detecting-rdp-brute-force-attacks
Fail
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The 'Steps' section (items 1-5) explicitly directs the agent to perform 'Exploitation' and 'Post-exploitation' actions, such as attempting to exploit vulnerabilities and lateral movement, which are diametrically opposed to the skill's claimed purpose of RDP brute force detection. This creates a severe risk of the agent performing unauthorized offensive actions.
- [COMMAND_EXECUTION]: The skill instructs the use of
wevtutil epl Security C:\logs\security.evtx, a command that requires administrative privileges and interacts with sensitive system security logs. - [DATA_EXFILTRATION]: The skill targets Windows Security Event Logs (Event IDs 4624, 4625), which contain sensitive authentication data, including usernames, logon types, and source IP addresses.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by processing external .evtx files. Evidence: 1. Ingestion points: .evtx log files (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: High-privilege command execution via wevtutil and file access. 4. Sanitization: Absent; the skill parses log data using python-evtx and lxml without explicit validation.
Recommendations
- AI detected serious security threats
Audit Metadata