pentester-mcp-security-tools

Warn

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The installation instructions require cloning a repository from an unverified source (https://github.com/halilkirazkaya/pentester-mcp.git).
  • [COMMAND_EXECUTION]: The skill makes extensive use of subprocess.run to execute high-risk security tools such as nmap, sqlmap, ffuf, and hydra. These commands are executed based on AI-generated parameters, which can lead to unauthorized network activity if the AI is misdirected.
  • [REMOTE_CODE_EXECUTION]: The server implementation in server.py uses importlib.import_module to dynamically load tool modules at runtime based on a YAML configuration file. This dynamic loading mechanism could be exploited if the configuration or the tool files are tampered with.
  • [DATA_EXFILTRATION]: The skill includes wrappers for tools specifically designed for sensitive data extraction, such as secretsdump.py from the Impacket suite, which is used to harvest credentials from domain controllers.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes output from external security tools (such as web fuzzers or port scanners) and feeds this data directly back into the AI context without explicit boundary markers or sanitization, potentially allowing a target system to influence the AI's behavior via malicious tool responses.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 12, 2026, 09:38 PM
Security Audit — agent-trust-hub — pentester-mcp-security-tools