external/anthropic-cybersecurity-skills/skills/performing-s7comm-protocol-security-analysis/SKILL.md
Perform security analysis of Siemens S7comm and S7CommPlus protocols used by SIMATIC S7 PLCs to identify vulnerabilities including replay attacks, integrity bypass, unauthorized CPU stop commands, and program download manipulation exploiting weaknesses in S7-300, S7-400, S7-1200, and S7-1500 controllers.
npx skillsauth add seikaikyo/dash-skills performing-s7comm-protocol-security-analysisInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Do not use for scanning production Siemens PLCs without authorization and a test plan (this can crash controllers), for non-Siemens protocol analysis (see detecting-modbus-command-injection-attacks for Modbus), or for modifying PLC programs in a production environment.
#!/usr/bin/env python3
"""S7comm Protocol Security Analyzer.
Analyzes Siemens S7comm protocol traffic to identify security
vulnerabilities, unauthorized access patterns, and potential
attack indicators against SIMATIC S7 PLCs.
"""
import struct
import sys
import json
from collections import defaultdict
from datetime import datetime
from typing import Dict, List, Optional
try:
from scapy.all import rdpcap, IP, TCP
except ImportError:
print("Install scapy: pip install scapy")
sys.exit(1)
# S7comm ROSCTR (PDU type) definitions
S7_ROSCTR = {
0x01: "Job (Request)",
0x02: "Ack",
0x03: "Ack_Data (Response)",
0x07: "Userdata",
}
# S7comm function codes
S7_FUNCTIONS = {
0x00: "CPU services",
0x04: "Read Variable",
0x05: "Write Variable",
0x1A: "Request Download (Program)",
0x1B: "Download Block",
0x1C: "Download Ended",
0x1D: "Start Upload (Read Program)",
0x1E: "Upload Block",
0x1F: "Upload Ended",
0x28: "PI Service (Start/Stop CPU)",
0x29: "PLC Stop",
0xF0: "Setup Communication",
}
# Critical security-relevant operations
CRITICAL_FUNCTIONS = {0x1A, 0x1B, 0x1C, 0x28, 0x29, 0x05}
PROGRAM_FUNCTIONS = {0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F}
class S7commSecurityFinding:
"""Represents a security finding in S7comm traffic."""
def __init__(self, severity: str, finding_type: str, src_ip: str,
dst_ip: str, function: str, description: str,
cve: str = "", recommendation: str = ""):
self.timestamp = datetime.now().isoformat()
self.severity = severity
self.finding_type = finding_type
self.src_ip = src_ip
self.dst_ip = dst_ip
self.function = function
self.description = description
self.cve = cve
self.recommendation = recommendation
class S7commAnalyzer:
"""Analyzes S7comm protocol traffic for security vulnerabilities."""
def __init__(self):
self.findings: List[S7commSecurityFinding] = []
self.sessions: Dict[str, dict] = defaultdict(lambda: {
"packets": 0,
"functions_seen": set(),
"writes": 0,
"program_downloads": 0,
"cpu_commands": 0,
"first_seen": None,
"last_seen": None,
})
self.authorized_engineering: set = set()
self.packet_count = 0
def set_authorized_stations(self, ips: List[str]):
"""Set list of authorized engineering workstation IPs."""
self.authorized_engineering = set(ips)
def parse_s7comm(self, payload: bytes) -> Optional[dict]:
"""Parse S7comm protocol data from TCP payload."""
# TPKT header: version(1) + reserved(1) + length(2)
if len(payload) < 4:
return None
tpkt_version = payload[0]
if tpkt_version != 3:
return None
tpkt_length = struct.unpack(">H", payload[2:4])[0]
# COTP header follows TPKT
if len(payload) < 7:
return None
cotp_length = payload[4]
cotp_type = payload[5]
# S7comm starts after COTP
s7_offset = 4 + 1 + cotp_length
if len(payload) < s7_offset + 10:
return None
# S7comm header
protocol_id = payload[s7_offset]
if protocol_id != 0x32: # S7comm magic byte
return None
rosctr = payload[s7_offset + 1]
redundancy = struct.unpack(">H", payload[s7_offset + 2:s7_offset + 4])[0]
pdu_ref = struct.unpack(">H", payload[s7_offset + 4:s7_offset + 6])[0]
param_length = struct.unpack(">H", payload[s7_offset + 6:s7_offset + 8])[0]
data_length = struct.unpack(">H", payload[s7_offset + 8:s7_offset + 10])[0]
result = {
"rosctr": rosctr,
"rosctr_name": S7_ROSCTR.get(rosctr, f"Unknown (0x{rosctr:02x})"),
"pdu_ref": pdu_ref,
"param_length": param_length,
"data_length": data_length,
}
# Parse function code from parameters
param_offset = s7_offset + 10
if rosctr in (0x01, 0x03) and param_length > 0 and len(payload) > param_offset:
func_code = payload[param_offset]
result["function_code"] = func_code
result["function_name"] = S7_FUNCTIONS.get(func_code, f"Unknown (0x{func_code:02x})")
return result
def analyze_packet(self, pkt):
"""Analyze a packet for S7comm security issues."""
self.packet_count += 1
if not pkt.haslayer(IP) or not pkt.haslayer(TCP):
return
tcp = pkt[TCP]
if tcp.dport != 102 and tcp.sport != 102:
return
payload = bytes(tcp.payload)
if not payload:
return
s7 = self.parse_s7comm(payload)
if not s7:
return
src_ip = pkt[IP].src
dst_ip = pkt[IP].dst
session_key = f"{src_ip}->{dst_ip}"
session = self.sessions[session_key]
session["packets"] += 1
if session["first_seen"] is None:
session["first_seen"] = float(pkt.time)
session["last_seen"] = float(pkt.time)
func_code = s7.get("function_code")
if func_code is not None:
session["functions_seen"].add(func_code)
# Check 1: Unauthorized engineering station
if tcp.dport == 102 and func_code in CRITICAL_FUNCTIONS:
if self.authorized_engineering and src_ip not in self.authorized_engineering:
self.findings.append(S7commSecurityFinding(
severity="CRITICAL",
finding_type="UNAUTHORIZED_ENGINEERING_ACCESS",
src_ip=src_ip, dst_ip=dst_ip,
function=s7.get("function_name", "Unknown"),
description=(
f"Critical S7comm operation from unauthorized source {src_ip}. "
f"Function: {s7.get('function_name')}. Only authorized TIA Portal "
f"workstations should issue these commands."
),
recommendation="Block unauthorized sources at industrial firewall. Investigate source host for compromise.",
))
# Check 2: CPU Stop command
if func_code == 0x29:
session["cpu_commands"] += 1
self.findings.append(S7commSecurityFinding(
severity="CRITICAL",
finding_type="CPU_STOP_COMMAND",
src_ip=src_ip, dst_ip=dst_ip,
function="PLC CPU Stop (0x29)",
description=f"CPU STOP command sent to PLC at {dst_ip}. This halts PLC program execution.",
cve="MITRE T0881 - Service Stop",
recommendation="Verify if this is an authorized maintenance action. If not, isolate source immediately.",
))
# Check 3: Program download
if func_code in (0x1A, 0x1B, 0x1C):
session["program_downloads"] += 1
self.findings.append(S7commSecurityFinding(
severity="CRITICAL",
finding_type="PROGRAM_DOWNLOAD",
src_ip=src_ip, dst_ip=dst_ip,
function=s7.get("function_name", "Download"),
description=(
f"PLC program download operation to {dst_ip}. "
f"This modifies the running control logic on the PLC."
),
cve="MITRE T0843 - Program Download",
recommendation="Verify against change management records. Compare with known-good program backup.",
))
# Check 4: Write variable operation
if func_code == 0x05:
session["writes"] += 1
# Check 5: Program upload (exfiltration of PLC code)
if func_code in (0x1D, 0x1E, 0x1F):
self.findings.append(S7commSecurityFinding(
severity="HIGH",
finding_type="PROGRAM_UPLOAD_EXFILTRATION",
src_ip=src_ip, dst_ip=dst_ip,
function=s7.get("function_name", "Upload"),
description=f"PLC program upload (read) from {dst_ip}. Source {src_ip} is extracting PLC control logic.",
recommendation="Verify if this is authorized maintenance. Unauthorized uploads indicate reconnaissance.",
))
def check_known_vulnerabilities(self):
"""Check for known Siemens S7 vulnerabilities based on observed behavior."""
vuln_checks = [
{
"name": "S7-300/400 Replay Attack Vulnerability",
"cve": "CVE-2019-13945",
"description": "S7-300/400 PLCs lack integrity checks on S7comm sessions, allowing replay attacks",
"affected": "S7-300, S7-400 (all firmware versions)",
"severity": "HIGH",
},
{
"name": "S7CommPlus Integrity Bypass",
"cve": "Research finding (Biham et al.)",
"description": "S7CommPlusV3 integrity mechanism can be bypassed by attackers who can observe one legitimate session",
"affected": "S7-1200 (< V4.5), S7-1500 (< V2.9)",
"severity": "HIGH",
},
{
"name": "Unpatchable Hardware Root of Trust",
"cve": "CVE-2022-38773",
"description": "Hardware vulnerability allows bypassing protected boot and persistent firmware modification",
"affected": "S7-1500 (specific hardware revisions)",
"severity": "CRITICAL",
},
{
"name": "Remote DoS via Port 102",
"cve": "CVE-2019-10929",
"description": "Specially crafted packets on TCP port 102 can crash S7 PLCs remotely",
"affected": "S7-300, S7-400, S7-1200, S7-1500 (specific firmware)",
"severity": "HIGH",
},
]
return vuln_checks
def generate_report(self):
"""Generate comprehensive S7comm security analysis report."""
print(f"\n{'='*70}")
print("S7COMM PROTOCOL SECURITY ANALYSIS REPORT")
print(f"{'='*70}")
print(f"Analysis Time: {datetime.now().isoformat()}")
print(f"Packets Analyzed: {self.packet_count}")
print(f"S7comm Sessions: {len(self.sessions)}")
print(f"Security Findings: {len(self.findings)}")
print(f"\n--- SESSION SUMMARY ---")
for key, session in self.sessions.items():
funcs = [S7_FUNCTIONS.get(f, f"0x{f:02x}") for f in session["functions_seen"]]
print(f"\n {key}")
print(f" Packets: {session['packets']}")
print(f" Functions: {', '.join(funcs)}")
print(f" Writes: {session['writes']}")
print(f" Program Downloads: {session['program_downloads']}")
print(f" CPU Commands: {session['cpu_commands']}")
if self.findings:
print(f"\n--- SECURITY FINDINGS ---")
for f in self.findings:
print(f"\n [{f.severity}] {f.finding_type}")
print(f" Source: {f.src_ip} -> {f.dst_ip}")
print(f" Function: {f.function}")
print(f" Detail: {f.description}")
if f.cve:
print(f" Reference: {f.cve}")
if f.recommendation:
print(f" Action: {f.recommendation}")
print(f"\n--- KNOWN VULNERABILITY ASSESSMENT ---")
for vuln in self.check_known_vulnerabilities():
print(f"\n [{vuln['severity']}] {vuln['name']}")
print(f" CVE: {vuln['cve']}")
print(f" Affected: {vuln['affected']}")
print(f" Detail: {vuln['description']}")
if __name__ == "__main__":
analyzer = S7commAnalyzer()
analyzer.set_authorized_stations(["10.10.2.50", "10.10.2.51"])
if len(sys.argv) >= 2:
print(f"[*] Analyzing capture: {sys.argv[1]}")
packets = rdpcap(sys.argv[1])
for pkt in packets:
analyzer.analyze_packet(pkt)
analyzer.generate_report()
else:
print("Usage: python s7comm_analyzer.py <capture.pcap>")
print(" Analyzes S7comm traffic for security vulnerabilities")
| Term | Definition | |------|------------| | S7comm | Siemens proprietary protocol for communication with SIMATIC S7 PLCs over TCP port 102, layered on COTP/TPKT | | S7CommPlus | Enhanced version of S7comm used by S7-1200/1500 with integrity protection mechanisms | | ROSCTR | Remote Operating Service Control field in S7comm header indicating PDU type (Job, Ack, Ack_Data, Userdata) | | TIA Portal | Totally Integrated Automation Portal -- Siemens engineering software for programming S7 PLCs | | CPU Stop (0x29) | S7comm function that halts PLC program execution, a critical denial-of-service operation | | Program Download (0x1A) | S7comm function initiating transfer of new control logic to a PLC, representing the highest risk operation |
Context: A Dragos sensor alerts on S7comm program download traffic from an IP address that is not the authorized TIA Portal engineering workstation.
Approach:
Pitfalls: S7-300/400 PLCs have no cryptographic integrity protection -- any device that can reach TCP port 102 can send commands. Do not rely solely on PLC passwords as they are transmitted in cleartext in S7comm (not S7CommPlus). Network segmentation is the primary defense.
S7COMM SECURITY ANALYSIS REPORT
===================================
Date: YYYY-MM-DD
Scope: [Network segments analyzed]
SESSION INVENTORY:
Engineering stations: [count and IPs]
PLCs communicating: [count and IPs]
Unauthorized sources: [count]
CRITICAL FINDINGS:
CPU Stop commands: [count]
Program downloads: [count from unauthorized sources]
Replay attack potential: [assessment]
VULNERABILITY ASSESSMENT:
S7-300/400 (no integrity): [count of affected PLCs]
S7-1200/1500 (S7CommPlus): [firmware assessment]
Known CVEs applicable: [list]
RECOMMENDATIONS:
1. [Highest priority remediation]
2. [Network segmentation improvement]
3. [Monitoring enhancement]
development
拋棄式 HTML mockup 比稿:產出 2 到 3 個設計立場不同的變體(密度 / 版式 / 強調軸,不是換色),各附取捨說明,最後給有立場的對比結論。適用:「畫個草圖」「比較 A 版 B 版」「先看方向再做」「給我看幾種做法」。要 production 元件或設計已定案時不適用。
tools
需求不明時的意圖萃取訪談:一次一題、每題附上自己的猜測、聽出「真正想要 vs 覺得應該要」,直到能預測使用者反應(約 95% 信心)才動工。適用:需求缺少對象 / 動機 / 成功標準 / 約束,或使用者點名「訪談我」「先確認一下」「我們確定嗎」。明確自足的指示、純資訊查詢、機械性操作不適用。
development
對非平凡決策啟動新鮮 context 對抗審查(找碴不背書),在修正還便宜的時候抓出錯誤方向。適用:高風險改動(production、資安敏感邏輯、不可逆操作)、不熟的程式碼、要宣稱「這樣是安全的 / 可行的」之前。機械性操作與一行修改不適用。
testing
Reference for writing and editing agent skills well — the vocabulary and principles that make a skill predictable. Consult when authoring, reviewing, or pruning a SKILL.md.