Security

Your clusters are the crown jewels.

Klystiq sits between you and production, so it's built to be careful. Here's how it protects your clusters and credentials, and how to tell us if you find a problem.

How Klystiq protects you

Careful by design.

Runs on your computer

Klystiq talks to your API servers directly, with your own credentials. There's no account, no telemetry and no service in the middle that could see your clusters.

Your kubeconfig, untouched

Klystiq reads your kubeconfig and never writes it. The local terminal gets a small kubeconfig of its own that names a context, with no credentials in it.

Read-only means read-only

Mark a context read-only and every write is refused in the Rust command layer, not only hidden in the interface: applies, scaling, deletes, shells, GitOps actions and agent proposals.

No blind changes

Edits get a server-side dry run and a diff before anything is written, fields owned by other tools raise a conflict, and every action shows the kubectl or helm command it's equivalent to.

Secrets stay masked

Secret values are masked until you reveal one, compared by fingerprint, and never sent to an AI model or returned by the MCP server.

AI only when you ask

Nothing goes to an AI provider until you ask a question. Each question shows exactly what was sent, lookups are read-only, and API keys live in your system keychain.

Local services stay local

Port forwards and the MCP server listen on 127.0.0.1 only. The MCP server needs a random bearer token and refuses requests from web pages, and agents' changes need your approval.

Privileged actions are explicit

A Node shell is confirmed every time, and its Pod is labeled as Klystiq's and removed when you close it, or after 12 hours at most. Debug copies expire the same way.

A local audit log

Every change Klystiq makes to a cluster is recorded on your computer, in a file only you can read, with Secret values masked.

Signed, offline license keys

License keys are signed with Ed25519 and checked on your computer. Klystiq never phones home to check a license, so it works in air-gapped networks.

Verifiable downloads

Every release has SHA-256 checksums. Builds aren't code-signed yet; signed and notarized builds are on the way, and we'll say so here when they ship.

A quiet website

klystiq.com sets no cookies and runs no analytics, and is served with a strict Content-Security-Policy and HSTS. Payments are handled by Paddle; we never see card details.

Vulnerability disclosure

Reporting a security issue

If you think you've found a vulnerability in Klystiq, please email security@klystiq.com rather than posting it publicly. Please include:

  • what's affected, and the version of Klystiq and your operating system if it's the app;
  • the steps to reproduce it, and a proof of concept if you have one;
  • what an attacker could do with it;
  • how you'd like to be credited, if at all.

What you can expect from us

  • We'll acknowledge your report within 3 business days.
  • We'll give you our assessment within 10 business days, and keep you updated until it's fixed.
  • We'll fix confirmed issues as fast as their severity calls for, and tell you before we publish.
  • With your permission, we'll credit you in the release notes.

We don't run a paid bug bounty at the moment.

In scope

  • the Klystiq desktop app for macOS, Windows and Linux;
  • klystiq.com, licenses.klystiq.com (license keys) and downloads.klystiq.com (installers);
  • the license key format and how the app verifies it.

Out of scope

  • Paddle's checkout and billing pages: report those to Paddle directly;
  • tools Klystiq runs on your behalf, such as kubectl, helm, Trivy or your cloud CLI, and Kubernetes itself: report those to their maintainers;
  • denial of service, volumetric or rate-limit testing, spam and social engineering;
  • issues that need an attacker to already control your computer, user account or kubeconfig;
  • missing headers or best practices without a demonstrated impact.

Safe harbor

We won't pursue legal action against good-faith research that follows this policy: test only against your own installations and accounts, don't access, change or keep other people's data, don't degrade our services, and give us a reasonable time to fix an issue (90 days, or sooner once a fix ships) before you disclose it.