workvivo-fpx

Read your organization's Workvivo (Zoom Workvivo employee-experience platform, .workvivo.us/.com/.io) from a shell — the home feed, company news articles, posts and shoutouts with their comments, spaces and their feeds/news/events/content pages, linked documents, upcoming events, the people directory and profiles, your team, notifications, and full-text federated search — through your signed-in Workvivo browser tab with the fpx CLI. Read-only. Use when asked about Workvivo, the company intranet feed, internal news, a Workvivo space, colleagues' profiles, or what's happening at work.

Workvivo via fpx

Workvivo sits behind your company's SSO, so there is no server-side login: every call rides your already-signed-in Workvivo tab through the ContextMint Bridge extension and the fpx CLI (npm i -g @fetchproxy/cli).

One-time setup

sh

fpx profile add workvivo --domain workvivo.us      # the apex of YOUR host (.us / .com / .io)
fpx profile declare workvivo --cookie XSRF-TOKEN --allow-download
fpx get https://acme.workvivo.us/ -p workvivo >/dev/null   # prints a pair code on stderr

Approve the pair code in the ContextMint Bridge popup of the browser where you are signed in to Workvivo. If a second browser also runs the extension, every call prints a fresh pair code on stderr while stdout still carries the data — that is the other browser being turned away; do not approve it (fpx trust list shows the pin, fpx trust clear fpx-workvivo re-pairs).

Call pattern

Source the helper — it reads the session CSRF token once and adds the headers Workvivo's /api/* routes require (X-CSRF-Token from the page's <meta name="csrf_token"> + X-Requested-With: XMLHttpRequest; without them every call is 401 {"message":"Unauthenticated."}, GETs included):

sh

export WORKVIVO_HOST=acme.workvivo.us
source ~/.claude/skills/workvivo-fpx/scripts/wv.sh     # or wherever this skill lives
wv_get 'activity' | jq '.data[] | {type: .item_type, by: (.item.creator // .item.user // .item.giver).full_name, title: .item.title, text: (.item.short_text // .item.text)}'
wv_search 'benefits enrollment' article event | jq '.data[] | {type, title, url}'

• wv_get <path> — GET https://$WORKVIVO_HOST/api/<path>.
• wv_search <query> [type…] — POST federated-search (WV_PAGE=2 for more). It adds an entry to your own Workvivo search history; that is its only side effect.
• Return codes: 0 ok · 1 usage · 2 bridge down · 3 bot wall · 4 upstream HTTP error (body on stdout) · 5 not signed in to Workvivo in the paired browser.

Resolve first. Most detail endpoints take a numeric id: find it from a list or search result (url like news/7001, events/8001, comments/update/5001), then read the detail. Times are UTC (*_utc fields where present).

Never dump raw

Several responses carry fields that are not yours to relay — always project with jq: user (your own record) includes realtime/analytics tokens and org config; events include zoom_webinar_password; people records include phone numbers. The recipes in references/endpoints.md select named fields only.

Everything readable is in references/endpoints.md — live-verified paths, paging, and jq recipes for each.