port-mass-scan
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to process external target files into command arguments for masscan and rustscan. Specifically, commands like
masscan -iL "$RANGE_FILE"andrustscan -a "$TARGET"execute based on variables that typically originate from untrusted user input. - [EXTERNAL_DOWNLOADS]: The 'IP Camera Hunting' section uses
curlto probe discovered IP addresses for snapshots (http://$ip:8010/axis-cgi/jpg/image.cgi). This involves making unauthenticated network requests to potentially external infrastructure based on scan results. - [PROMPT_INJECTION]: The skill documents an 'Indirect Prompt Injection' surface (Category 8). It ingests untrusted data from target files (
$RANGE_FILE,hits.txt) and scan results (masscan_common.json) which are then used to drive subsequent automated actions, such as full scans or curl probes, without explicit sanitization or boundary markers.
Audit Metadata