mcpField
T
tls-fingerprint-impersonation

tls-fingerprint-impersonation

github.com/uphiagoSecurity1.3k44 installsUpdated 8w ago
Repository

About tls-fingerprint-impersonation

A skill for using impit to send HTTP requests with browser-matched TLS fingerprints. It helps authorized testing and API probing when a target responds differently to non-browser TLS, without requiring a full browser.

Key features

  1. 01Browser TLS fingerprint profiles
  2. 02JA3/JA4 validation guidance
  3. 03Proxy and cookie configuration
  4. 04HTTP/3 and fallback options
  5. 05Python, Node.js, and Rust usage

Use cases

  1. 01Investigating blocks tied to TLS fingerprints
  2. 02Testing mobile API behavior with OkHttp profiles
  3. 03Comparing target responses across browser profiles

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/uphiago/recon-skills --skill tls-fingerprint-impersonation

After installing, type/tls-fingerprint-impersonationin Claude Code to confirm it loaded.

Details

From the project README

TLS Fingerprint Impersonation

Spoof TLS ClientHello parameters — cipher suites, key exchange groups, signature algorithms, and extension order — to match real browsers at the JA3/JA4 fingerprint level. Uses patched rustls to rebuild the TLS layer with browser-identical configurations. Bypasses TLS fingerprinting detection (Cloudflare, Akamai, F5) that flags non-browser TLS stacks. Supports 20 browser profiles including Chrome 100-142, Firefox 128-144, Safari iOS 18, and OkHttp 3-5 (Android).

When to Use

  • Target returns 403/blocked on curl/httpx even with correct User-Agent headers.
  • Cloudflare or Akamai is fingerprinting TLS ClientHello (JA3/JA4 mismatch with browser).
  • API probing requires mobile-app impersonation (OkHttp fingerprint for Android).
  • Need high-throughput HTTP requests that pass TLS fingerprint checks without running a full browser.
  • Target shows different behavior based on TLS fingerprint (mobile vs desktop endpoints).

Prerequisites

  • terminal with python3.
  • Python: pip install impit (wraps the Rust library via PyO3).
  • Or Node.js: npm install impit (native binding).
  • Or Rust: impit crate with patched dependencies in Cargo.toml.

Quick Detection

# Check if a target blocks non-browser TLS
curl --max-time 30 --connect-timeout 10 -sk https://target.com | head -1
# If 403, test with browser impersonation:
python3 -c "
from impit import Impit
impit = Impit.builder().with_fingerprint('chrome142').build()
r = impit.get('https://target.com').text
print(r[:200])
"

Procedure

Phase 1 — Browser Profile Selection

Choose the right fingerprint for your target:

ProfileUse caseKey differentiator
chrome142Modern desktopLatest Chrome, post-quantum KEX (X25519MLKEM768), GREASE
chrome100Legacy systemsOlder cipher suites, no GREASE in key exchange
firefox144Firefox desktopDifferent pseudo-header order, FFDHE groups, SHA-1 signatures
safari_ios18iOS mobile3DES ciphers, duplicate signature algorithm, no session tickets
okhttp4Android appsBoringSSL profile, no GREASE, no ECH, simpler cipher suites
chrome124Common defaultGood balance of modern compatibility and detection pass rate
from impit import Impit

# Chrome 142 (latest)
client = Impit.builder().with_fingerprint("chrome142").build()

# Firefox 144
client = Impit.builder().with_fingerprint("firefox144").build()

# Safari iOS 18 (mobile API endpoints)
client = Impit.builder().with_fingerprint("ios18").build()

# OkHttp 4 (Android app impersonation)
client = Impit.builder().with_fingerprint("okhttp4").build()
Phase 2 — Basic Request
from impit import Impit

impit = (
    Impit.builder()
    .with_fingerprint("chrome142")
    .with_ignore_tls_errors(True)  # for self-signed certs during recon
    .with_http3()                   # HTTP/3 support
    .with_fallback_to_vanilla(True) # retry without fingerprint if blocked
    .build()
)

response = impit.get("https://target.com")
print(response.status_code)
print(response.headers)
print(response.text()[:500])
Phase 3 — With Proxy
impit = (
    Impit.builder()
    .with_fingerprint("chrome142")
    .with_proxy("http://user:pass@residential-proxy:8080")
    .with_default_timeout(15_000)  # 15 seconds in milliseconds
    .build()
)

response = impit.get("https://target.com/api/endpoint")
Phase 4 — Custom Headers
# Custom headers are merged with fingerprint defaults
# Fingerprint headers have lower priority — custom headers win on conflict
response = impit.get(
    "https://target.com/api/v1",
    headers={
        "X-Forwarded-For": "[REDACTED_IP]",
        "Authorization": "Bearer token",
    }
)
Phase 5 — Cookie Session
from impit.cookie import Jar

impit = (
    Impit.builder()
    .with_fingerprint("chrome142")
    .with_cookie_store(Jar())  # persistent cookie jar
    .build()
)

# Login
impit.post("https://target.com/login", json={
    "username": "admin", "password": "admin"
})

# Authenticated request — cookies preserved automatically
response = impit.get("https://target.com/dashboard")
Phase 6 — Fingerprint Selection Logic

Choose based on target characteristics:

def select_fingerprint(target_url):
    """Auto-select browser fingerprint based on target."""
    if "mobile" in target_url or "api/v2" in target_url:
        return "ios18"
    elif "android" in target_url or "play.google" in target_url:
        return "okhttp4"
    elif target_url.startswith("https://"):
        return "chrome142"  # default for modern HTTPS
    return "chrome124"
Phase 7 — JA4 Hash Validation

Verify your TLS fingerprint is working correctly:

# Test against a site that returns JA4 hash in response headers
python3 -c "
from impit import Impit
impit = Impit.builder().with_fingerprint('chrome142').build()
r = impit.get('https://cloudflare.com/cdn-cgi/trace')
print(r.text())
# Look for JA4 hash in trace output or response headers
"

Browser Fingerprint Reference

TLS Configuration per Browser
FeatureChrome 142Firefox 144Safari iOS 18OkHttp 4
TLS versions1.3 + 1.21.3 + 1.21.3 + 1.21.3 + 1.2
GREASE cipher✅ (pos 1)❌✅❌
GREASE key exchange✅ (pos 1)❌✅❌
Post-quantum (MLKEM768)✅✅✅❌
FFDHE groups❌✅ (2048/3072)❌❌
SHA-1 signatures❌✅✅ (legacy)✅ (RSA only)
ECH GREASE✅✅❌❌
3DES ciphers❌❌✅❌
Certificate compressionBrotliZlib+Brotli+ZstdZlib❌
Delegated credentials❌✅❌❌
Session tickets✅✅❌✅
Duplicate signatures❌❌✅ (RsaPssRsaSha384)❌
HTTP/2 Settings per Browser
BrowserStream WindowConnection WindowPseudo-Header Order
Chrome6,291,45615,663,105:method :authority :scheme :path
Firefox131,07212,517,377:method :path :authority :scheme
Safari iOS2,097,15210,485,760:method :scheme :authority :path
OkHttp16,777,21616,777,216:method :path :authority :scheme
Multipart Boundary Format
BrowserFormatExample
Chrome----WebKitFormBoundary + 16 alphanumeric----WebKitFormBoundaryx8fH3kLm9pQr2sTv
Firefox----geckoformboundary + hex u64 values----geckoformboundary3fa8c10e5d6b2904
OkHttpUUID v4550e8400-e29b-41d4-a716-446655440000

Pitfalls

  • Not all sites use TLS fingerprinting. Test with curl first — if it works, TLS fingerprinting is not the blocker.
  • Fingerprint must match User-Agent. Using Chrome TLS with Firefox UA headers will be detected.
  • HTTP/1.1-only sites don't use HTTP/2 impersonation. The with_http3() flag only matters for sites that support it.
  • OkHttp 3 is TLS 1.2 only. Some modern servers reject TLS 1.2 connections.
  • TLS fingerprint caching means first request is slowest. CryptoProvider instances are cached per fingerprint — subsequent requests are fast.
  • Vanilla fallback may leak your real TLS fingerprint. Disable vanilla_fallback if stealth is critical.

Verification

  1. Confirm TLS fingerprint matches target browser using https://www.howsmyssl.com/ or Cloudflare trace endpoint.
  2. Check cf-ja4 response header when hitting Cloudflare-protected sites.
  3. Verify response status changes from 403 → 200 when using browser fingerprint vs vanilla curl.
  4. Test with multiple browser profiles to find the one that passes the target's detection.

Related Skills

  • http2-header-impersonation — HTTP/2 pseudo-header ordering and SETTINGS frame matching.
  • stealth-browser-launch — Full browser automation with C++ fingerprint patches for JS-heavy targets.
  • humanize-automation — Human-like interaction patterns for behavioral detection bypass.

Comments

Markdown supported