ux-skill is a deterministic, offline design-intelligence engine. It runs locally or in CI, makes no network calls, and handles no credentials or user data. The attack surface is small by design, but we still take reports seriously.

Fixes land on the latest 3.x release. Please upgrade before reporting an issue against an older version.

Do not open a public issue for a security vulnerability.

Use GitHub's private vulnerability reporting:

1. Go to the Security tab.
2. Click Report a vulnerability.
3. Describe the issue, affected version, and reproduction steps.

This keeps the report private until a fix ships. (If the "Report a vulnerability" button is not visible, private reporting has not yet been enabled in repository settings — see the note below.)

What to expect:

• An acknowledgement, typically within a few days (this is a small, community-maintained project — best effort, not a guaranteed SLA).
• A fix or mitigation plan for confirmed issues, released on the 3.x line.
• Credit in the release notes if you would like it.

In scope:

• The Python engine (engine/) — recommender, linter, synthesizer, evaluator.
• The CLI (bin/uxskill.mjs, engine/cli/).
• The MCP server (engine/mcp/server.py).

Out of scope:

• The static documentation site under docs/ (no server-side code).
• Third-party dependencies — report those upstream; we track them via Dependabot.
• Findings that require a malicious local environment the engine already trusts (e.g. a hostile PATH or a poisoned local data manifest the user installed).

• The engine is pure dict -> dict Python with no eval, exec, shell execution, or deserialization of untrusted input.
• The CLI spawns the Python engine via child_process.spawn with an argument array (no shell), so there is no shell-injection path.
• Code scanning (CodeQL) and dependency review run in CI on every push and pull request to main.