android-pentesting

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill package contains multiple explicit offensive tools and scripts (SSL-pinning/root/biometric bypasses, Frida hooks that dump SharedPreferences/KeyStore and log crypto ops, mitmproxy scripts that exfiltrate/modify API traffic and force premium access, anti-Frida hiding) that enable credential theft, data exfiltration, and stealthy remote manipulation — high-risk functionality that can be abused as backdoors or for unauthorized access.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The skill ingests outsider-authored free text via runtime traffic interception: it starts mitmproxy/mitmdump and prints/logs request/response bodies (e.g., modify_requests.py decodes flow.request.content), where those bodies originate from external parties’ network traffic and are fed into the agent’s LLM context through captured console/log output.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs bypassing security controls (SSL pinning, root detection, biometric and anti-Frida bypasses) and performing privileged device modifications (adb root, remount, push certs into /system/etc/security/cacerts), which alter system state and require elevated privileges.