Skip to content

Report a vulnerability

How to report a security problem in Klav privately, what to include, what happens next, and what is in scope.

Updated

This page explains how to report a security problem in Klav privately, what to include, and what happens next. It summarizes Klav’s security policy, SECURITY.md at the root of Klav’s source code, which has the full text and wins where they differ.

How to report

  1. Preferred: GitHub’s private vulnerability reporting, under Security › Report a vulnerability in Klav’s repository. The report stays private between you and the maintainers; the fix is made privately, and the advisory is published with it.
  2. If you can’t use GitHub: email security@klav.dev with the subject “Klav security”. Don’t put details in the first email; ask for a private channel.

Include:

  • the affected commit or version, and the component: core, server, CLI, klavd, extension, iOS, Mac or web;
  • the steps to reproduce, or a proof of concept;
  • the impact: which secret or which guarantee breaks, and for whom (another user, an agent, the server, another process on the device);
  • whether you want credit, and under which name.

Using AI tools to research is fine; sending what they produce without checking it is not. Only reports that a person has reproduced, with a concrete proof of concept against Klav’s code or a running instance, are accepted. Scanner output and unverified write-ups by a language model are closed without a detailed reply.

What to expect

Step Target
Acknowledgement 72 hours
First assessment and severity 7 days
Fix or mitigation plan Depends on severity; critical issues first
Coordinated disclosure When the fix is released, and no later than 90 days after the acknowledgement unless agreed otherwise

Fixed vulnerabilities are published as GitHub Security Advisories, requesting a CVE when one applies, and you are credited unless you’d rather stay anonymous. There is no paid bug bounty yet.

Scope

In scope: all of Klav’s code, the formats in its cryptography specification, and the hosted server at api.klav.dev and the site at klav.dev, under the testing rules below.

Out of scope (the threat model explains why):

  • malware running as root, or as your user with full control of the device;
  • attacks that need the victim to give away both their master password and their Secret Key;
  • denial of service by volume against any server;
  • servers that other people host;
  • reports that only say a header, a TLS setting or a dependency version is “best practice”, with no way to exploit it in Klav.

Testing rules

  • Use only accounts and vaults that you own. Prefer your own server to api.klav.dev.
  • No denial of service, spam, social engineering, phishing or physical attacks.
  • Don’t access, change or keep other people’s data. If you reach any by accident, stop, delete it and report it.
  • Give the maintainers the time in the table above before you disclose anything.

Research in good faith that follows the policy is covered by its safe harbor; SECURITY.md has its exact terms.

Supported versions

Until version 1.0, only the main branch of the source code and the latest published build of each app get security fixes.

Type to search…

↑↓ navigate↵ selectEsc close