Skip to content

Skills Library

Overwatch includes 34 offensive methodology guides searchable via TF-IDF through the get_skill tool.

Available Skills

Network & Infrastructure

Skill File Description
Network Recon network-recon.md Host discovery, port scanning, service enumeration
DNS Enumeration dns-enumeration.md DNS zone transfers, subdomain enumeration
SNMP Enumeration snmp-enumeration.md SNMP community string brute-force, MIB walking
SMB Enumeration smb-enumeration.md Share enumeration, null sessions, user listing
NetExec netexec.md Multi-protocol enumeration and exploitation with NXC
Subnet Enumeration subnet-enumeration.md Per-CIDR agent methodology for parallel subnet sweeps

Active Directory

Skill File Description
AD Discovery ad-discovery.md AD enumeration, BloodHound, trusts, delegation, ADCS
Kerberoasting kerberoasting.md SPN enumeration, TGS cracking
ADCS Exploitation adcs-exploitation.md Certificate abuse (ESC1–ESC8)
Domain Trust Attacks domain-trust-attacks.md Cross-domain and forest trust abuse
AD Persistence ad-persistence.md Golden/Silver tickets, skeleton key, DCShadow
SCCM Attacks sccm-attacks.md SCCM/MECM credential harvesting and lateral movement
Exchange Attacks exchange-attacks.md ProxyLogon, ProxyShell, mailbox access
MSSQL Attacks mssql-attacks.md xp_cmdshell, linked servers, impersonation, credential extraction

Credentials

Skill File Description
Password Spraying password-spraying.md Domain and local password spraying
Credential Dumping credential-dumping.md LSASS, SAM, NTDS.dit, DCSync
SMB Relay smb-relay.md NTLM relay attacks, coercion methods

Lateral Movement & Pivoting

Skill File Description
Lateral Movement lateral-movement.md WMI, PSExec, WinRM, DCOM, RDP
Pivoting pivoting.md SSH tunnels, SOCKS proxies, port forwarding

Container & CI/CD

Skill File Description
Container & Kubernetes container-attacks.md Pod escape, RBAC abuse, service account theft, etcd access
CI/CD Exploitation cicd-exploitation.md Jenkins, GitLab CI, GitHub Actions — secret extraction, runner abuse

Web Application

Skill File Description
Web Discovery web-discovery.md Web app fingerprinting, directory enumeration
Web App Attacks web-app-attacks.md Authentication bypass, file upload, SSRF
Web Vuln Scanning web-vuln-scanning.md Automated vulnerability scanning
CMS Exploitation cms-exploitation.md WordPress, Joomla, Drupal
SQL Injection sql-injection.md SQL injection detection and exploitation

Privilege Escalation

Skill File Description
Windows Privilege Escalation privilege-escalation.md Service misconfigs, token abuse, UAC bypass
Linux Enumeration linux-enumeration.md System info, SUID, cron, capabilities
Linux Privilege Escalation linux-privesc.md Kernel exploits, sudo abuse, Docker escape

Cloud

Skill File Description
AWS Exploitation aws-exploitation.md IAM abuse, metadata service, S3 misconfiguration
Azure Exploitation azure-exploitation.md Azure AD, managed identities, storage
GCP Exploitation gcp-exploitation.md Service accounts, metadata, storage buckets

Post-Exploitation

Skill File Description
Persistence persistence.md Startup scripts, scheduled tasks, services
Data Exfiltration data-exfiltration.md File staging, transfer, and exfil techniques

Searching Skills

The get_skill tool uses TF-IDF search. Tips for effective queries:

  • Use specific technique names: "kerberoast", "ntlm relay", "adcs esc1"
  • Use service names: "smb", "ldap", "mssql", "http", "jenkins", "kubernetes"
  • Use attack patterns: "privilege escalation", "lateral movement", "credential dumping"

Writing Custom Skills

Skills are markdown files in the skills/ directory following this template:

# Skill Name

tags: keyword1, keyword2, keyword3

## Objective
What this skill accomplishes.

## Prerequisites
What's needed before using this skill.

## Methodology
Step-by-step approach with exact commands.

## Reporting
What to report via report_finding — node types, edge types, properties.

## OPSEC Notes
Noise considerations and stealth alternatives.

Best Practices

  • Tags improve search ranking — use specific terms the LLM might search for
  • Commands should be exact and copy-pasteable
  • OPSEC Notes should include noise ratings (0.0–1.0 scale)
  • Reporting sections should reference report_finding format with specific node/edge types
  • Detection signatures help the LLM understand risk before executing