Skip to content

Security: Nova-Violet-Role/.github

Security

SECURITY.md

Security Policy

Applies to every repository in the Nova-Violet Role organisation unless that repository ships its own SECURITY.md, which then takes precedence.


Reporting a vulnerability

Do not open a public issue.

Use GitHub's private vulnerability reporting on the affected repository:

Repository → SecurityAdvisoriesReport a vulnerability

That channel is private between you and the maintainer, gives us a place to work on a fix before it is public, and lets us credit you in the advisory.

If the repository has advisories disabled and you believe you have found something real, contact @Saimonokuma on GitHub and say only that you have a security report — no details in the open.

What we will do

Acknowledge Within 7 days. If you have not heard back, assume it was missed and ping again — that is not rude.
Assess We reproduce it, or tell you exactly where we failed to.
Fix With a regression test that we have watched fail before it passed.
Disclose Via a GitHub Security Advisory, crediting you unless you ask us not to.

We will tell you honestly if we decide not to fix something, and why. A disclosed limitation is preferable to a silent one.

What counts as a vulnerability here

These projects are developer tooling. The interesting attack surface is not a web server — it is the trust boundary between a claim and its evidence:

  • Gate evasion — making a check report success while the property it checks is false.
  • Record or attestation forgery — producing an artefact that a verifier accepts as evidence of work that never happened.
  • Fence escape — content crossing a boundary that is documented as holding it, including untrusted text reaching a place where it is treated as instruction rather than data.
  • Injection through tool output — data from a file, log, network response or another agent being executed, expanded, or obeyed.
  • Privilege or path escape — a hook, checker or installer touching anything outside its declared footprint, or writing outside the paths it declares.
  • Credential exposure — any path by which a token, key or secret could be written to a log, an artefact, a commit, or a published archive.
  • Supply chain — a release archive whose contents do not match its tracked sources.

A proof-of-concept that makes a gate lie is the highest-value report we accept.

What is not a vulnerability

  • A documented limitation. Check the repository's review or limits document first; we publish what these tools do not claim.
  • A finding that requires an attacker who already controls the machine the tool runs on. If they own the shell, they do not need our bug.
  • Output from an automated scanner with no demonstrated impact. Show the path.

Reporting one of these in good faith is not a problem — you will get a real answer explaining which category it fell into.

Supported versions

The default branch, and the most recent tagged release of each repository. Older tags are not patched.

Safe harbour

We will not pursue or support action against anyone who, in good faith:

  • researches vulnerabilities in these repositories,
  • reports them privately through the channel above,
  • gives us reasonable time to fix before public disclosure,
  • and does not access, modify or exfiltrate anyone else's data in the process.

Testing against your own checkout is always fine. These are local developer tools; there is no production service of ours to attack.


Nova-Violet Role · Reality is the judge.

There aren't any published security advisories