This is the default security policy that propagates to my repositories without their own. Repos that handle credentials, network surfaces, or sensitive workflows ship a more specific
SECURITY.md(e.g., gone-phishing).
For most of my repos:
| Version | Supported |
|---|---|
main / master (latest) |
✅ |
| Latest tagged release | ✅ |
| Older tagged releases |
Repos that ship security-sensitive functionality may have a more specific support matrix in their own SECURITY.md.
Do not report security vulnerabilities through public GitHub issues, discussions, or PRs.
Use one of these private channels:
- Preferred: Open a GitHub Security Advisory on the affected repo (private — only visible to maintainers and you)
- Alternative: Reach me through my GitHub profile — current contact methods are listed there
- Clear description of the issue and its impact
- Steps to reproduce (proof-of-concept where applicable)
- The version / commit hash where you observed the issue
- Your assessment of severity (critical / high / medium / low) and reasoning
- Whether you've disclosed the issue elsewhere
- Acknowledgment within 72 hours
- Triage within 7 days with a rough timeline
- Fix or mitigation within 30 days for high/critical issues; lower-severity may take longer
- Coordinated disclosure. Once a fix ships, I'll publish a security advisory crediting you (with your permission). Anonymous reports are also fine.
- Vulnerabilities in upstream dependencies — please report to the respective maintainer (and feel free to ping me as well so I can update the affected repos)
- Vulnerabilities in third-party services my code integrates with — please report to the respective vendor
- Issues that require physical access to deployment hosts
- Social engineering attacks against me or contributors
- Denial-of-service attacks against your own deployment
Across the repos here:
- Credential leakage via error messages, logs, or response bodies
- Authentication bypass in any deployment configuration
- Prompt injection that escalates privilege or exfiltrates data in AI-integrated repos
- Path traversal in any file-handling code
- XSS that survives sanitization layers in chat/UI repos
- Dependency vulnerabilities I've missed in
requirements.txt/package.json/mix.exs
If you're not sure whether something qualifies as a vulnerability, err on the side of reporting privately.