Sidecall starts other AI command-line tools on your computer, so safety problems matter to us.

Please do not open a public issue. Open a private security advisory instead: GitHub repository page, then "Security", then "Report a vulnerability". Describe what you found, how to reproduce it, and what it lets someone do. You will get a reply as soon as possible.

• A way for a tool's answer to make Sidecall or Claude run commands, write files outside the run folder, or send data somewhere.
• A credential (key, token, password) reaching a tool or a run folder.
• A tool or its child processes surviving after the time limit or after Sidecall is stopped.
• A quota-spending call that happens without the notice line shown first.

Sidecall is not affiliated with Anthropic, OpenAI, Google, xAI or Moonshot AI; problems inside those tools should be reported to their makers.