Is LI Prospect Finder safe?

Low risk

LIPF is low risk. When LI Prospect Finder can't parse certain LinkedIn search-result pages locally, it posts the page HTML to li-prospect-finder.com. The posted JSON's html field carries the full document, including rendering data used to extract records.…

info@lipf.storev4.0.4Extension ID mdhjhplkmldfjgoegbdmallcdhlecfcnChrome Web Store
20Risk
AI-assisted analysis.

LIPF findings

LinkedIn search HTML sent to parsing endpoints

MediumunwantedCWE-359Dynamic sandbox

When LI Prospect Finder can't parse certain LinkedIn search-result pages locally, it posts the page HTML to li-prospect-finder.com.

What actually happens

You did this

You run a LinkedIn company or people search through the extension.

The fallback paths appear when LinkedIn pages use the newer rendering data or when local people parsers return no results.

The extension did this

The extension posts the LinkedIn search-result HTML to its vendor server for parsing.

The request body is JSON with an html field containing the page document.

Data placed in the parsing request

  • LinkedIn page HTML
    Illustrative HTML: <html><script>window.__como_rehydration__ = {"data":{"searchDashClustersByAll":{"metadata":{"primaryResultType":"COMPANIES"}}}}</script></html>

    This can include the search-result markup and embedded LinkedIn rendering data for the profiles or companies shown on the page.

  • Request field name
    html

    The server receives the page document under a single html field, which is enough to reconstruct the search results being parsed.

Captured request

POSThttps://li-prospect-finder.com/extension/api/companies/parse
Headers
Accept
application/json
Ext-Version
3.2.12
Content-Type
application/json

Captured request

POSThttps://li-prospect-finder.com/extension/api/peoples/parse
Headers
Accept
application/json
Ext-Version
3.2.12
Content-Type
application/json

The code that does this

Fallbacks that send LinkedIn search-result HTML

Readable version

Company search fallback in readable form

deobfuscated/js/li/search/searchCompanies.js
async function getCompaniesFromDefaultSearch() {  let e = WEB_REQUEST_LI_API_URLS.COMPANIES_SEARCH;  !0 === tab.url?.toLowerCase().includes("schools") && (e = WEB_REQUEST_LI_API_URLS.SCHOOLS_SEARCH);  var a = await getLocalStorageValue(e);  a && a[tab.id] && (searchUrl = a[tab.id], csrfToken) && (a = {    "csrf-token": this.csrfToken,    "x-restli-protocol-version": "2.0.0"  }, 0 < searchUrl.indexOf(WEB_REQUEST_LI_API_URLS.NEW_API) && (a.accept = "application/vnd.linkedin.normalized+json+2.1"), $.ajaxSetup({    headers: a,    global: !1,    type: "GET"  }));  let t = !1,    s = "";  tab.url?.toLowerCase().includes("companies") && (s = await $.get(tab.url) || "", t = s.includes("window.__como_rehydration__")), t ? companies = await apiCompaniesInfoFromHtml({    html: s  }) || [] : (a = await $.get(searchUrl), (companies = liSearchCompaniesParser.handle(a)) && 0 !== companies.length || (companies = liSearchCompaniesParser.getCompanies(a))), renderCompaniesList()}

People parser fallback in readable form

deobfuscated/js/li/search/searchProspectsParser.js
async getPeopleListV4(e) {  return {    people: (await apiProspectsInfoFromHtml({      html: e    }) || []).map(e => ({      ...e,      source: "linkedIn",      searchLink: `https://www.linkedin.com/voyager/api/identity/dash/profiles?q=memberIdentity&memberIdentity=${e.source_id_2}&decorationId=com.linkedin.voyager.dash.deco.identity.profile.FullProfileWithEntities-26`    })),    hasUnadded: !1  }}

Automated company search pagination in readable form

deobfuscated/js/li/autoSearch/autoSearchCompanyTask.js
async goToNextSearchPage() {  let t = !1;  if (!this.isFinished && !this.isPaused && this.isStarted) {    var s;    if (this.currentCompanies = [], (!0 === this.urlStart?.toLowerCase().includes("results/companies") || !0 === this.urlStart?.toLowerCase().includes("results/schools") || 0 < this.urlCurrent.indexOf(WEB_REQUEST_LI_API_URLS.NEW_API)) && (t = !0), +this.pageCurrent != +this.pageStart) return !(+this.pageCurrent <= +this.pageFinish) || (this.errorInternetCount = 0, this.urlCurrent = this.getCurrentSearchUrl(this.urlStart), 0 < this.urlCurrent.indexOf(WEB_REQUEST_LI_API_URLS.NEW_API) && $.ajaxSetup({      headers: {        accept: "application/vnd.linkedin.normalized+json+2.1"      },      global: !1,      type: "GET"    }), "string" == typeof(s = await $.get(this.urlCurrent)) && s.includes("window.__como_rehydration__") ? this.companies = await apiCompaniesInfoFromHtml({      html: s    }) : this.companies = liSearchCompaniesParser.handle(s), this.companies && 0 !== this.companies.length || (this.list.length ? this.companies = this.list : this.companies = t ? liSearchCompaniesParser.getCompanies(s) : getCompaniesSN(s), this.checkIsPageAbsentError(s, t)), this.isPageAbsentError) ? void await this.finishPageHandler() : (await taskManager.setUnfinishedTask(getDeepCopyOfObject({      ...this,      currentCompanies: []    })), void this.sendCompanyBlock());    this.list.length ? this.companies = this.list : this.companies = localStorage.companies ? JSON.parse(localStorage.companies) : [], this.needsSendCompaniesCount = this.companies.length + (this.pageFinish - this.pageStart) * this.companyCountOnPage, await this.sendCompanyBlock()  }}

Automated people search pagination in readable form

deobfuscated/js/li/autoSearch/autoSearchTask.js
async goToNextSearchPage() {  if (!this.isFinished && !this.isPaused && this.isStarted)    if (+this.pageCurrent > +this.pageFinish) this.finish && this.finish();    else {      this.errorInternetCount = 0, this.urlCurrent = this.getCurrentSearchUrl(this.urlStart);      var t = this;      if (this.list) i();      else {        let e = this.urlCurrent;        var s = {            "csrf-token": this.csrfToken,            "x-restli-protocol-version": "2.0.0"          },          s = (0 < e.indexOf(WEB_REQUEST_LI_API_URLS.NEW_API) && (s.accept = "application/vnd.linkedin.normalized+json+2.1"), this.requestMethod === REQUEST_METHODS.POST && (e = this.urlCurrent.split("?")[0], s["Content-Type"] = "application/x-www-form-urlencoded", s["x-http-method-override"] = "GET"), {            method: this.requestMethod,            headers: s          });        this.requestMethod === REQUEST_METHODS.POST && (s.body = this.urlCurrent.split("?")[1]);        try {          i(await (await fetch(e, s)).text())        } catch (e) {          0 === e.status ? (t.error = !0, t.errorCode = 0, t.errorOnPageGlobal()) : t.errorOnPage()        }      }    } async function i(e) {    !e && t.list ? (t.people = t.list, t.list = void 0) : (t.people = t.getPeopleList(e), t.people && t.people?.length || t.isSalesNavInt || (t.people = (await (new searchProspectsParser).getPeopleListV4(e))?.people)), t.getTrackingInfo && (t.trackingInfo = t.getTrackingInfo(e)), t.people && 0 < t.people.length ? (e = await t.checkPrevAddedPeople(t.people) || !1, t.needRecheckPrevAdded = e, await t.checkRecentlyUpdatedPeoples(t.people), t.itemsParsed = 0, t.waitAndGetPerson(), await reRenderTask(t)) : (t.itemsParsed = 0, t.waitAndGetPerson(), reRenderTask(t)), await reRenderTask(t)  }}

The code that does this

The API wrapper builds POST requests to the parsing endpoints

Readable version

Endpoint constants and POST helper in readable form

deobfuscated/js/apiMethods.js
let API_HOST = APP_HOST + "/extension/api",  USER_BALANCE = "/user/balance",  PEOPLE_CREATE = "/peoples/create",  PEOPLE_CONTACTS = "/peoples/contacts",  PEOPLE_RECENTLY_UPDATED = "/peoples/recently-updated",  COMPANIES_CREATE = "/companies/create",  LISTS_BY_USER = "/lists/get-by-user-id",  LISTS_BY_PROSPECTS = "/lists/get-by-peoples-ids",  LISTS_BY_COMPANIES = "/lists/get-by-companies-ids",  NEWS_GET = "/news/get-last",  LI_HTML = "/linkedin/collect",  COMPANIES_PARSE = "/companies/parse",  PROSPECTS_PARSE = "/peoples/parse";function post(e, a) {  return fetch(e, {    method: "POST",    body: JSON.stringify(a),    headers: {      "Content-Language": getUILanguage(),      "Ext-Version": chrome.runtime.getManifest().version,      "Content-Type": "application/json",      Accept: "application/json"    }  })}

Parsing endpoint wrappers in readable form

deobfuscated/js/apiMethods.js
async function apiCompaniesInfoFromHtml(e) {  e = await post(API_HOST + COMPANIES_PARSE, e);  return e.ok && (await e.json())?.companies || []}async function apiProspectsInfoFromHtml(e) {  e = await post(API_HOST + PROSPECTS_PARSE, e);  return e.ok && (await e.json())?.peoples || []}

Server that receives the parsing requests

    • li-prospect-finder.com

    LI Prospect Finder vendor host that receives LinkedIn search-result HTML at /extension/api/companies/parse and /extension/api/peoples/parse.

LinkedIn Prospect Finder records LinkedIn CSRF tokens and search request bodies

MediumunwantedCWE-200Static analysis

Code analysis shows the extension reads the CSRF token header and Sales Navigator search request and response data on LinkedIn pages and keeps it for its own requests.

What actually happens

You did this

You browse LinkedIn or run a Sales Navigator people or company search.

The extension did this

The extension reads the CSRF token header and the search request body and response.

It keeps them so it can build its own LinkedIn requests.

The code that does this

Listener that reads the CSRF token and POST body

Readable version

Readable names are illustrative.

service-worker.js
chrome.webRequest.onBeforeSendHeaders.addListener(({tabId, url, method, requestHeaders}) => {  if (!isLinkedInApi(url)) return;  const csrf = requestHeaders?.find(h => h.name.toLowerCase() === 'csrf-token')?.value;  const kind = classify(url);  if (csrf) saveCsrf(tabId, csrf);  if (kind && method === 'GET') handleCapturedRequest(tabId, kind, {url, method});}, {urls: ['https://*.linkedin.com/*']}, ['requestHeaders']);chrome.webRequest.onBeforeRequest.addListener(({tabId, url, method, requestBody}) => {  const kind = classify(url);  if (!kind || method !== 'POST') return;  const bytes = requestBody?.raw?.[0]?.bytes;  const body = bytes ? new TextDecoder('utf-8').decode(bytes) : '';  if (body) handleCapturedRequest(tabId, kind, {url: withBody(url, method, body), method});}, {urls: ['https://*.linkedin.com/*']}, ['requestBody']);

The code that does this

Page hook that relays Sales search responses

Readable version

Matches /sales-api/salesApi(People|Lead|Account|Company)Search; names are illustrative.

sales-search-page-hook.js
XMLHttpRequest.prototype.send = function(...args) {  const meta = tracked.get(this);  if (meta) {    const info = {...meta, body: bodyToString(args[0])};    this.addEventListener('load', () => {      if (isOk(this.status)) readResponse(this).then(text => relay(info, text));    });  }  return origSend.apply(this, args);};

Data the extension reads on LinkedIn pages

  • LinkedIn CSRF token
    ajax:4821937650123847561

    A token that lets requests act as your signed-in session.

  • Search request body
    {"keywords":"head of finance","start":0,"count":25}

    The filters you used in a people or company search.

  • Search results
    {"elements":[{"fullName":"Jane Doe"}]}

    The result list LinkedIn returned for that search.

Observation

Static analysis finding. This behaviour was identified by reading the shipped extension code and has not yet been reproduced in a live run. The trigger conditions and the exact data sent are read from the code, not from an observed capture.

Prospect Finder sends LinkedIn page HTML to li-prospect-finder.com after load

MediumunwantedCWE-359Static analysis

Code analysis shows the extension reads LinkedIn page HTML about 5 seconds after load and sends it to li-prospect-finder.com.

What actually happens

You did this

You open a LinkedIn profile, company or search page.

The extension did this

About 5 seconds later the extension sends the page HTML and URL to li-prospect-finder.com.

No click is needed on each page.

Fields in the upload request

  • Page address
    https://www.linkedin.com/in/jane-doe-1a2b3c/

    Tells the vendor which LinkedIn page you viewed.

  • Page HTML
    <div class="scaffold-layout">...</div>

    The full rendered page body, including names and job titles shown.

  • Page language
    en

    The language of the page.

Where the page HTML goes

    • li-prospect-finder.com

    Vendor backend; receives page HTML at /extension/api/linkedin/collect.

Version caveat

The 4.0.3 files on disk do not contain the page-HTML upload described by this claim. It was seen in the analysed code and not captured at runtime.

Observation

Static analysis finding. This behaviour was identified by reading the shipped extension code and has not yet been reproduced in a live run. The trigger conditions and the exact data sent are read from the code, not from an observed capture.

LIPF version history

  1. 3.2.6

    Mar 4, 2026

  2. +6 more

  3. 4.0.2

    Oct 19, 2026

  4. 4.0.3

    Oct 23, 2026

  5. 4.0.4

    Nov 20, 2026

Where LIPF sends data

  • fonts.gstatic.com/s/notosans/…4 calls

    From a script it injects into websites · 319 other extensions

  • li-prospect-finder.com/assets/domain.json3 calls

    From its background script · only this extension

  • li-prospect-finder.com/extension/api/…3 calls

    From its background script · only this extension

  • li-prospect-finder.com/knowledgebase/ext/…3 calls

    From its background script · only this extension

  • app.snov.io/api/checkAuth2 calls

    From a script it injects into websites · only this extension

  • www.google-analytics.com/collect1 call

    From its background script · 238 other extensions

Chrome Web Store listing

Versions
4.0.4 latest of 10 seen since March 4, 2026
Store rating
4.9 from 110 ratings
Last updated
November 20, 2026
Size
559 KiB
Updated November 20, 2026