What is collected
Standard server logs may record requests for operational purposes such as abuse prevention and capacity planning. No account is created when you browse, and search happens entirely in your browser against an index shipped with the page. Nothing beyond those logs is collected unless you allow analytics.
What is not collected
- No account, email or profile, because there is nothing to sign in to.
- No tracking of which plugins you view or copy while analytics is off — and analytics is off until you allow it.
- No data from your local DeepSeek Harness installation. This site is a website, not a plugin.
- No advertising or cross-site tracking. Those signals are denied for every visitor and cannot be granted.
Analytics
Analytics is optional here and is off until you allow it. If you allow it, the site loads Google Analytics, which records the page you viewed, the site that referred you, a coarse location derived from your IP address, and a random identifier stored in a first-party cookie. With the same permission it also records when an install command is copied — the listing, the page language and the verification level, never an identity — which is how the directory measures whether a listing leads to an install decision. With the same permission it records when a README badge snippet is copied — the listing and the badge variant — which measures whether authors share their listings. Nothing else is collected.
Google's advertising storage, ad personalisation and Signals features are denied for every visitor, before consent and after it. You can change or withdraw your decision at any time on theanalytics settings page, which is linked from the footer of every page. Declining costs you nothing: every page, the catalog and the search index work identically either way.
Third party requests
Plugin listings show an author's initials rendered locally, so no avatar image is requested from GitHub or anywhere else. Links to repositories take you to GitHub. Beyond the optional Google Analytics script described above, no advertising networks or third-party tracking scripts are loaded.
When this site refreshes repository metadata, it reads public GitHub API endpoints. Those requests leave your browser only if you follow a link out; they are not proxied through a first-party tracker on every page view.
Repository metadata
Listings describe public repositories: names, descriptions, licenses, languages, star counts and update dates. That information is already public. If you maintain a repository listed here and want a description corrected, use the contact page.
Your messages
Email to the addresses on the contact page is used only to answer the message and, when relevant, to update a listing. We do not sell contact addresses and do not add submitters to a marketing list. If you ask for a removal of your listing, we keep the minimum correspondence needed to confirm authorship and complete the removal.
Changes
If this policy changes, the updated text appears on this page with the date below. Continued use of the site after a change means the new policy applies.
Last updated 2026-09-16.