Skip to content

Security: OkamiOps/okami-sentinel

Security

SECURITY.md

Security policy

English · Português (Brasil) · Deutsch · Français

Supported version

Security fixes currently target the latest commit on main. Before a stable release, older commits and local forks are not covered by a backport guarantee.

Report a vulnerability

Do not disclose exploitable details in a public issue. Use the repository’s Security → Report a vulnerability flow when available, or contact the repository owner privately through GitHub.

Include:

  • affected component and commit;
  • reproduction steps or a minimal proof of concept;
  • expected impact and attack prerequisites;
  • whether local scanner output, GitHub Checks, or repository data is exposed;
  • any mitigation already tested.

Please allow time for validation before public disclosure. No response-time SLA is promised while the project remains pre-stable, but actionable reports will be triaged as a priority.

Security boundaries

  • Scanner output, findings, paths, logs, and repository content are untrusted input.
  • The default product is local-first. Data leaves the machine only through an explicitly requested integration such as GitHub Checks or an API-backed GitHub Actions run.
  • Operational failures must never become passing security decisions.
  • OPENAI_API_KEY is a GitHub Actions secret. The application diagnoses its presence but does not read or persist its value.
  • Managed scan deletion may remove local scan output; the target and effect must remain explicit in the UI.

Sensitive data

Never attach real secrets, private source code, full scanner state, database files, or personal filesystem paths to a public issue. Redact logs and provide the smallest artifact that reproduces the problem.

There aren't any published security advisories