skills/mukul975/anthropic-cybersecurity-skills/building-c2-redirector-infrastructure/Gen Agent Trust Hub
building-c2-redirector-infrastructure
Warn
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends cloning and executing code from external GitHub repositories such as 'threatexpress/cs2modrewrite', 'mgeeky/RedWarden', and 'taherio/redi'.
- [COMMAND_EXECUTION]: The instructions include multiple 'sudo' commands for installing packages, enabling system modules, and modifying firewall settings.
- [DATA_EXFILTRATION]: The provided 'scripts/agent.py' script performs outbound HTTP requests to verify the redirector's status, which involves communicating with external domains.
- [REMOTE_CODE_EXECUTION]: The workflow involves cloning third-party repositories and immediately executing the contained Python scripts.
- [PROMPT_INJECTION]: The skill processes external input through command-line arguments in 'scripts/agent.py' which are interpolated into configurations. Evidence Chain: 1. Ingestion points: user-supplied --uri and --user-agent arguments in 'scripts/agent.py'; 2. Boundary markers: absent; 3. Capability inventory: network requests via requests.get and file writing via open(); 4. Sanitization: absent.
Audit Metadata