souspilotSecurity Back to Souspilot
Security / Disclosure policy

Coordinated vulnerability disclosure for Souspilot discovered vulnerabilities

A framework for working with maintainers to turn security findings into fixes and keep users informed.

Last updated Effective date

Purpose & scope

This policy covers human-validated vulnerabilities Souspilot discovers in open-source software or authorized research on proprietary software. It excludes unreviewed AI findings and incoming reports about Souspilot.

Disclosure timelines

We notify maintainers promptly. Public disclosure is targeted at the applicable deadline or patch release, whichever comes first, unless security considerations warrant a change.

Calendar days from our initial report
FindingTarget
High or critical severity30days
Other severities45days
Active exploitation7days

For active exploitation, we also target a patch or mitigation within seven days. An engaged maintainer may request seven additional days.

After 15 days without a response, we aim to involve an external vulnerability coordinator; the disclosure deadline still applies.

These timelines apply to reports sent on or after November 16, 2026. Earlier reports retain their existing deadlines.

Coordination & exceptions

Maintainers progressing toward remediation may request a 14-day extension near the standard deadline. We generally accept their severity assessment; credible exploitation evidence overrides it.

For ecosystem-wide issues, we coordinate notification and support before publication. Exceptional circumstances or complex remediation may require adjusted deadlines; we communicate changes and reasons as appropriate.

Technical details

We generally hold full technical details for 30 days after patch availability, allowing users to update. This is separate from initial disclosure.

We may shorten that buffer when details are already public or defenders need them against ongoing attacks, or extend it for difficult deployment or widespread impact.

Report quality

A researcher validates findings before submission. We identify AI-assisted reports, label candidate patches by origin, and help maintainers develop production fixes where possible. Reports include actionable guidance and available mitigations.

We agree on a sustainable submission pace before sending many findings to one project. Active exploitation takes priority.

For all communications: jcc@security.souspilot.com