The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/agent.py utilizes the subprocess.run module to execute tshark commands for deep packet inspection and conversation analysis. This interaction is performed using structured argument lists without a shell wrapper, effectively mitigating shell injection risks. Evidence: subprocess.run(cmd, capture_output=True, text=True) in scripts/agent.py.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by processing untrusted network traffic data (PCAP files) and presenting it to the agent for analysis.
Ingestion points: Raw packet data is ingested via scapy.all.rdpcap() in both scripts/agent.py and scripts/process.py.
Boundary markers: The scripts output results in JSON format, which provides some structural separation, but no explicit delimiters or warnings are used to protect the agent from adversarial strings embedded in packet headers or payloads.
Capability inventory: The skill has the capability to execute system commands via tshark and write report files to the local file system.
Sanitization: String decoding uses errors="replace" to prevent encoding-related crashes, but no logical sanitization is performed to filter out potentially malicious language instructions contained within the analyzed traffic.

performing-network-packet-capture-analysis