Open a private security advisory on GitHub (Security tab → Report a vulnerability). Please do not file public issues for anything that could lead to data exposure or remote code execution.

We will acknowledge receipt within 72 hours and aim to ship a fix within 30 days for high-severity reports.

In scope:

• The serve HTTP surface (auth bypass, request smuggling, path traversal, pickle deserialization, etc.).
• The exporter's secret scrub (false negatives are bugs, not features).
• Any code path that writes outside data/, adapters/, runs/, or hf_cache/ without an explicit user flag.

Out of scope:

• Third-party dependencies; please report upstream.
• Misconfigured deployments (default API key in production, exposed /metrics on the internet, etc.).
• Model output content. CodeClone does not run content moderation.

• The serve endpoint is intended for 127.0.0.1 or a private network. If you expose it publicly, you must front it with a reverse proxy that performs additional auth and rate limiting.
• The training data is your own commit history. CodeClone does not protect against you intentionally feeding it someone else's code; the safety filters are belt and suspenders, not a sandbox.
• The CLI is trusted with shell access (it shells out to git). Do not run codeclone export against untrusted repo URLs without your own review.