AGENT SKILL

performing-malware-persistence-investigation

Systematically investigate all persistence mechanisms on Windows and Linux systems to identify how malware survives reboots and maintains access.

License 许可
Apache-2.0
最近更新
2026-08-31
Stars
★ 3.6w
Forks
4.4k

01 前言 Front Matter

name
performing-malware-persistence-investigation
tags
[ "forensics", "malware-persistence", "autoruns", "registry", "scheduled-tasks", "rootkit-detection", "incident-response" ]
author
mahipal
domain
cybersecurity
license
Apache-2.0
version
1.0
nist_csf
[ "RS.AN-03", "DE.AE-02", "RS.MA-01" ]
subdomain
digital-forensics
description
Systematically investigate all persistence mechanisms on Windows and Linux systems to identify how malware survives reboots and maintains access.
mitre_attack
[ "T1005", "T1074", "T1119", "T1070", "T1547" ]

02 资源文件 3

03 SKILL.md 正文

Performing Malware Persistence Investigation

When to Use

  • When investigating how malware maintains presence on a compromised system after reboots
  • During incident response to identify all persistence mechanisms for complete remediation
  • For threat hunting to discover unauthorized autostart entries across endpoints
  • When analyzing malware behavior to understand its persistence strategy
  • For verifying that all persistence has been removed after incident remediation

Prerequisites

  • Forensic image or live system access with administrative privileges
  • Autoruns (Sysinternals) for Windows persistence enumeration
  • RegRipper for offline registry analysis
  • Understanding of Windows and Linux persistence mechanisms
  • YARA rules for scanning persistence locations
  • Baseline of known-good autorun entries for comparison

Workflow

Step 1: Enumerate Windows Registry Persistence

# Extract registry hives from forensic image
mount -o ro,loop,offset=$((2048*512)) /cases/case-2024-001/images/evidence.dd /mnt/evidence

# Key registry persistence locations
python3 << 'PYEOF'
from Registry import Registry
import json

results = {'registry_persistence': []}

# SYSTEM hive analysis
system_reg = Registry.Registry("/cases/case-2024-001/registry/SYSTEM")
select = system_reg.open("Select")
current = select.value("Current").value()
cs = f"ControlSet{current:03d}"

# Services (very common persistence)
services = system_reg.open(f"{cs}\\Services")
for svc in services.subkeys():
    try:
        start_type = svc.value("Start").value()
        image_path = ""
        try:
            image_path = svc.value("ImagePath").value()
        except:
            pass
        # Start types: 0=Boot, 1=System, 2=Auto, 3=Manual, 4=Disabled
        if start_type in (0, 1, 2) and image_path:
            svc_type = svc.value("Type").value() if svc.values() else 0
            results['registry_persistence'].append({
                'location': f'HKLM\\SYSTEM\\{cs}\\Services\\{svc.name()}',
                'type': 'Service',
                'value': image_path,
                'start_type': start_type,
                'timestamp': str(svc.timestamp())
            })
    except Exception:
        pass

# SOFTWARE hive analysis
sw_reg = Registry.Registry("/cases/case-2024-001/registry/SOFTWARE")

# Machine Run keys
run_keys = [
    "Microsoft\\Windows\\CurrentVersion\\Run",
    "Microsoft\\Windows\\CurrentVersion\\RunOnce",
    "Microsoft\\Windows\\CurrentVersion\\RunServices",
    "Microsoft\\Windows\\CurrentVersion\\RunServicesOnce",
    "Microsoft\\Windows\\CurrentVersion\\Policies\\Explorer\\Run",
    "Wow6432Node\\Microsoft\\Windows\\CurrentVersion\\Run",
    "Wow6432Node\\Microsoft\\Windows\\CurrentVersion\\RunOnce",
]

for key_path in run_keys:
    try:
        key = sw_reg.open(key_path)
        for value in key.values():
            results['registry_persistence'].append({
                'location': f'HKLM\\SOFTWARE\\{key_path}',
                'type': 'Run Key',
                'name': value.name(),
                'value': str(value.value()),
                'timestamp': str(key.timestamp())
            })
    except Exception:
        pass

# NTUSER.DAT analysis
import glob
for ntuser in glob.glob("/cases/case-2024-001/registry/NTUSER*.DAT"):
    try:
        user_reg = Registry.Registry(ntuser)
        user_run_keys = [
            "Software\\Microsoft\\Windows\\CurrentVersion\\Run",
            "Software\\Microsoft\\Windows\\CurrentVersion\\RunOnce",
            "Software\\Microsoft\\Windows\\CurrentVersion\\Explorer\\Shell Folders\\Startup",
        ]
        for key_path in user_run_keys:
            try:
                key = user_reg.open(key_path)
                for value in key.values():
                    results['registry_persistence'].append({
                        'location': f'HKCU\\{key_path}',
                        'type': 'User Run Key',
                        'name': value.name(),
                        'value': str(value.value()),
                        'timestamp': str(key.timestamp()),
                        'hive': ntuser
                    })
            except Exception:
                pass
    except Exception:
        pass

print(f"Total registry persistence entries: {len(results['registry_persistence'])}")
for entry in results['registry_persistence']:
    print(f"  [{entry['type']}] {entry.get('name', '')} -> {entry.get('value', '')[:100]}")

with open('/cases/case-2024-001/analysis/registry_persistence.json', 'w') as f:
    json.dump(results, f, indent=2)
PYEOF

Step 2: Check Scheduled Tasks and WMI Persistence

# Extract scheduled tasks from forensic image
mkdir -p /cases/case-2024-001/persistence/tasks/
cp -r /mnt/evidence/Windows/System32/Tasks/* /cases/case-2024-001/persistence/tasks/ 2>/dev/null

# Parse scheduled task XML files
python3 << 'PYEOF'
import os, xml.etree.ElementTree as ET

tasks_dir = '/cases/case-2024-001/persistence/tasks/'
suspicious_tasks = []

for root_dir, dirs, files in os.walk(tasks_dir):
    for fname in files:
        fpath = os.path.join(root_dir, fname)
        try:
            tree = ET.parse(fpath)
            root = tree.getroot()
            ns = {'t': 'http://schemas.microsoft.com/windows/2004/02/mit/task'}

            actions = root.findall('.//t:Exec', ns)
            for action in actions:
                command = action.find('t:Command', ns)
                args = action.find('t:Arguments', ns)
                cmd_text = command.text if command is not None else ''
                args_text = args.text if args is not None else ''

                # Flag suspicious commands
                suspicious_indicators = [
                    'powershell', 'cmd.exe', 'wscript', 'cscript', 'mshta',
                    'regsvr32', 'rundll32', 'certutil', 'bitsadmin',
                    '/c ', '-enc', '-e ', 'hidden', 'bypass', 'downloadstring',
                    'invoke-', 'iex', '/tmp/', 'appdata', 'programdata',
                    'temp\\', '.ps1', '.vbs', '.hta', 'base64'
                ]

                is_suspicious = any(s in (cmd_text + ' ' + args_text).lower() for s in suspicious_indicators)

                task_info = {
                    'name': fname,
                    'path': fpath.replace(tasks_dir, ''),
                    'command': cmd_text,
                    'arguments': args_text,
                    'suspicious': is_suspicious
                }

                if is_suspicious:
                    suspicious_tasks.append(task_info)
                    print(f"SUSPICIOUS TASK: {fname}")
                    print(f"  Command: {cmd_text}")
                    print(f"  Arguments: {args_text}")
                    print()

        except Exception as e:
            pass

print(f"\nTotal suspicious scheduled tasks: {len(suspicious_tasks)}")
PYEOF

# Check WMI event subscriptions (common APT persistence)
# WMI repository: C:\Windows\System32\wbem\Repository\
cp -r /mnt/evidence/Windows/System32/wbem/Repository/ /cases/case-2024-001/persistence/wmi/ 2>/dev/null

# Parse WMI persistence using PyWMIPersistenceFinder
python3 << 'PYEOF'
import os, re

# Search WMI OBJECTS.DATA for event subscriptions
wmi_db = '/cases/case-2024-001/persistence/wmi/OBJECTS.DATA'
if os.path.exists(wmi_db):
    with open(wmi_db, 'rb') as f:
        data = f.read()

    # Search for EventFilter strings
    filters = re.findall(b'__EventFilter.*?(?=\x00\x00)', data)
    consumers = re.findall(b'CommandLineEventConsumer.*?(?=\x00\x00)', data)
    bindings = re.findall(b'__FilterToConsumerBinding.*?(?=\x00\x00)', data)

    print("=== WMI PERSISTENCE ===")
    print(f"Event Filters: {len(filters)}")
    print(f"Command Consumers: {len(consumers)}")
    print(f"Filter-Consumer Bindings: {len(bindings)}")

    for consumer in consumers:
        decoded = consumer.decode('utf-8', errors='ignore')
        print(f"  Consumer: {decoded[:200]}")
else:
    print("WMI repository not found")
PYEOF

Step 3: Check File System and Boot Persistence

# Startup folders
echo "=== STARTUP FOLDER CONTENTS ===" > /cases/case-2024-001/analysis/startup_items.txt

ls -la "/mnt/evidence/ProgramData/Microsoft/Windows/Start Menu/Programs/Startup/" \
   >> /cases/case-2024-001/analysis/startup_items.txt 2>/dev/null

for userdir in /mnt/evidence/Users/*/; do
    username=$(basename "$userdir")
    echo "--- User: $username ---" >> /cases/case-2024-001/analysis/startup_items.txt
    ls -la "$userdir/AppData/Roaming/Microsoft/Windows/Start Menu/Programs/Startup/" \
       >> /cases/case-2024-001/analysis/startup_items.txt 2>/dev/null
done

# Check DLL search order hijacking locations
echo "=== DLL HIJACKING CHECK ===" > /cases/case-2024-001/analysis/dll_hijack.txt
# Check for DLLs in application directories that should only be in System32
find /mnt/evidence/Program\ Files/ /mnt/evidence/Program\ Files\ \(x86\)/ \
   -name "*.dll" -newer /mnt/evidence/Windows/System32/ntdll.dll 2>/dev/null \
   >> /cases/case-2024-001/analysis/dll_hijack.txt

# Check for COM object hijacking
python3 << 'PYEOF'
from Registry import Registry

reg = Registry.Registry("/cases/case-2024-001/registry/SOFTWARE")

# Check for suspicious CLSID entries
try:
    clsid = reg.open("Classes\\CLSID")
    for key in clsid.subkeys():
        try:
            server = key.subkey("InprocServer32")
            dll_path = server.value("(default)").value()
            if any(s in dll_path.lower() for s in ['temp', 'appdata', 'programdata', 'downloads', 'tmp']):
                print(f"SUSPICIOUS COM: {key.name()} -> {dll_path}")
        except:
            pass
except:
    pass
PYEOF

# Check boot configuration for bootkits
# BCD (Boot Configuration Data)
ls -la /mnt/evidence/Boot/BCD 2>/dev/null
# Check for modified bootmgr or winload.exe
sha256sum /mnt/evidence/Windows/System32/winload.exe 2>/dev/null

Step 4: Check Linux Persistence Mechanisms

# If analyzing a Linux system
LINUX_ROOT="/mnt/evidence"

echo "=== LINUX PERSISTENCE CHECK ===" > /cases/case-2024-001/analysis/linux_persistence.txt

# Cron jobs
echo "--- Cron Jobs ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
cat $LINUX_ROOT/etc/crontab >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null
ls -la $LINUX_ROOT/etc/cron.d/ >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null
cat $LINUX_ROOT/etc/cron.d/* >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null
cat $LINUX_ROOT/var/spool/cron/crontabs/* >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# Systemd services
echo "--- Custom Systemd Services ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
find $LINUX_ROOT/etc/systemd/system/ -name "*.service" -not -type l \
   >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# SSH authorized keys
echo "--- SSH Authorized Keys ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
find $LINUX_ROOT/home/ $LINUX_ROOT/root/ -name "authorized_keys" -exec cat {} \; \
   >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# Init scripts and rc.local
echo "--- RC Scripts ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
cat $LINUX_ROOT/etc/rc.local >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# Shell profile scripts
echo "--- Profile Scripts ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
cat $LINUX_ROOT/etc/profile.d/*.sh >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# LD_PRELOAD
echo "--- LD_PRELOAD ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
cat $LINUX_ROOT/etc/ld.so.preload >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null
grep -r "LD_PRELOAD" $LINUX_ROOT/etc/ >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# Kernel modules
echo "--- Loaded Kernel Modules ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
cat $LINUX_ROOT/etc/modules-load.d/*.conf >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

# PAM backdoors
echo "--- PAM Configuration ---" >> /cases/case-2024-001/analysis/linux_persistence.txt
find $LINUX_ROOT/etc/pam.d/ -exec grep -l "pam_exec\|pam_script" {} \; \
   >> /cases/case-2024-001/analysis/linux_persistence.txt 2>/dev/null

Step 5: Compile Persistence Report

# Generate comprehensive persistence report
python3 << 'PYEOF'
import json

with open('/cases/case-2024-001/analysis/registry_persistence.json') as f:
    reg_data = json.load(f)

report = """
MALWARE PERSISTENCE INVESTIGATION REPORT
==========================================

PERSISTENCE MECHANISMS FOUND:

1. REGISTRY RUN KEYS:
"""

run_keys = [e for e in reg_data['registry_persistence'] if 'Run' in e.get('type', '')]
for entry in run_keys:
    report += f"   [{entry['timestamp']}] {entry.get('name', 'N/A')} -> {entry.get('value', '')[:100]}\n"

services = [e for e in reg_data['registry_persistence'] if e.get('type') == 'Service']
report += f"\n2. SERVICES ({len(services)} auto-start services):\n"
for entry in services[:20]:
    report += f"   {entry['location'].split('\\')[-1]}: {entry['value'][:100]}\n"

report += """
3. SCHEDULED TASKS: [See scheduled_tasks analysis]
4. WMI SUBSCRIPTIONS: [See WMI analysis]
5. STARTUP FOLDER: [See startup_items.txt]
6. COM HIJACKING: [See COM analysis]

SUSPICIOUS ENTRIES REQUIRING INVESTIGATION:
"""

# Flag suspicious entries
for entry in reg_data['registry_persistence']:
    value = str(entry.get('value', '')).lower()
    suspicious_indicators = ['powershell', 'cmd /c', 'wscript', 'certutil',
                             'programdata', 'appdata\\local\\temp', 'base64',
                             '.ps1', '.vbs', '.hta', '/tmp/', 'hidden']
    if any(s in value for s in suspicious_indicators):
        report += f"   SUSPICIOUS: {entry.get('name', 'N/A')} -> {entry.get('value', '')[:100]}\n"

with open('/cases/case-2024-001/analysis/persistence_report.txt', 'w') as f:
    f.write(report)

print(report)
PYEOF

Key Concepts

Concept Description
Run keys Registry keys executing programs at user logon (HKLM and HKCU)
Scheduled tasks Windows Task Scheduler entries that execute on triggers (time, event, logon)
WMI event subscriptions Persistent WMI queries that trigger actions (stealthy persistence)
COM hijacking Redirecting COM object loading to execute malicious DLLs
DLL search order hijacking Placing malicious DLLs in directories searched before System32
Service persistence Installing Windows services that auto-start with the system
Boot-level persistence Modifying boot configuration or MBR/VBR for pre-OS execution
Living-off-the-land Using legitimate system tools (PowerShell, WMI, certutil) for persistence

Tools & Systems

Tool Purpose
Autoruns Sysinternals comprehensive autostart enumeration tool
RegRipper Automated registry persistence artifact extraction
KAPE Automated persistence artifact collection and analysis
Velociraptor Endpoint agent with persistence hunting artifacts
OSQuery SQL-based system querying for persistence enumeration
PersistenceSniper PowerShell tool for Windows persistence detection
RECmd Eric Zimmerman registry command-line analysis tool
Volatility Memory forensics for in-memory only persistence

Common Scenarios

Scenario 1: APT Persistence After Initial Compromise Check all registry Run keys, enumerate scheduled tasks for encoded PowerShell commands, examine WMI event subscriptions for event-triggered execution, check COM object registrations for hijacked CLSIDs, review services for recently installed entries with suspicious image paths.

Scenario 2: Ransomware Pre-Encryption Persistence Identify how the ransomware maintains access for re-encryption or monitoring, check for scheduled tasks that would re-launch encryption, examine services installed by the ransomware operator, verify no additional backdoor persistence exists before declaring remediation complete.

Scenario 3: Fileless Malware Persistence Focus on registry-based persistence storing payload in registry values, check WMI subscriptions executing PowerShell from event triggers, examine scheduled tasks using encoded command arguments, check for mshta/rundll32 based persistence loading remote content.

Scenario 4: Post-Remediation Verification Run Autoruns comparison against known-good baseline, verify all identified persistence mechanisms have been removed, check for additional persistence that may have been missed, confirm services, tasks, and registry entries are clean, monitor for re-infection indicators.

Output Format

Persistence Investigation Summary:
  System: DESKTOP-ABC123 (Windows 10 Pro)
  Analysis Date: 2024-01-20

  Persistence Mechanisms Found:
    Registry Run Keys (HKLM):    5 entries (1 SUSPICIOUS)
    Registry Run Keys (HKCU):    3 entries (1 SUSPICIOUS)
    Services (Auto-Start):       142 entries (2 SUSPICIOUS)
    Scheduled Tasks:             67 entries (3 SUSPICIOUS)
    WMI Subscriptions:           1 entry (SUSPICIOUS)
    Startup Folder:              4 items (1 SUSPICIOUS)
    COM Objects:                 0 hijacked entries
    DLL Hijacking:               0 detected

  Suspicious Entries:
    1. HKCU\Run\WindowsUpdate -> powershell -ep bypass -e <base64>
       Timestamp: 2024-01-15 14:35:00
       Action: Encoded PowerShell download cradle

    2. Service: WinDefenderUpdate -> C:\ProgramData\svc\update.exe
       Timestamp: 2024-01-15 14:40:00
       Action: Unknown executable in ProgramData

    3. Task: \Microsoft\Windows\Maintenance\SecurityUpdate
       Command: cmd.exe /c powershell -w hidden -e <base64>
       Trigger: On system startup

    4. WMI: __EventFilter "ProcessStart" -> CommandLineEventConsumer
       Action: Execute C:\Windows\Temp\svc.exe on WMI event

  Remediation Required: 4 persistence mechanisms to remove
  Report: /cases/case-2024-001/analysis/persistence_report.txt

04 同仓库更多 Skills

mukul975
mukul975/Anthropic-Cybersecurity-Skills

verifying-build-provenance-with-slsa-sigstore

Verifies artifact signatures and SLSA provenance using Sigstore's cosign (verify, verify-attestation, verify-blob-attestation) and slsa-verifier (verify-artifact), enforcing keyless OIDC builder identity and source repo against SLSA Build levels. Use in CI/CD before deploying artifacts, when consuming third-party attestations, establishing a SLSA Build L3 pipeline, or confirming provenance during incident response or admission control.

Skill与智能体开发
★ 3.6w
mukul975
mukul975/Anthropic-Cybersecurity-Skills

validating-tpm-measured-boot-attestation

Verifies TPM 2.0 measured-boot integrity and remote attestation with tpm2-tools -- reading PCRs (tpm2_pcrread), replaying the boot event log, generating and checking signed quotes (tpm2_quote/tpm2_checkquote), and sealing secrets to a PCR policy. Use to confirm a system booted trusted firmware/kernel for Zero Trust device posture, detect boot-chain tampering via PCR drift, or build a golden-value baseline for fleet attestation.

Skill与智能体开发
★ 3.6w
mukul975
mukul975/Anthropic-Cybersecurity-Skills

validating-backup-integrity-for-recovery

Validates backup integrity through cryptographic hash verification, automated restore testing, corruption detection, and recoverability checks to confirm backups are reliable for disaster recovery and ransomware response. Use before relying on backups for recovery, when building post-backup validation pipelines, auditing backup infrastructure for compliance, or checking immutable/air-gapped backups for silent corruption or tampering.

测试与质量Skill与智能体开发
★ 3.6w
mukul975
mukul975/Anthropic-Cybersecurity-Skills

triaging-windows-with-kape

Runs KAPE (Kroll Artifact Parser and Extractor) to collect targeted forensic artifacts (registry hives, $MFT, event logs, prefetch, browser data) via Targets and parse them with Modules wrapping Eric Zimmerman's EZ Tools (PECmd, MFTECmd, RECmd). Use during early incident containment/triage when full disk imaging is impractical but a defensible, parseable Windows artifact set is needed quickly, including at-scale remote collection.

Skill与智能体开发
★ 3.6w
mukul975
mukul975/Anthropic-Cybersecurity-Skills

triaging-vulnerabilities-with-ssvc-framework

Triages and prioritizes vulnerabilities with CISA's Stakeholder-Specific Vulnerability Categorization (SSVC) decision tree, weighing exploitation status (via the CISA KEV catalog and FIRST EPSS API), technical impact, automatability, and mission prevalence to output Track/Track*/Attend/Act decisions. Use when prioritizing vulnerability scan results (OpenVAS, Nessus, Qualys) for remediation planning beyond raw CVSS scores.

Skill与智能体开发
★ 3.6w
mukul975
mukul975/Anthropic-Cybersecurity-Skills

triaging-security-incident

Performs initial triage of security incidents using the NIST SP 800-61r3 and SANS PICERL frameworks, classifying incident type, assigning priority by business impact, and routing to the appropriate response team. Use when a SIEM/EDR alert needs human classification, concurrent alerts must be prioritized, or a user report or threat-intel IOC match requires initial incident categorization.

Skill与智能体开发
★ 3.6w