Kodrix runs code, downloads toolchains and handles GitHub and AI-provider credentials on people's phones, so security problems matter here. Thank you for helping keep users safe.

Only the latest release gets security fixes. Please check that you can reproduce the problem on the newest APK from Releases or on the main branch before reporting.

Do not open a public issue, discussion or pull request for a security problem. Public reports put users at risk before a fix is out.

Report it privately instead:

1. Go to the repository's Security tab → Report a vulnerability.
2. Include:
   • what the problem is and what an attacker could do with it,
   • the Kodrix version and Android version / device you tested on,
   • steps to reproduce, or a proof of concept,
   • any logs, with tokens and personal data removed.

Kodrix is run by its owner, @Zohaib8090, so these are targets rather than guarantees:

You'll get updates in the private advisory thread. Once a fix is released, the advisory is published and you're credited by name (or kept anonymous if you prefer).

Please give us a reasonable chance to fix the issue before sharing it anywhere else.

In scope:

• The Android app (androidApp/, shared/) and the desktop app (desktopApp/)
• The native code (androidApp/src/main/cpp/), including the exec shim kodrix_exec.c
• The agent server (agentServer/) and how the app launches it
• How runtimes and extensions are downloaded, verified and installed (the Termux package installer, the marketplace, the versions.json registry)
• Handling of GitHub OAuth tokens, AI-provider API keys and other stored credentials
• The GitHub Actions workflows in this repository

Examples of what we want to hear about:

• A way for a web page, project file, extension or downloaded package to run code or read files it shouldn't
• Credentials (GitHub token, API keys) being leaked, logged, or stored unencrypted
• Downloads that can be tampered with (missing or bypassable checksum checks, HTTP instead of HTTPS)
• Shell or command injection through file names, project names, git URLs or agent tool calls
• The local agent server or tunnel exposing something to other devices or the internet without the user choosing to

Out of scope:

• Code a user deliberately runs in their own terminal. Kodrix is an IDE, and running what you type is the point.
• Problems that need a rooted device, a device already compromised, or physical access to an unlocked phone
• Vulnerabilities in third-party packages (Termux packages, npm modules, language servers). Report those upstream. Tell us too if Kodrix makes them worse.
• Anything exposed by a port-forwarding tunnel (bore.pub) that the user opened on purpose
• Missing security headers on the project website, or automated scanner output with no demonstrated impact

We won't pursue or support legal action against anyone who researches and reports in good faith under this policy. That means:

• only testing on your own devices and accounts,
• not accessing, changing or deleting other people's data,
• not degrading the service for others,
• giving us time to fix the issue before disclosing it.

All code merged into Kodrix must follow these rules. Reviewers will ask for changes if a pull request breaks them.

1. No secrets in the repository. Never commit API keys, tokens, passwords, signing keystores (*.jks, *.keystore), local.properties, .env files or private keys. Use environment variables, Android's encrypted storage, or the agent server's apiKeyRef references (env:, file:), never literal keys.
2. Verify everything you download. Runtimes, packages and extensions must be fetched over HTTPS and checked against a SHA-256 hash (or a signed index) before they're installed or run. Never turn TLS verification off in new code.
3. Don't build shell commands from untrusted text. File names, project names, git URLs, package names and anything an AI model produced must be passed as separate arguments (ProcessBuilder with a list) or properly quoted, never pasted into a shell string.
4. Keep credentials out of logs and files. Don't log tokens, Authorization headers or full request bodies. Don't write credentials to .netrc, hosts.yml or other plain-text files. Pass them in memory (see git-credential-kodrix).
5. Local servers listen on 127.0.0.1 only. The agent server, language servers and anything else Kodrix starts must not listen on all interfaces. Exposing something to the network must be a clear choice by the user (like starting a tunnel).
6. Ask for permissions only when they're needed, with an explanation, and never more than the feature requires.
7. AI agent tools that change files, run commands or touch the network must go through the user-approval flow unless the user has explicitly turned on auto mode.
8. Native code (kodrix_exec.c, native-lib.cpp) must check buffer sizes and return values. Changes to the exec shim need a clear explanation of what they allow to run.
9. New dependencies must come from the official source (Maven Central, Google, npm) and be pinned to a version. Mention every new dependency in the pull request description.

If you're not sure whether a change is safe, say so in the pull request. Asking is always fine.