Skip to content

Security: alanmaizon/nebula

Security

SECURITY.md

Security Policy

Supported Versions

Security updates are currently provided for:

Version Supported
v1.0.x Yes
main (next changes) Best effort
v0.x snapshots Best effort

Reporting a Vulnerability

Please report vulnerabilities through GitHub private vulnerability reporting:

  • Go to the repository Security tab.
  • Choose "Report a vulnerability" to open a private advisory.

Do not create public issues for suspected vulnerabilities.

If private reporting is unavailable, contact the maintainers directly and include:

  • affected component and version/commit
  • reproduction steps
  • impact assessment
  • proposed mitigation (if known)

Response Targets

  • Initial triage acknowledgement: within 2 business days.
  • Risk assessment and remediation plan: within 5 business days.
  • Critical vulnerability fix target: within 7 calendar days.
  • High severity fix target: within 14 calendar days.

Targets may vary for complex infrastructure dependencies, but status updates will be provided.

Disclosure Process

  • Reports are validated and severity-scored.
  • A fix is prepared in a private branch when possible.
  • Coordinated disclosure occurs after a patch is available.
  • Release notes summarize impact, affected versions, and mitigation.

Security Baseline for Contributors

  • Never commit secrets or credentials.
  • Use least-privilege IAM and scoped credentials for cloud resources.
  • Avoid logging raw document contents or personally identifiable information.
  • Validate all model outputs against explicit schemas.
  • Reject citations that cannot be traced to indexed evidence.
  • Keep dependencies up to date and address CI security alerts promptly.

Logging and Redaction Rules

  • Every API response includes a request correlation header (X-Request-ID by default).
  • Logs are structured and scoped to request metadata (method, path, status, latency).
  • Do not log request or response bodies containing uploaded document text.
  • Redact sensitive keys before logging: authorization, cookie, token, password, secret, API keys, email, phone, and SSN fields.
  • Redact sensitive inline patterns in any logged string (email addresses, phone numbers, SSNs, bearer tokens, AWS access keys).

There aren’t any published security advisories