognition

With the reporter's consent, we credit valid reports in the relevant release notes or security advisory.

There aren't any published security advisories