Trace which local processes and hosts are consuming bandwidth during incidents with bandwhich
Identify which processes, connections, and remote hosts are actually using bandwidth before you chase the wrong incident hypothesis.
npx skills add agentskillexchange/skills --skill trace-which-local-processes-and-hosts-are-consuming-bandwidth-during-incidents-with-bandwhich
Use bandwhich when an agent needs a fast, local answer to a narrow incident question: what process or remote host is consuming network bandwidth right now. This is not a generic network monitor or platform card. It is a bounded diagnostic workflow around sniffing one interface, mapping traffic back to processes, and separating process, connection, and address views during live troubleshooting.
What this skill actually does
Invoke it instead of using a full observability stack when the job is immediate host-level attribution on the machine in front of you. Typical runs inspect one interface, optionally disable reverse DNS for speed, and narrow the view to processes, connections, or remote addresses before escalation.