Mcp Pcapy Ng
Let your coding agent capture and inspect live network packets when you are debugging protocol issues, API traffic, or connectivity in dev or staging.
Overview
io.github.daedalus/mcp-pcapy-ng is an Operate-phase MCP server that exposes pcapy-ng packet capture tools to AI coding agents over stdio.
What is this MCP server?
- Stdio MCP server (v0.1.0) packaging pcapy-ng for agent-driven capture workflows
- PyPI distribution as mcp-pcapy-ng with Model Context Protocol tool exposure
- Useful for reproducing HTTP/TCP issues without leaving the agent chat
- Fits solo builders who self-host APIs and need quick PCAP-style introspection
- GitHub source at daedalus/mcp-pcapy-ng with MCP server schema metadata
- Server version 0.1.0 in published MCP manifest
- Single PyPI package identifier mcp-pcapy-ng
- Transport type stdio per Model Context Protocol server schema
What problem does it solve?
You cannot see what your app actually sent on the wire, so agents guess at network failures from logs alone.
Who is it for?
Self-hosted backend builders debugging intermittent API or socket issues who already comfort with Python and local capture permissions.
Skip if: Teams needing compliant enterprise PCAP storage, cloud-native service mesh tracing, or builders with no elevated capture rights on their OS.
What do I get? / Deliverables
Your agent can request packet capture sessions through MCP so you validate hypotheses about traffic, ports, and protocols with evidence.
- Registered stdio MCP server entry pointing at mcp-pcapy-ng
- Agent-callable packet capture workflow backed by pcapy-ng
- Local evidence files or summaries agents can reference in incident notes
Recommended MCP Servers
Journey fit
Packet capture is most actionable once something is running and traffic is failing mysteriously—classic production and staging operations work rather than greenfield ideation. Monitoring is the canonical shelf because pcapy-ng surfaces wire-level visibility you use to correlate errors, latency, and unexpected hosts with application behavior.
How it compares
Wire-level MCP bridge to pcapy-ng, not a hosted APM product or a Claude Code skill markdown workflow.
Common Questions / FAQ
Who is mcp-pcapy-ng for?
Solo builders and small teams running Python stacks who want their MCP-enabled agent to run packet capture when debugging live services.
When should I use mcp-pcapy-ng?
Use it during operate or ship hardening when logs are inconclusive and you need to confirm hosts, ports, or payload shapes on the network.
How do I add mcp-pcapy-ng to my agent?
Install the mcp-pcapy-ng package from PyPI, register the stdio MCP server in Claude Code Cursor or Codex per your client’s MCP config, and ensure capture permissions on your machine.