ClawMuse runs AI agents that can read files, run shell commands and drive a browser on the user's own computer, so security reports are taken seriously and handled privately.

Do not open a public issue, discussion or pull request for a security problem.

Report it through GitHub's private vulnerability reporting:

1. Open the repository's Security tab.
2. Choose Report a vulnerability (direct link).
3. Describe the problem, the affected version (Settings → About), your OS, and the steps to reproduce it. A proof of concept helps.

Only the maintainers can see the report. We will:

• acknowledge it within 3 working days;
• confirm or rule out the issue and share our assessment within 10 working days;
• agree a disclosure date with you, normally once a fixed release is available and at most 90 days after the report;
• credit you in the published advisory unless you prefer not to be named.

In scope:

• the ClawMuse desktop app in this repository: the Electron main, preload and renderer processes, IPC, the secure store, deep links (clawmuse://), the installers and the update path;
• how ClawMuse installs, configures and connects to the local OpenClaw gateway, including the gateway token and device identity it generates.

Examples: escaping context isolation, a web page or chat message making the app run commands without an approval, reading another app's secrets, a key ending up in logs or in the app bundle.

Out of scope, and better reported elsewhere:

• vulnerabilities in OpenClaw itself; report those to that project;
• vulnerabilities in Electron or Chromium that ClawMuse does not make worse;
• actions an agent takes with permissions the user granted it (a bot with shell access can run shell commands; that is the documented design, see the README's Limits);
• unsigned-build warnings from Gatekeeper or SmartScreen.

Security fixes go into the latest release. Please reproduce on it before reporting.