Security & Trust
Kanban Pro is a local-first desktop app. Your boards live as plain Markdown files on your own machine — there is no Kanban Pro account, server, or cloud database, and your project data never passes through us. This page explains, in plain terms, how we protect your data and your device, and how the embedded terminal and AI-agent features are built to stay safe.
We are a small independent team, so we keep this honest and specific: everything below describes how the shipping app actually behaves. We would rather be trusted than impressive.
Read the full, technical Security Policy on GitHub →
What Kanban Pro does to protect you
| Area | What we do |
|---|---|
| Your data | Boards are plain Markdown and YAML on your disk. No accounts, no servers, no cloud database. Syncing is your choice, through your own iCloud, Dropbox, OneDrive, or Git. |
| Privacy | No telemetry, no analytics, no ads, no tracking. Nothing about your projects is sent to us. |
| Network | Offline by default. The only network activity in a release build is the update check on GitHub and the links you click. A strict Content Security Policy stops the app from making other network requests. |
| App isolation | The interface runs in a hardened, sandboxed shell — context isolation on, Node integration off, OS-level sandbox on. It cannot reach your files or run programs except through a small, explicit, audited bridge. |
| The terminal | The system terminal is owned by the trusted core of the app, not the sandboxed interface. It runs with your normal user permissions, exactly like the Terminal or PowerShell you already use. |
| AI and keys | Kanban Pro never calls an AI service and never stores your API keys. AI agents are separate tools that you run; their credentials stay in your operating system, not in Kanban Pro. |
| Distribution | macOS builds are signed and notarised with an Apple Developer ID and run under the Apple App Sandbox. Updates are verified by your operating system before they install. |
How to think about trust
A tool that can open a terminal and work alongside AI agents is powerful, so it helps to know exactly where the boundaries sit.
- You and your computer are in charge. Kanban Pro adds guardrails, but you decide which folders to open and which commands to run. A terminal that can run your commands is doing precisely what a terminal is for.
- Opening a board never runs anything on its own. Kanban Pro reads and displays files; it does not execute programs from a folder on your behalf. Commands run only when you launch the terminal or an agent and choose to run them.
- Treat boards from other people like any other download. A board you received, cloned, or synced from someone else is untrusted content until you have looked it over.
We are open about what we have not built yet: Kanban Pro does not force a "trust this folder?" prompt, does not add its own approval pop-up in front of AI agents, and does not store credentials in the OS keychain. The checklist below covers those gaps on your side today — and they are firmly on our minds for the future.
Staying safe — a short checklist
- Before you point the terminal or an AI agent at a board someone else made, skim its instruction files (such as
CLAUDE.md) and its tickets for anything that reads like a command. - Keep your AI tools in their normal "ask me first" mode. Save any auto-approve mode for throwaway, isolated environments you would not mind losing.
- Do not paste real API keys into tickets or notes. Hand them to your agent through operating-system environment variables instead.
- Keep Kanban Pro up to date, and only install it from our official Releases page.
Found a security issue?
We welcome private, responsible disclosure and will work with you on it.
- Preferred: open a private report from the Security tab of our GitHub repository, using Report a vulnerability.
- Email: contact.goodguyapps@gmail.com
We aim to acknowledge every report within 5 business days, and we appreciate coordinated disclosure. Please do not post security issues publicly until we have had a chance to fix them.
Read the full Security Policy on GitHub → · See our Privacy Policy →