LLM-Dash ships from main; the latest release on the default branch is the only supported version. Security fixes land there and are published in the next tagged release.

Please report security issues privately — do not open a public issue for a suspected vulnerability.

Use GitHub's private reporting: go to the repository's Security tab → Report a vulnerability to open a private advisory. This keeps the report confidential until a fix ships. If private reporting is disabled, open a minimal public issue asking a maintainer to enable it — do not include exploit details.

Include what you can: affected version/commit, reproduction steps, impact, and any proof-of-concept. We aim to acknowledge within a few days and will keep you updated as we triage and fix.

LLM-Dash runs locally and binds to 127.0.0.1 by default. The most relevant security surfaces are:

• Credential handling — provider, Exa, and LLM Stats API keys are stored via environment, the Voidware credential broker, or the OS keychain, and are never written to the repo, logs, or API responses. Reports of keys leaking into any of those are in scope.
• --host 0.0.0.0 exposure — binding to a LAN/public interface widens the attack surface; treat the dashboard as trusted-network only.
• Update-run injection — the agent update contract (skill/SKILL.md) fetches and parses external content. Issues that let researched content execute or corrupt local state are in scope.

Out of scope: vulnerabilities in third-party dependencies (report those upstream), and risks that require an already-compromised local machine.