Custom Skills

Custom skills extend Cyberstrike with your own slash commands, automating common workflows and tasks.

📸 SCREENSHOT: custom-skills.png

Custom skill definition

Overview

Custom skills allow you to:

Create reusable prompts
Define workflow automations
Share team conventions
Standardize security tests

Skill Definition

Basic Skill

---
name: quick-scan
description: Run a quick security scan on a target
args:
  - name: target
    description: Target URL or IP
    required: true
---

Perform a quick security scan on {{target}}:

1. Check if the target is reachable
2. Identify open ports (top 100)
3. Detect web technologies
4. Run basic vulnerability checks
5. Summarize findings

Focus on speed over thoroughness.

Usage

/quick-scan example.com

Skill Configuration

Frontmatter Options

Field	Type	Description
`name`	string	Command name (required)
`description`	string	Help text description
`args`	array	Command arguments
`model`	string	Override model for skill
`agent`	string	Override agent for skill
`permission`	string	Permission mode

Arguments

args:
  - name: target
    description: Target to scan
    required: true
    type: string

  - name: depth
    description: Scan depth
    required: false
    default: "standard"
    type: string
    choices:
      - quick
      - standard
      - deep

Template Variables

---
name: scan
args:
  - name: target
  - name: ports
    default: "1-1000"
---

Scan {{target}} on ports {{ports}}.

Skill Examples

Full Web Scan

---
name: full-web-scan
description: Comprehensive web application security scan
args:
  - name: url
    description: Target URL
    required: true
agent: web-application
---

Perform a comprehensive security assessment on {{url}}:

## Phase 1: Reconnaissance
1. Identify subdomains
2. Enumerate technologies
3. Map attack surface

## Phase 2: Vulnerability Scanning
1. Test for OWASP Top 10
2. Check for misconfigurations
3. Identify outdated components

## Phase 3: Authentication Testing
1. Test login mechanisms
2. Check session management
3. Verify access controls

## Phase 4: Reporting
1. Compile all findings
2. Prioritize by severity
3. Provide remediation guidance

Document all findings with evidence.

Code Review

---
name: code-review
description: Security-focused code review
args:
  - name: path
    description: File or directory to review
    required: true
---

Perform a security code review on {{path}}:

Focus areas:
1. **Injection vulnerabilities**
   - SQL injection
   - Command injection
   - XSS

2. **Authentication/Authorization**
   - Credential handling
   - Session management
   - Access controls

3. **Data Protection**
   - Sensitive data exposure
   - Encryption usage
   - Secret management

4. **Error Handling**
   - Information leakage
   - Proper logging
   - Fail-safe defaults

Report findings with:
- File and line number
- Severity rating
- Code snippet
- Remediation recommendation

Network Recon

---
name: network-recon
description: Network reconnaissance and enumeration
args:
  - name: target
    description: Network range or host
    required: true
  - name: depth
    default: standard
agent: internal-network
---

Perform network reconnaissance on {{target}}:

{% if depth == "quick" %}
Quick scan:
- Host discovery
- Top 100 ports
- Service detection
{% elif depth == "deep" %}
Deep scan:
- Full port scan
- Service enumeration
- OS detection
- Vulnerability scanning
- Banner grabbing
{% else %}
Standard scan:
- Host discovery
- Top 1000 ports
- Service detection
- Basic vulnerability checks
{% endif %}

Document:
- Live hosts
- Open ports and services
- Potential vulnerabilities
- Recommended next steps

Bug Bounty Recon

---
name: bb-recon
description: Bug bounty reconnaissance workflow
args:
  - name: domain
    required: true
agent: bug-hunter
---

Bug bounty reconnaissance for {{domain}}:

## Subdomain Enumeration
1. Passive sources (crt.sh, SecurityTrails)
2. DNS brute force
3. Permutation scanning

## Asset Discovery
1. Probe live hosts
2. Identify technologies
3. Screenshot endpoints

## Content Discovery
1. Directory enumeration
2. Parameter mining
3. JavaScript analysis
4. Wayback Machine

## Vulnerability Identification
1. Subdomain takeover checks
2. Open redirect testing
3. CORS misconfiguration
4. Information disclosure

Store all findings in memory for future reference.

Skill Organization

Directory Structure

.cyberstrike/
└── skills/
    ├── recon/
    │   ├── subdomain.md
    │   ├── port-scan.md
    │   └── tech-detect.md
    ├── testing/
    │   ├── sqli.md
    │   ├── xss.md
    │   └── auth.md
    └── reporting/
        ├── executive.md
        └── technical.md

Namespaced Skills

/recon/subdomain example.com
/testing/sqli https://target.com/page?id=1
/reporting/executive

Global Skills

Store skills globally for all projects:

~/.cyberstrike/skills/
├── my-quick-scan.md
└── my-report.md

Conditional Logic

If/Else

{% if args.verbose %}
Provide detailed output with all evidence.
{% else %}
Provide concise summary.
{% endif %}

Loops

{% for check in ["sqli", "xss", "idor"] %}
- Test for {{check}}
{% endfor %}

Including Other Skills

Chain Skills

---
name: full-assessment
---

Run the following in sequence:

1. First: /network-recon {{target}}
2. Then: /full-web-scan {{target}}
3. Finally: /reporting/executive

Compile all findings into a comprehensive report.

Skill Variables

Environment Variables

Scan using API key: {{env.API_KEY}}

Session Variables

Continue scanning {{session.target}} from previous findings.

Config Variables

Using model: {{config.model}}

Publishing Skills

# Copy to shared location
cp -r .cyberstrike/skills/* /shared/cyberstrike-skills/

# Or use git
git add .cyberstrike/skills/
git commit -m "Add security testing skills"

npm Package

{
  "name": "@team/cyberstrike-skills",
  "cyberstrike": {
    "skills": [
      "skills/*.md"
    ]
  }
}

Best Practices

Clear descriptions - Help users understand what skills do
Sensible defaults - Provide good defaults for optional args
Consistent naming - Use clear, descriptive names
Documentation - Include examples in skill description
Modularity - Create focused, composable skills
Error handling - Account for edge cases

Tip

Start with simple skills and gradually add complexity as needed.

Slash Commands Overview - All commands
Agents - Agent configuration
Configuration - Skill settings