skills/mukul975/anthropic-cybersecurity-skills/conducting-wireless-network-penetration-test/Gen Agent Trust Hub
conducting-wireless-network-penetration-test
Warn
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/agent.pyuses thesubprocess.run()function to execute several system utilities, includingiwconfig,airodump-ng,aircrack-ng, andwash, for capturing and cracking wireless traffic. - [PRIVILEGE_ESCALATION]: The documentation in
SKILL.mdandreferences/api-reference.mdrecommends usingsudowith theairmon-ngcommand, which allows the execution of tools with root-level access to the system. - [INDIRECT_PROMPT_INJECTION]: The skill captures and logs environmental wireless data, which creates a potential surface for prompt injection attacks if an agent consumes the generated reports.
- Ingestion points: The
scan_access_pointsanddetect_client_probesfunctions inscripts/agent.pycapture SSID strings and probe requests from the air using Scapy's sniffing capabilities. - Boundary markers: The captured SSID data is not wrapped in protective delimiters or accompanied by warnings to ignore instructions embedded within the text.
- Capability inventory: The skill has the ability to execute system commands through
subprocessand perform file writing operations. - Sanitization: The script decodes captured information using
errors="ignore"but performs no validation, filtering, or sanitization of the SSID content before including it in the output JSON report.
Audit Metadata