a1073097082/dsh-model-search

Adds searchable filtering to the model selector by provider, model name, and model ID; spaces act as wildcards.

Adds search to the DeepSeek Harness model selector. You open the selector and filter by provider name, model name or model ID; matching is case-insensitive and spaces act as wildcards, so GPT 6 matches both GPT-5.6 and GPT-6. Because it is a client-side overlay it never touches model routing or stored credentials.

UI Enhancements ★ 1 updated 2026-08-15
View on GitHub ↗

Install

dsh plugin --profile web add github:a1073097082/dsh-model-search

README Install block quoted verbatim (the README lists the npm-market form dsh plugin --profile web add dsh-model-search first and the GitHub form second; it also says to refresh the Harness page after installation). Registry check 2026-09-16: the npm name dsh-model-search resolves at 0.1.1 but its registry repository field points at github.com/HeathHe/dsh-model-search — a different repository — so no npm ownership is asserted and this page uses the README's own GitHub form.

Compatibility

README: a client-only overlay for the Web profile's model selector; it does not change model routing or credentials. It needs nothing beyond a Web profile whose model selector is present.

Details

Recent updates

The README documents the search behaviour rather than a release-by-release table.

FAQ

How do I install it?
The README's GitHub form is dsh plugin --profile web add github:a1073097082/dsh-model-search; a marketplace/npm form is also listed. Refresh the Harness page afterwards.
How does the search matching work?
Per the README you search by provider, model name or model ID; matching is case-insensitive and a space is treated as a wildcard, so GPT 6 matches GPT-5.6 and GPT-6.
Does it change how models are routed?
No — the README says it is a client-only overlay that does not change model routing or credentials.

More plugins in UI Enhancements

Browse more in UI Enhancements

Guides for UI Enhancements plugins