wifi-pentest

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to run numerous wireless security utilities (aircrack-ng suite, reaver, hashcat). These commands are used for interface management, traffic sniffing, and packet injection.
  • [EXTERNAL_DOWNLOADS]: The evil-twin-attack-guide.md file directs the agent to clone the eaphammer tool from a non-trusted GitHub repository (https://github.com/s0lst1c3/eaphammer.git) and install it using python3 setup.py install. This execution of unverified external code is a significant risk vector.
  • [PRIVILEGE_ESCALATION]: The skill requires frequent use of sudo to perform low-level network operations, such as switching wireless cards to monitor mode (airmon-ng) or sending deauthentication frames (aireplay-ng). The instructions explicitly provide sudo-prefixed commands.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by ingesting and processing untrusted wireless network metadata (SSIDs, MAC addresses, probe requests) from the surrounding environment. Maliciously crafted SSIDs could potentially influence the behavior of the agent or the tools it manages.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 11:07 PM