apt-emulation

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This content is a high-risk, dual-use red team playbook that explicitly instructs on building C2 infrastructure, deploying remote agents, credential theft (LSASS dumps, mimikatz), persistence (registry, scheduled tasks), process injection/hollowing, lateral movement (Pass-the-Hash, WinRM), and exfiltration over C2—features that constitute deliberate malicious behavior and backdoor capabilities usable for unauthorized compromise and data theft.

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow explicitly instructs fetching and using public threat-intel and code from external sites (e.g., "访问 attack-evals.mitre-engenuity.org" and Atomic Red Team GitHub/raw URLs shown in references/emulation-frameworks.md) which the agent would read and use to choose TTPs and execution steps, exposing it to untrusted third‑party content.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill contains runtime commands that fetch and execute remote code—e.g., "git clone https://github.com/mitre/caldera.git --recursive" (followed by pip3 install -r requirements.txt and python3 server.py) and "git clone https://github.com/redcanaryco/atomic-red-team.git"—so these external repos are pulled at runtime and result in executing remote code required for the described operations.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The prompt explicitly instructs offensive actions that change system state—persistence (registry Run keys, scheduled tasks), privilege escalation/UAC bypass, LSASS dumping, process hollowing and C2/tool setup—i.e. techniques that would compromise or modify the host and require elevated privileges.