The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts (scripts/agent.py and code blocks in SKILL.md) that use the socket library to perform active network connectivity testing and port scanning. These functions are used to audit network boundaries and detect exposed OT protocols (e.g., Modbus, S7comm, DNP3), which is the primary and legitimate purpose of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill features data ingestion points where it reads asset inventory from CSV files and zone mappings from JSON files.
Ingestion points: PurdueSegmentationPlanner.load_asset_inventory in SKILL.md and load_zone_map in scripts/agent.py.
Boundary markers: Not present; the skill assumes the provided configuration files are well-formed.
Capability inventory: The ingested data is used to define target IPs for network connectivity tests via the socket library.
Sanitization: The skill uses standard csv and json libraries for parsing. No data is passed to dangerous execution sinks like eval() or os.system(), making the risk of exploitation low.
[EXTERNAL_DOWNLOADS]: The references/api-reference.md file recommends installing scapy and requests via pip. These are well-known, trusted libraries commonly used for network analysis and HTTP operations in cybersecurity workflows.

implementing-purdue-model-network-segmentation