Pre-1.0, only the latest release receives security fixes.

Please do not report security vulnerabilities through public GitHub issues.

Instead, report them privately via GitHub's security advisory flow: Report a vulnerability (the "Security" tab of the repository, then "Report a vulnerability").

Include what you can of the following:

• A description of the issue and its impact.
• Steps to reproduce, or a proof-of-concept.
• Affected version(s) and platform.

You will get an acknowledgment as soon as possible, normally within a few days. Please allow time for a fix and release before any public disclosure.

anycli is a local library that drives agent CLIs running on the user's own machine. Reports we consider in scope include:

• Subprocess handling flaws: command or argument injection into the agent CLI launch, leaked or unreaped subprocesses that a caller cannot bound.
• Concurrency flaws that break the documented safety guarantees (e.g. the subprocess cap being bypassable).
• Any code path that logs, persists, or transmits credential material.

anycli never manages credentials by design. Authentication belongs entirely to the agent CLI on the user's machine: anycli does not read, store, transport, or proxy OAuth tokens or API keys — its only interaction with auth is detecting misconfiguration locally (for example, warning when ANTHROPIC_API_KEY is set and would override subscription auth). Any observed behavior to the contrary is a bug; please report it as a vulnerability.

Vulnerabilities in the agent CLIs themselves (Claude Code and its SDK) should be reported to their vendors, not here.