Skills
Navigate
skills/jd-opensource/joysafeter/pentest-network-internal/Gen Agent Trust Hub

pentest-network-internal

Fail

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The execute_command tool in references/tools.md allows the execution of arbitrary shell commands on the server environment.
  • [REMOTE_CODE_EXECUTION]: The execute_python_script tool in references/tools.md enables the execution of arbitrary Python code, and metasploit_run allows running exploitation modules.
  • [DYNAMIC_EXECUTION]: The install_python_package tool in references/tools.md allows the installation of arbitrary third-party libraries.
  • [DATA_EXFILTRATION]: The responder_credential_harvest tool in references/tools.md captures network credentials from the local network environment.
  • [CREDENTIALS_UNSAFE]: The workflows in references/workflows.md involve collecting hashes and credentials which are processed and stored by the skill.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from network targets which may influence agent behavior.
  • Ingestion points: Outputs from nmap_scan, httpx_probe, and responder in references/tools.md are returned to the agent context.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat tool outputs as untrusted data.
  • Capability inventory: The agent has access to execute_command, execute_python_script, modify_file, and metasploit_run.
  • Sanitization: There is no evidence of output validation or sanitization before the data is processed by the agent.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 28, 2026, 02:37 PM