The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data files, creating a vulnerability surface where malicious instructions embedded in those files could influence agent behavior.
Ingestion points: The scripts/agent.py script reads asset inventory and firewall rule CSV files. The scripts/process.py script reads PCAP capture files and performs live traffic sniffing.
Boundary markers: The scripts do not employ delimiters or explicit instructions to ignore instructions embedded within the processed data.
Capability inventory: The skill has the capability to execute shell commands (nmap) and perform file system writes.
Sanitization: Content extracted from CSV fields (such as asset names, vendors, or firewall rules) and PCAP metadata is used in reports without validation or sanitization.
[COMMAND_EXECUTION]: The skill performs dynamic command execution using the nmap network scanning utility.
Evidence: In scripts/agent.py, the scan_ot_protocols function (line 89) constructs a command list including a target_subnet derived from user input and executes it via subprocess.run(). While the use of a list argument rather than a raw shell string mitigates traditional shell injection, it still allows for the execution of external binaries with dynamic parameters.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of external dependencies to function.
Evidence: The documentation in SKILL.md and the import checks in scripts/process.py instruct the user to install the scapy library via pip for packet analysis functionality.

performing-ot-network-security-assessment