as we can see here
https://unpkg.com/@pnp/[email protected]/dist/sp-clientsvc.es5.umd.js
and there
https://unpkg.com/@pnp/[email protected]/dist/sp-taxonomy.es5.umd.js
the global name is in fact pnp["sp-taxonomy"] and pnp["sp-clientsvc"] (we're currently suggesting pnp.sptaxonomy and pnp.spclientsvc but it doens't work, pnp.taxonomy doesn't seem to work either)
It contrasts with the other modules like https://unpkg.com/@pnp/[email protected]/dist/common.es5.umd.js (pnp.common) which work fine.

The global name syntax of SPFx doesn't seem to understand anything with brackets (unless there's a special syntax I'm not aware of).

Additionnally those packages won't make it to v2 https://pnp.github.io/pnpjs/documentation/v2-beta/transition-guide/#drop-client-svc-and-sp-taxonomy-libraries

I was thinking of adding a bypass for those two modules (so they don't show up in recommendations), thoughts before I proceed?