francescostabile/numasec/forensics-kit

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Digital forensics and incident response toolbox. Load when the operator asks about a pcap, a binary, a memory dump, a suspicious file, malware triage, IOC hunting, or post-incident analysis. Covers network (tshark), binaries (radare2, strings, binwalk, file, exiftool), memory (volatility), and pattern matching (YARA). All commands assume the artifact is local and disposable; never analyze in-place on a production system.

Install

npx vskill@latest install francescostabile/numasec/forensics-kit

SKILL.md

Forensics & IR kit

Rule of engagement. Forensics work runs on copies. Never analyze a live victim system in place — cp/dd to a workspace, hash the original, and work on the copy. Every finding must reference the SHA256 of the artifact so chain-of-custody survives.

Output contract. Each IOC, carved artifact, or suspicious observation is a candidate observation (severity medium+ for IOCs, confidence high for exact matches, medium for heuristic hits).

When to load

SituationWhy
Operator drops a .pcap, .pcapng, .bin, .dmp, .elf, .exeYou need these tools
Post-compromise triage of a Linux/Windows host imageMemory + binaries
Suspicious file discovered by a scannerstrings / binwalk / file first
IOC hunt across a datasetYARA rules

Toolbox

1. file + exiftool + strings — first-3-minutes triage

Before anything heavier, always identify the artifact and skim it.

bash
file suspicious.bin
exiftool suspicious.bin
strings -n 8 suspicious.bin | head -200
strings -eL -n 8 suspicious.bin | head -100 # UTF-16LE (Windows PE)

Interpretation. file gives format/arch. exiftool surfaces author/tool metadata. strings reveals hard-coded URLs, C2 hostnames, file paths, error messages, library imports.

2. binwalk — Carve embedded content
bash
binwalk -B firmware.bin # signature scan
binwalk -e firmware.bin # extract what it recognizes
binwalk -E firmware.bin # entropy (encryption/compression indicator)

Severity hints. Firmware that carves cleanly to a shell script + a busybox → high if it contains credentials. Very-high uniform entropy → encrypted blob; flag and move on.

3. tshark — Packet capture analysis
bash
# 1-line summary of protocols
tshark -r capture.pcapng -q -z io,phs
# HTTP requests
tshark -r capture.pcapng -Y http.request -T fields -e ip.dst -e http.host -e http.request.uri | sort -u
# DNS queries
tshark -r capture.pcapng -Y dns -T fields -e dns.qry.name | sort -u | head -50
# Suspicious destinations (non-RFC1918)
tshark -r capture.pcapng -T fields -e ip.dst | sort -u | grep -vE '^(10\.|192\.168\.|172\.(1[6-9]|2[0-9]|3[01])\.)'
# TLS SNI (see what hosts are being contacted even under HTTPS)
tshark -r capture.pcapng -Y "tls.handshake.type == 1" -T fields -e tls.handshake.extensions_server_name | sort -u
# Extract objects (HTTP, SMB)
mkdir -p /tmp/tshark-export
tshark -r capture.pcapng --export-objects http,/tmp/tshark-export

Interpretation. Unknown SNI + sustained beacon interval ≈ C2 candidate. Non-RFC1918 destinations in an internal capture → exfiltration candidate.

4. volatility (Volatility 3) — Memory forensics
bash
# Identify the profile implicitly (Volatility 3 auto-detects)
vol -f memory.dmp windows.info
vol -f memory.dmp windows.pslist
vol -f memory.dmp windows.pstree
vol -f memory.dmp windows.cmdline
vol -f memory.dmp windows.netscan
vol -f memory.dmp windows.malfind # heuristics for injected code
vol -f memory.dmp windows.dlllist --pid <PID>
vol -f memory.dmp windows.hashdump # SAM hashes
# Linux equivalents
vol -f memory.dmp linux.pslist
vol -f memory.dmp linux.bash # recovered shell history

Severity hints. malfind hit with RWX region and MZ header → high, confidence medium. hashdump success → critical (credential exposure).

5. radare2 — Binary RE
bash
r2 -A suspicious.elf
# Inside r2:
# afl list functions
# s main; pdf disassemble main
# iz print strings
# ii imports
# aaaa deeper analysis

For one-liners:

bash
rabin2 -z suspicious.elf # strings in data sections
rabin2 -I suspicious.elf # binary info
rabin2 -i suspicious.elf # imports
6. YARA — Signature-based hunting
bash
# Install: apt install yara
# Quick hunt using an existing ruleset
yara -r /opt/rules/index.yar suspicious.bin
# Directory scan
yara -r /opt/rules/index.yar /mnt/image/
# Author a rule from observed strings
cat > /tmp/demo.yar <<'EOF'
rule demo_c2 {
meta:
author = "numasec"
description = "heuristic: hard-coded C2 host"
strings:
$s1 = "evil-c2.example.com"
$s2 = { 4D 5A } // MZ header
condition:
$s1 and $s2
}
EOF
yara /tmp/demo.yar suspicious.bin

Good rulesets to start from (no key required):

  • YARA Rules project: https://github.com/Yara-Rules/rules
  • Neo23x0 Signature-Base: https://github.com/Neo23x0/signature-base
7. Hashing + Chain-of-custody hygiene
bash
sha256sum artifact.bin > artifact.bin.sha256
# Every observation references $(cut -d' ' -f1 artifact.bin.sha256)

Also keep a copy of ls -la artifact.bin and stat artifact.bin so mtime/ctime are recorded.

Standard triage opening

For an unknown artifact:

  1. sha256sum + file + exiftool + strings | head.
  2. If it's a binary: binwalk -E (entropy) → binwalk -e if entropy is moderate → radare2 -A if interesting.
  3. If it's a pcap: tshark -z io,phs → HTTP/DNS dumps → SNI → export objects.
  4. If it's a memory image: vol windows.info → pslist → netscan → malfind → hashdump.
  5. Run YARA with a generic ruleset.
  6. Summarize IOCs (hosts, hashes, paths, mutexes, registry keys) into observations.
  7. Cross-reference IOCs against cve tool (T10) and any open-source threat intel the operator has.

What this skill does NOT do

  • No live acquisition. Operator pre-acquires memory/disk.
  • No cloud forensics (AWS CloudTrail parsing, GCP audit logs). Those need a dedicated skill pack.
  • No automated malware detonation / sandboxing. Use Cuckoo/CAPE externally if needed.

Escalation

After triage:

  • If a binary has exploitable defects → hand off to the hacking agent via task tool.
  • If an IR finding needs remediation code changes → hand off to the appsec agent.
  • If the pcap shows sustained C2 → open a parallel threat-intel subtask (security agent).

Security

  • Security scanPassed, 91 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
372
Repository forks
47
7-day trend score
30

Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 24 May 2026, 5 months ago.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/francescostabile/numasec/forensics-kit/badge)](https://verified-skill.com/skills/francescostabile/numasec/forensics-kit)

Something wrong with this skill or its page? Tell us.