agentleFS
sentineldesksentineldesk/agent/skills/network-recon/SKILL.md

network-recon

Skill0 starsChanged 3 months agoApache-2.0View on GitHub

Map and diagnose networks and hosts you are authorized to reach — local interfaces and listeners, reachability, open ports, service versions, TLS, DNS, and packet capture — through the MCP tools. For authorized diagnostics and security assessment only. Every probe runs through a visible, logged tool; do not touch hosts nobody named. Read this before scanning or capturing.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read it back with run_command "tcpdump -r /tmp/cap.pcap -n | head" or the Wireshark CLI, and do not leave the capture running.

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Authorization is the first step, not a footnote

Recon is defensible when it is scoped and traced, and indefensible when it is neither. Two rules, before any packet leaves:

  1. Only hosts you were explicitly asked to assess. The machine you are on and targets a person named. Never a range somebody might have meant, never "the rest of the subnet while I'm here." If the scope is unclear, ask_human.
  2. Everything runs through the catalogue, which means everything is on the shared screen and the action log. That is the feature: an authorized scan you can point to afterwards is the difference between an assessment and an intrusion. There is no quiet path and you must not look for one.

Start with the machine you are on — free and always in scope

  • Interfaces and routes: run_command "ip -brief addr", run_command "ip route".
  • Local listeners: run_command "ss -tulpn" — every open port cross-checked against list_processes. This is often the whole assessment for the host itself.
  • DNS: run_command "dig +short <name>", run_command "cat /etc/resolv.conf".

Reaching a named target

  • Alive and how far: run_command "ping -c3 <host>", run_command "traceroute <host>" (install via install_packages if absent).
  • Ports and services: nmap is the tool. Install it with install_packages, and run a real scan as a job, not a command — a full scan is slower than a few seconds: job_start "nmap -sV -Pn <host>", then job_wait. A wait that times out has not stopped the scan; call job_wait again rather than launching a second nmap.
  • TLS: run_command "openssl s_client -connect <host>:443 -servername <host> </dev/null 2>/dev/null | openssl x509 -noout -dates -subject -issuer".
  • HTTP surface: run_command "curl -sSI https://<host>" for headers; the browser_* tools when you need to see a page render.

Packet capture is sharp — scope it tightly

tcpdump captures whatever crosses the interface, which on a shared machine can include other people's traffic. So:

  • Capture a named host and port, time- or count-bounded, never "everything": job_start "tcpdump -n -c 500 host <host> and port <port> -w /tmp/cap.pcap".
  • Say you are capturing and why, before you start — the same courtesy a recording gets, for the same reason.
  • Read it back with run_command "tcpdump -r /tmp/cap.pcap -n | head" or the Wireshark CLI, and do not leave the capture running.

Report against the scope you were given

List what you found per host: ports, versions, TLS, anything notable — and nothing about hosts you were not asked about, even if they answered. The trail shows exactly what you probed, which is what keeps an authorized assessment authorized.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.