hf-vhf-radio-attack

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external RF sources, including aircraft transponders (ADS-B), maritime vessel broadcasts (AIS), and pager networks (POCSAG/FLEX), via tools like dump1090, AIS-catcher, and multimon-ng. While the provided parsing scripts use regular expressions and standard JSON loaders to analyze this data for red-teaming purposes, vulnerabilities in the underlying decoder tools or complex parsing logic could potentially be targeted by specially crafted, malicious RF transmissions.
  • [DYNAMIC_EXECUTION]: Multiple files, notably payloads.md and the deep-dive guide, utilize shell heredocs to dynamically generate Python scripts on the local filesystem. These generated scripts are then executed to perform tasks such as anomaly detection, signal generation for authorized research, and data visualization. While these scripts are based on visible templates, the mechanism represents runtime code generation and execution.
  • [COMMAND_EXECUTION]: The skill frequently invokes powerful command-line tools for signal processing (e.g., hackrf_transfer, rtl_sdr, dump1090, rtl_fm) and system-level package management (sudo apt install). These commands are necessary for the skill's primary function of radio-layer assessment and hardware interfacing.
  • [EXTERNAL_DOWNLOADS]: The documentation provides instructions to clone and build various specialized SDR tools from public GitHub repositories (e.g., AIS-catcher, dumpvdl2, gr-adsb, dumphfdl) and install specific Python dependencies such as pyModeS and pyais. These are community-maintained tools for radio protocol analysis used to extend the agent's capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 04:36 PM