The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/agent.py tool uses the subprocess module to execute tshark for protocol analysis. The code uses a secure list-based argument passing method which prevents command injection vulnerabilities by not invoking a shell.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external network capture data (PCAP files), which constitutes an untrusted input surface. Evidence: (1) Ingestion point: PCAP file parsing via Scapy and tshark in agent.py and process.py. (2) Boundary markers: Not present. (3) Capability inventory: Subprocess execution and file writing. (4) Sanitization: The code uses errors='replace' during data decoding to handle malformed packet content safely.\n- [EXTERNAL_DOWNLOADS]: The documentation references official Wireshark docs, SANS Institute courses, and Netresec PCAP repositories. These are well-known, reputable resources within the cybersecurity community and are documented neutrally for educational use.

performing-network-packet-capture-analysis