Skip to content

Security: ScriptKittyOS/kotiko

SECURITY.md

Security policy

Supported versions

Only the latest release gets security fixes: from 1.0.0 on (the first store release, still pending), the latest 1.x. Release candidates (vX.Y.Z-rc.N pre-releases, for review) and versions older than 1.0.0 get none.

Reporting a vulnerability

Please report privately. Don't open a public issue.

  • Preferred: GitHub's private vulnerability reporting. Open the repository's Security tab and choose Report a vulnerability.
  • No GitHub account: email security@scriptkittyos.com.

Please include:

  • what an attacker can do, and what they need first (network access, a web page the user visits, access to the user's machine);
  • steps to reproduce, ideally with a request or a small page;
  • the Kotiko version, browser and operating system.

We acknowledge reports within 7 days and send a fix or a plan within 30. If you'd like credit, we'll name you in the release notes.

What happens after you report

  1. Acknowledge within 7 days, from a maintainer named in MAINTAINERS.md.
  2. Triage: we reproduce the problem, decide whether it is in scope, and rate it with CVSS v4.0. We tell you the rating and the plan.
  3. Fix privately in a GitHub draft security advisory with a temporary private fork. You can be added to the advisory to review the fix.
  4. Targets from triage to a released fix: critical 7 days, high 30 days, medium 60 days, low in the next planned release. If we will miss a target, we tell you why.
  5. CVE: for medium severity and above, we request a CVE through GitHub when we publish the advisory.
  6. Release: a patch release whose notes have a "Security" section listing the advisory, the CVE, the severity and, unless you ask otherwise, your name.
  7. Disclose: the advisory is published with the release. We ask reporters to keep details private until then or for 90 days from the report, whichever comes first.
  8. Afterwards: the fix includes a regression test, and the assurance case is updated if the problem crossed a trust boundary.

If the problem is in a dependency, we report it upstream and ship a pinned or patched version in the meantime.

We also look for problems ourselves: every pull request and every release is checked for known vulnerabilities in our dependencies and for security weaknesses in our code. What blocks a merge or a release, how fast findings are fixed, and how a false positive is accepted are in the dependency and static analysis policy.

The security response lead named in MAINTAINERS.md runs this process; GOVERNANCE.md describes the role.

Scope

In scope:

  • the server: authentication, token handling, request parsing, anything reachable over the network;
  • the extension's own pages (popup, dashboard, welcome tab) and how they store keys and tokens;
  • anything Kotiko runs on web pages, including what a page can learn about the user from it.

Out of scope:

  • a self-hosted server deliberately exposed without a token or over plain HTTP against the documentation's advice;
  • problems in Telegram, OpenRouter or other model providers themselves.

There aren't any published security advisories