Skip to content

Security: AlexShang1992/dsh-offpeak

Security

SECURITY.md

Security policy

Supported versions

Fixes land on the latest release only. Upgrade before reporting a problem you found on an older version.

Reporting a vulnerability

Report privately through GitHub's private vulnerability reporting: Security → Report a vulnerability. Please do not open a public issue for a security problem.

Include the affected version, your OS and Node.js version, the dsh version, and a minimal reproduction. You will get an acknowledgement within a week, and credit in the release notes unless you would rather stay anonymous.

What this plugin does

Installing any plugin runs third-party code with your own permissions. This one:

  • Writes no files. Its only durable state is its own offpeak settings section, persisted by the harness's settings provider.
  • Reads nothing but that section.
  • Makes no network calls, and collects no analytics, crash reports, or usage data.
  • Registers no tools and no commands, so it can neither be invoked by a model nor put anything into a model request.

Anything else — a file write, an outbound request, any data collection — is a security bug worth reporting under this policy.

Threat model

  • Untrusted settings values. Every value crossing the Remote is validated against the zod codecs in src/contract.ts in both directions; a section the schema rejects keeps the last good value rather than reaching the UI.
  • Untrusted text in the UI. The plugin renders only its own translated strings and numbers formatted from settings, as React text nodes; it uses no dangerouslySetInnerHTML and builds no HTML from strings.
  • The prices are not authoritative. They are what a user typed. Never treat the pill as a billing record or a spend control.
  • Dependencies. Runtime dependencies are zod and @deepseek-ai/schemastery, pinned by pnpm-lock.yaml. Review Dependabot updates before merging them.

Disclosure

Please give a reasonable window — 90 days by default — before publishing, and do not exploit an issue beyond what demonstrating it requires.

There aren't any published security advisories