Skip to content

Security: studio-11-co/falsify

Security

.github/SECURITY.md

Security policy

Supported versions

Currently 0.1.x is the only supported line. Older pre-releases are not maintained; please upgrade before reporting.

Reporting a vulnerability

If you find a way to:

  • Make falsify guard pass on a contradicting claim.
  • Make two semantically different specs produce the same SHA-256 canonical hash (a collision breaks pre-registration).
  • Execute arbitrary code through a crafted spec.yaml or experiment.command beyond what the user explicitly authored.

Please do not open a public issue. Email the maintainer directly at hello@studio-11.co. A fix typically lands within 14 days; we'll coordinate a disclosure window with you.

What counts as a vulnerability

Any break of the determinism contract counts:

  • Exit codes — getting a 0 when the criterion was violated, or 10 when it held.
  • Canonical hash — two distinct specs hashing the same, or the same spec hashing differently across machines.
  • Guard semantics — commit-msg guard passing text that affirmatively references a FAIL / INCONCLUSIVE verdict.
  • Leakage — anything that exposes secrets from spec.yaml, run output, or the verdict store to unauthorized readers.

Cosmetic bugs, typos, or missing features don't qualify — use a bug report or feature request instead.

Credit

Disclosed issues are credited in CHANGELOG.md under the release that fixes them, unless the reporter requests otherwise.

Full threat model

For the complete enumeration of attacks falsify defends against, attacks it explicitly does NOT defend against, and the mitigations for each, see docs/ADVERSARIAL.md.

There aren't any published security advisories