phantom-frida

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). The URL points to an unvetted GitHub repository (TheQmaks/phantom-frida) that provides tooling to build modified Frida servers with explicit anti-detection features — a dual-use capability often abused to evade security controls and distribute executables from an unknown author, so it represents a high-risk download source.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This content automates building and deploying modified Frida servers with explicit anti-detection and binary-cleaning techniques (string obfuscation, SELinux label changes, disabling libc hooks, syscall/stack/frame modifications, randomized name/port) and provides deployment commands, effectively enabling stealthy remote code execution/backdoor access on Android and intentional evasion of detection.

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

• Third-party content exposure detected (high risk: 1.00). SKILL.md explicitly instructs the agent to clone and execute code from a public GitHub repository (git clone https://github.com/TheQmaks/phantom-frida.git) and to run scripts that fetch latest release data from the public GitHub API (scripts/get-latest-version.sh), meaning untrusted third‑party content is ingested at runtime and can materially change build/tooling behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly clones and relies on https://github.com/TheQmaks/phantom-frida.git at runtime and then runs its build scripts (e.g., python3 build.py), meaning fetched remote code will be executed and directly controls the agent's behavior.

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 using sudo (apt-get install, sudo chown) and guides building/deploying modified Frida binaries with system-level changes, which asks the agent to obtain elevated privileges and modify the machine state.