Ai-Chef is a small personal tool. Only the latest released version receives fixes.

Please don't open a public issue for a security problem.

Use GitHub's private reporting: go to the Security tab on this repository and choose "Report a vulnerability". Include:

• what the issue is and which file/feature it affects
• steps to reproduce, or the relevant prompt/model output if it's a parsing bug
• your OS, Python version, and Ai-Chef version

I'll aim to acknowledge within a week and to ship a fix or a mitigation plan reasonably promptly.

In scope:

• Anything involving OPENAI_API_KEY — leaking, logging, or transmitting it.
• Path traversal in the data directory or in grocery-list export paths.
• Parsing untrusted model output in a way that executes code or writes outside the intended directory.
• Corrupting or losing user state through json_store.

Out of scope:

• Vulnerabilities in third-party dependencies with no reachable path through this project's own code — please report those upstream instead.
• Malicious input used to make a bad recipe. The model will believe anything.

A few things are deliberate and worth knowing before you report them:

• User data lives outside the repo, in the platform data directory (~/.local/share/ai-chef on Linux, ~/Library/Application Support/ai-chef on macOS). data_dir.get_data_dir() is the only thing that constructs a path.
• Writes are atomic. json_store writes to a temp file and renames, so an interrupted write can't corrupt state.
• The API key is read from the environment or .env and is never written to the data directory or included in exports.
• This is a local single-user CLI. It has no authentication, no network listener, and no multi-user support. Reports about missing auth on a local tool aren't vulnerabilities.