The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/agent.py utility invokes external command-line tools including tshark, crackle, and ubertooth-btle to perform packet analysis and capture operations.
Evidence: Subprocess calls in analyze_pcap_for_ble_attacks and run_ubertooth_capture functions in scripts/agent.py.
Context: These calls are implemented using argument lists rather than shell strings, which is a secure practice for calling external utilities in a controlled manner.
[INDIRECT_PROMPT_INJECTION]: The skill processes advertising data and characteristic values from external BLE devices, which constitutes an ingestion point for untrusted data.
Ingestion points: scripts/agent.py functions scan_ble_devices and enumerate_gatt_services.
Boundary markers: The skill does not implement specific delimiters or warnings to ignore instructions embedded within the BLE data.
Capability inventory: The script can execute specific external commands via subprocess and write results to a local JSON file.
Sanitization: The script converts most BLE data to hex strings or uses basic UTF-8 decoding with error suppression.
[EXTERNAL_DOWNLOADS]: The documentation references the installation of security tools from external sources.
Evidence: SKILL.md instructs the user to build the crackle tool from source at github.com/mikeryan/crackle.
Context: This targets a well-known repository for a standard security auditing tool, which is considered a safe reference in this domain.

detecting-bluetooth-low-energy-attacks