如何在 CefSharp OffScreen 控制台应用中识别 reCAPTCHA

# create a console app, then edit CefApp.csproj as shown below before adding packages
dotnet new console -n CefApp
cd CefApp

<PropertyGroup>
  <OutputType>Exe</OutputType>
  <!-- CefSharp's .NET packages only install into a Windows target (net8.0-windows works too) -->
  <TargetFramework>net10.0-windows</TargetFramework>
  <!-- without it, the Chromium binaries never reach bin and the app fails at startup -->
  <RuntimeIdentifier>win-x64</RuntimeIdentifier>
  <ImplicitUsings>enable</ImplicitUsings>
  <Nullable>enable</Nullable>
</PropertyGroup>

# the first restore downloads about 250 MB, most of it Chromium for win-x64
dotnet add package CefSharp.OffScreen.NETCore
dotnet add package CapSkip

// Program.cs. Cef.Initialize and Cef.Shutdown must run on the same thread, so the
// top-level code stays synchronous and blocks while the async work runs
using System.Text.Json;
using CapSkip;
using CefSharp;
using CefSharp.OffScreen;

const string SiteUrl = "https://example.com/login";

var settings = new CefSettings
{
    // CEF locks this folder: give each app, and each running copy, its own
    RootCachePath = Path.Combine(
        Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "CefApp"),
};
// InitializeAsync returns only when CEF is fully up
if (!Cef.InitializeAsync(settings).GetAwaiter().GetResult())
    throw new InvalidOperationException("CEF did not start: " + Cef.GetExitCode());
try
{
    RunAsync(SiteUrl).GetAwaiter().GetResult();
}
finally
{
    Cef.Shutdown();
}

static async Task RunAsync(string url)
{
    using var browser = new ChromiumWebBrowser(url);
    // the wait has no timeout of its own: a malformed URL never finishes loading
    var load = await browser.WaitForInitialLoadAsync().WaitAsync(TimeSpan.FromSeconds(30));
    Console.WriteLine($"Loaded {url}: HTTP {load.HttpStatusCode}");
}

// data-sitekey covers most pages. A grecaptcha.render() call leaves no attributes,
// but Google's anchor iframe carries the key in k= and the widget size in size=
static async Task<(string Sitekey, bool Invisible, string? Callback)?> FindRecaptchaAsync(
    ChromiumWebBrowser browser)
{
    var r = await browser.GetMainFrame().EvaluateScriptAsync("""
        (() => {
          const el = document.querySelector('.g-recaptcha[data-sitekey]');
          return el && [el.dataset.sitekey, el.dataset.size ?? '', el.dataset.callback ?? ''];
        })()
        """);
    var attrs = r.Result as List<object>;
    var anchor = browser.GetBrowser().GetAllFrames()
        .Select(f => f.Url)
        .FirstOrDefault(u => u.Contains("/recaptcha/") && u.Contains("/anchor?"));
    var query = anchor is null
        ? null
        : System.Web.HttpUtility.ParseQueryString(new Uri(anchor).Query);
    var sitekey = attrs?[0] as string ?? query?["k"];
    if (sitekey is null)
        return null;
    var invisible = (query?["size"] ?? attrs?[1] as string) == "invisible";
    var callback = attrs?[2] as string;
    return (sitekey, invisible, string.IsNullOrEmpty(callback) ? null : callback);
}

// CapSkip solves the widget for this sitekey and page and returns the token.
// CapSkip.TimeoutException is written in full: System has a TimeoutException too
static async Task<string?> SolveAsync(string sitekey, string pageUrl, bool invisible)
{
    var solver = new CapSkipClient(host: "127.0.0.1", port: 8080);
    var options = invisible ? new Dictionary<string, object?> { ["invisible"] = 1 } : null;
    try
    {
        var result = await solver.RecaptchaAsync(sitekey, pageUrl, options);
        return result.Code;
    }
    catch (NetworkException ex) { Console.WriteLine("Is CapSkip running? " + ex.Message); }
    catch (ApiException ex) { Console.WriteLine("CapSkip could not solve it: " + ex.Message); }
    catch (CapSkip.TimeoutException) { Console.WriteLine("No answer before the timeout"); }
    catch (CapSkipError ex) { Console.WriteLine("CapSkip error: " + ex.Message); }
    return null;
}

// fill every response field, call the widget's callback or submit its form,
// then wait for the page the form posts to and read it
static async Task<string> SubmitTokenAsync(
    ChromiumWebBrowser browser, string? callback, string token)
{
    // start waiting before the submit, or the navigation can finish first
    var navigation = browser.WaitForNavigationAsync(TimeSpan.FromSeconds(30));
    var r = await browser.GetMainFrame().EvaluateScriptAsync($$"""
        ((token, cb) => {
          const boxes = document.querySelectorAll('textarea[name="g-recaptcha-response"]');
          boxes.forEach(b => { b.value = token; });
          if (cb && typeof window[cb] === 'function') { window[cb](token); return 'called back'; }
          const form = boxes[0]?.form;
          if (!form) return 'no form found';
          if (!form.checkValidity()) return 'form has invalid or empty fields';
          form.requestSubmit();
          return 'submitted';
        })({{JsonSerializer.Serialize(token)}}, {{JsonSerializer.Serialize(callback)}})
        """);
    var how = r.Result as string ?? r.Message;
    if (how is not ("submitted" or "called back"))
        return how;
    try
    {
        var page = await navigation;
        var text = await browser.GetMainFrame().EvaluateScriptAsync("document.body.innerText");
        return $"{how}, HTTP {page.HttpStatusCode}: {text.Result}";
    }
    catch (System.TimeoutException) { return how + ", but the page did not navigate"; }
}

// Program.cs: load a page in a windowless CefSharp browser, solve its reCAPTCHA v2
// with CapSkip, submit the form and print the page that comes back
using System.Text.Json;
using CapSkip;
using CefSharp;
using CefSharp.OffScreen;

const string SiteUrl = "https://example.com/login";

var settings = new CefSettings
{
    // CEF locks this folder: give each app, and each running copy, its own
    RootCachePath = Path.Combine(
        Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "CefApp"),
};
// InitializeAsync returns only when CEF is fully up
if (!Cef.InitializeAsync(settings).GetAwaiter().GetResult())
    throw new InvalidOperationException("CEF did not start: " + Cef.GetExitCode());
try
{
    RunAsync(SiteUrl).GetAwaiter().GetResult();
}
finally
{
    Cef.Shutdown();
}

static async Task RunAsync(string url)
{
    using var browser = new ChromiumWebBrowser(url);
    // the wait has no timeout of its own: a malformed URL never finishes loading
    var load = await browser.WaitForInitialLoadAsync().WaitAsync(TimeSpan.FromSeconds(30));
    Console.WriteLine($"Loaded {url}: HTTP {load.HttpStatusCode}");
    if (await FindRecaptchaAsync(browser) is not { } widget)
    {
        Console.WriteLine("No reCAPTCHA on this page");
        return;
    }
    Console.WriteLine($"Sitekey {widget.Sitekey}, invisible: {widget.Invisible}");
    var token = await SolveAsync(widget.Sitekey, browser.Address, widget.Invisible);
    if (token is null)
        return;
    Console.WriteLine($"Token received, {token.Length} characters");
    Console.WriteLine(await SubmitTokenAsync(browser, widget.Callback, token));
}

// data-sitekey covers most pages. A grecaptcha.render() call leaves no attributes,
// but Google's anchor iframe carries the key in k= and the widget size in size=
static async Task<(string Sitekey, bool Invisible, string? Callback)?> FindRecaptchaAsync(
    ChromiumWebBrowser browser)
{
    var r = await browser.GetMainFrame().EvaluateScriptAsync("""
        (() => {
          const el = document.querySelector('.g-recaptcha[data-sitekey]');
          return el && [el.dataset.sitekey, el.dataset.size ?? '', el.dataset.callback ?? ''];
        })()
        """);
    var attrs = r.Result as List<object>;
    var anchor = browser.GetBrowser().GetAllFrames()
        .Select(f => f.Url)
        .FirstOrDefault(u => u.Contains("/recaptcha/") && u.Contains("/anchor?"));
    var query = anchor is null
        ? null
        : System.Web.HttpUtility.ParseQueryString(new Uri(anchor).Query);
    var sitekey = attrs?[0] as string ?? query?["k"];
    if (sitekey is null)
        return null;
    var invisible = (query?["size"] ?? attrs?[1] as string) == "invisible";
    var callback = attrs?[2] as string;
    return (sitekey, invisible, string.IsNullOrEmpty(callback) ? null : callback);
}

// CapSkip solves the widget for this sitekey and page and returns the token.
// CapSkip.TimeoutException is written in full: System has a TimeoutException too
static async Task<string?> SolveAsync(string sitekey, string pageUrl, bool invisible)
{
    var solver = new CapSkipClient(host: "127.0.0.1", port: 8080);
    var options = invisible ? new Dictionary<string, object?> { ["invisible"] = 1 } : null;
    try
    {
        var result = await solver.RecaptchaAsync(sitekey, pageUrl, options);
        return result.Code;
    }
    catch (NetworkException ex) { Console.WriteLine("Is CapSkip running? " + ex.Message); }
    catch (ApiException ex) { Console.WriteLine("CapSkip could not solve it: " + ex.Message); }
    catch (CapSkip.TimeoutException) { Console.WriteLine("No answer before the timeout"); }
    catch (CapSkipError ex) { Console.WriteLine("CapSkip error: " + ex.Message); }
    return null;
}

// fill every response field, call the widget's callback or submit its form,
// then wait for the page the form posts to and read it
static async Task<string> SubmitTokenAsync(
    ChromiumWebBrowser browser, string? callback, string token)
{
    // start waiting before the submit, or the navigation can finish first
    var navigation = browser.WaitForNavigationAsync(TimeSpan.FromSeconds(30));
    var r = await browser.GetMainFrame().EvaluateScriptAsync($$"""
        ((token, cb) => {
          const boxes = document.querySelectorAll('textarea[name="g-recaptcha-response"]');
          boxes.forEach(b => { b.value = token; });
          if (cb && typeof window[cb] === 'function') { window[cb](token); return 'called back'; }
          const form = boxes[0]?.form;
          if (!form) return 'no form found';
          if (!form.checkValidity()) return 'form has invalid or empty fields';
          form.requestSubmit();
          return 'submitted';
        })({{JsonSerializer.Serialize(token)}}, {{JsonSerializer.Serialize(callback)}})
        """);
    var how = r.Result as string ?? r.Message;
    if (how is not ("submitted" or "called back"))
        return how;
    try
    {
        var page = await navigation;
        var text = await browser.GetMainFrame().EvaluateScriptAsync("document.body.innerText");
        return $"{how}, HTTP {page.HttpStatusCode}: {text.Result}";
    }
    catch (System.TimeoutException) { return how + ", but the page did not navigate"; }
}

// several pages at once: one Cef.Initialize, one browser per page. Each browser gets
// its own RequestContext, or they all share one cookie jar and one session
static Task RunManyAsync(string[] urls) =>
    Task.WhenAll(urls.Select(async url =>
    {
        using var browser = new ChromiumWebBrowser(url, requestContext: new RequestContext());
        await browser.WaitForInitialLoadAsync().WaitAsync(TimeSpan.FromSeconds(30));
        if (await FindRecaptchaAsync(browser) is not { } widget)
            return;
        var token = await SolveAsync(widget.Sitekey, browser.Address, widget.Invisible);
        if (token is not null)
            Console.WriteLine(url + ": " + await SubmitTokenAsync(browser, widget.Callback, token));
    }));

// at the very end of Program.cs, then start CEF with
// Cef.InitializeAsync(settings, browserProcessHandler: new RelaunchHandler()).
// A second copy on the same RootCachePath still cannot start, but CEF no longer opens
// a window of its own in the running copy
class RelaunchHandler : CefSharp.Handler.BrowserProcessHandler
{
    protected override bool OnAlreadyRunningAppRelaunch(
        IReadOnlyDictionary<string, string> commandLine, string currentDirectory) => true;
}

// WinForms (CefSharp.WinForms.NETCore): CefSharp raises LoadingStateChanged, FrameLoadEnd
// and AddressChanged on its own thread, not the form's, so marshal control updates back
browser.LoadingStateChanged += (_, e) =>
{
    if (e.IsLoading)
        return;
    statusLabel.BeginInvoke(() => statusLabel.Text = "Loaded " + browser.Address);
};

// CapSkip in Server mode on another Windows machine: replace the client line in SolveAsync
var solver = new CapSkipClient(apiKey: "YOUR_API_KEY", host: "192.168.1.50", port: 8080);