Skip to main content

Cyberstrike is now open source! AI-powered penetration testing for security professionals. Star on GitHub

Bug Hunter Agent

The Bug Hunter agent is a specialized security testing agent equipped with comprehensive knowledge of reconnaissance methodology, attack surface discovery, and bug bounty hunting techniques.

Overview

The Bug Hunter agent combines deep reconnaissance knowledge with automated tooling to maximize attack surface discovery for bug bounty programs. It’s pre-loaded with specialized skills for subdomain enumeration, content discovery, and vulnerability pattern detection.

Agent Configuration

PropertyValue
Agent IDbug-hunter
ColorGreen
Pre-loaded Skillsrecon-methodology, owasp-wstg, subdomain-takeover, nuclei-templates
Allowed ToolsBrowser, Bash, WebFetch, WebSearch, Read, Glob, Grep, Memory

Launch the Agent

Terminal window
# Start Cyberstrike with bug hunter agent
cyberstrike --agent bug-hunter
# Or switch during session
/agent bug-hunter

Knowledge Base

The Bug Hunter agent has access to extensive security knowledge through the skills system.

Recon Methodology Skill

The agent is pre-loaded with comprehensive reconnaissance methodology including:

  • Passive reconnaissance techniques
  • Active enumeration methods
  • Asset discovery workflows
  • Technology fingerprinting
  • Content discovery strategies
  • Parameter mining techniques

OWASP WSTG Integration

Full OWASP Web Security Testing Guide v4.2 knowledge for systematic vulnerability testing after reconnaissance is complete.

Subdomain Takeover Skill

Complete subdomain takeover detection including:

  • 70+ vulnerable fingerprints
  • Cloud provider patterns (AWS, Azure, GCP)
  • Third-party service patterns
  • DNS configuration analysis
  • CNAME chain analysis

Nuclei Templates Skill

Knowledge of Nuclei template categories and usage:

  • CVE detection templates
  • Exposure detection
  • Misconfiguration checks
  • Takeover detection
  • Technology fingerprinting

Reconnaissance Methodology

The agent follows a structured multi-phase reconnaissance approach:

Phase 1: Passive Reconnaissance (15 techniques)

Technique IDTechnique NameData Sources
RECON-01Certificate Transparencycrt.sh, Censys, Google CT
RECON-02DNS HistorySecurityTrails, ViewDNS
RECON-03Reverse DNSPTR records, IP ranges
RECON-04ASN EnumerationBGP data, WHOIS
RECON-05WHOIS AnalysisDomain registration data
RECON-06Search Engine DorkingGoogle, Bing, Yandex
RECON-07Archive MiningWayback Machine, Common Crawl
RECON-08GitHub ReconCode leaks, secrets, endpoints
RECON-09S3 Bucket EnumerationBucket naming patterns
RECON-10JavaScript AnalysisEndpoints, API keys, secrets
RECON-11Social Media OSINTLinkedIn, Twitter, employee info
RECON-12Job Posting AnalysisTechnology stack discovery
RECON-13Public Breach DataCredential leaks, email patterns
RECON-14Shodan/Censys QueriesInternet-wide scan data
RECON-15Cloud Resource DiscoveryAzure, AWS, GCP resources

Phase 2: Active Enumeration (12 techniques)

Technique IDTechnique NameTools
ENUM-01DNS Brute Forceffuf, gobuster, amass
ENUM-02Virtual Host Discoveryffuf, gobuster
ENUM-03Port Scanningnmap, masscan, naabu
ENUM-04Service Fingerprintingnmap, httpx
ENUM-05Technology Detectionwappalyzer, whatweb
ENUM-06Directory Bruteforceffuf, feroxbuster
ENUM-07Parameter Discoveryarjun, paramspider
ENUM-08API Endpoint Discoverykiterunner, ffuf
ENUM-09Favicon Hash Analysisfavfreak, shodan
ENUM-10HTTP Header Analysiscurl, httpx
ENUM-11Robots.txt/Sitemap MiningCustom scripts
ENUM-12JavaScript Endpoint ExtractionLinkFinder, JSParser

Phase 3: Vulnerability Discovery (10 patterns)

Pattern IDVulnerability PatternDetection Method
VULN-01Subdomain TakeoverDNS CNAME analysis
VULN-02Open RedirectParameter fuzzing
VULN-03CORS MisconfigurationHeader analysis
VULN-04SSRF PatternsParameter testing
VULN-05IDOR IndicatorsResponse comparison
VULN-06Information DisclosurePattern matching
VULN-07Debug EndpointsPath discovery
VULN-08Default CredentialsCredential testing
VULN-09Version DisclosureBanner grabbing
VULN-10Security Header MissingHeader analysis

Subdomain Takeover Detection

The agent includes comprehensive subdomain takeover knowledge:

Vulnerable Fingerprints (70+)

ServiceFingerprintTakeover Possible
AWS S3NoSuchBucketYes
AWS CloudFrontBad RequestYes
AzureNXDOMAIN on *.azurewebsites.netYes
GitHub Pages404 There isn’t a GitHub Pages siteYes
HerokuNo such appYes
ShopifySorry, this shop is currently unavailableYes
TumblrThere’s nothing hereYes
WordPress.comDo you want to registerYes
GhostThe thing you were looking for is no longer hereYes
Surge.shproject not foundYes
FastlyFastly error: unknown domainYes
Pantheon404 error unknown siteYes
ZendeskHelp Center ClosedConditional
Readme.ioProject doesnt existYes
BitbucketRepository not foundYes
IntercomThis page is reserved forConditional
Cargo404 Not FoundYes
StatusPageStatus page launchedConditional
Uptimerobotpage not foundYes
VercelThe deployment could not be foundYes

Tip

Always verify takeover possibility manually before reporting. Some fingerprints may produce false positives.

Takeover Detection Example

> Check subdomains for takeover vulnerabilities
[Checking 147 subdomains...]
FINDING: Subdomain Takeover - AWS S3
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Subdomain: assets.target.com
CNAME: assets.target.com.s3.amazonaws.com
Status: NoSuchBucket error
Severity: High (CVSS 8.6)
Bounty Estimate: $500-$3,000
Claim Instructions:
1. Create S3 bucket: assets.target.com
2. Upload proof of ownership file
3. Report to program
FINDING: Subdomain Takeover - GitHub Pages
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Subdomain: docs.target.com
CNAME: target.github.io
Status: 404 - No GitHub Pages site
Severity: High (CVSS 8.6)
Bounty Estimate: $300-$1,500
Claim Instructions:
1. Create GitHub repo: target/target.github.io
2. Enable GitHub Pages
3. Add CNAME file with docs.target.com

Content Discovery Methodology

Wordlist Strategy

The agent uses optimized wordlist combinations:

Wordlist TypePurposeSize
CommonFrequently seen paths10K
Technology-specificFramework paths (Rails, Django, etc.)5K
APIREST/GraphQL endpoints15K
Backup files.bak, .old, .backup2K
Config files.env, .config, web.config1K
CustomGenerated from JS analysisDynamic

Recursive Discovery

> Perform recursive content discovery on target.com
[Phase 1: Initial discovery...]
Found 234 paths
[Phase 2: Recursive on interesting directories...]
/api/ → 47 new endpoints
/admin/ → 23 new paths
/internal/ → 12 new paths
[Phase 3: Technology-specific...]
Detected: Laravel
Running Laravel-specific wordlist...
Found: /telescope, /horizon, /log-viewer
FINDING: Debug Interface Exposed
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
URL: https://target.com/telescope
Technology: Laravel Telescope
Access: Unauthenticated
Severity: High (CVSS 7.5)
Evidence: Full request/response logging visible
CWE: CWE-215
Bounty Estimate: $500-$2,000

Parameter Mining

Sources for Parameters

SourceMethodExample
URL HistoryWayback Machine?id=, ?user=
JavaScriptStatic analysisapiKey, token
HTML FormsDOM parsingusername, password
API DocsSwagger/OpenAPIAll documented params
Traffic AnalysisProxy captureHidden parameters

Parameter Testing Workflow

> Discover and test parameters on target.com
[Mining parameters...]
URL Parameters: 47 unique
POST Parameters: 23 unique
Headers: 8 custom
[High-value parameters found...]
?redirect_url → Potential Open Redirect
?file → Potential LFI/Path Traversal
?url → Potential SSRF
?id → Potential IDOR
?debug → Potential Info Disclosure
[Testing redirect_url...]
FINDING: Open Redirect
━━━━━━━━━━━━━━━━━━━━━
URL: https://target.com/login?redirect_url=
Parameter: redirect_url
Payload: //evil.com
Severity: Medium (CVSS 5.4)
Evidence: Redirects to attacker-controlled domain
CWE: CWE-601
Bounty Estimate: $100-$500

Testing Methodology

The agent follows a structured approach:

┌─────────────────────────────────────────────────────────────┐
│ TESTING METHODOLOGY │
├─────────────────────────────────────────────────────────────┤
│ 1. SCOPE ANALYSIS │
│ └── Parse program policy, understand boundaries │
│ │
│ 2. PASSIVE RECON │
│ └── Certificate transparency, OSINT, archives │
│ │
│ 3. ACTIVE ENUMERATION │
│ └── Subdomain brute, port scan, content discovery │
│ │
│ 4. FINGERPRINTING │
│ └── Technology stack, version detection │
│ │
│ 5. VULNERABILITY DISCOVERY │
│ └── Pattern matching, fuzzing, template scanning │
│ │
│ 6. EXPLOITATION & PoC │
│ └── Demonstrate impact, document steps │
│ │
│ 7. REPORTING │
│ └── Write clear report, estimate bounty │
└─────────────────────────────────────────────────────────────┘

Integrated Security Tools

The agent leverages industry-standard bug bounty tools:

ToolPurpose
SubfinderPassive subdomain enumeration
AmassComprehensive asset discovery
httpxHTTP probing and fingerprinting
NucleiTemplate-based vulnerability scanning
ffufWeb fuzzing
ArjunHTTP parameter discovery
gauURL discovery from archives
WaybackurlsHistorical URL mining
KatanaCrawling and spidering
DalfoxXSS detection

Tool Chain Example

> Run full recon workflow on target.com
[Step 1: Subdomain enumeration...]
subfinder -d target.com -all -o subs.txt
[Step 2: DNS resolution and probing...]
cat subs.txt | httpx -silent -o live.txt
[Step 3: Port scanning on live hosts...]
naabu -l live.txt -top-ports 1000 -o ports.txt
[Step 4: Technology fingerprinting...]
cat live.txt | httpx -tech-detect -o tech.txt
[Step 5: Content discovery...]
ffuf -w wordlist.txt -u FUZZ -o content.txt
[Step 6: Nuclei scanning...]
nuclei -l live.txt -t nuclei-templates/ -o vulns.txt
Summary:
━━━━━━━━
Subdomains: 147
Live hosts: 89
Open ports: 234
Technologies: 23 unique
Directories: 1,247
Vulnerabilities: 12 potential

Nuclei Template Categories

CategoryTemplatesCoverage
cves3,000+Known CVE exploits
exposures500+Sensitive data exposure
misconfiguration400+Server misconfigurations
takeovers70+Subdomain takeover
default-logins200+Default credential checks
technologies600+Technology detection
file300+Sensitive file detection
fuzzing100+Parameter fuzzing
headless50+Browser-based checks

Nuclei Scanning Example

> Run Nuclei vulnerability scan
[Loading templates...]
Loaded: 4,847 templates
[Scanning 89 live hosts...]
FINDING: CVE-2023-44487 (HTTP/2 Rapid Reset)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Template: CVE-2023-44487
Host: api.target.com
Severity: High (CVSS 7.5)
Evidence: Server vulnerable to rapid reset
CWE: CWE-400
FINDING: Exposed .git Directory
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Template: git-config-exposure
Host: staging.target.com/.git/config
Severity: High (CVSS 7.5)
Evidence: Git config file accessible
CWE: CWE-200
FINDING: AWS Credentials in JavaScript
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Template: aws-access-key-exposure
Host: target.com/assets/app.js
Severity: Critical (CVSS 9.8)
Evidence: AKIA... pattern found
CWE: CWE-798

CVSS Severity Rating

The agent rates findings using CVSS 3.1 and bounty estimates:

SeverityCVSS ScoreExample FindingsBounty Range
Critical9.0 - 10.0RCE, Auth bypass, exposed creds$5,000-$50,000
High7.0 - 8.9Subdomain takeover, SSRF, SQLi$1,000-$10,000
Medium4.0 - 6.9XSS, Open redirect, IDOR$200-$2,000
Low0.1 - 3.9Info disclosure, missing headers$50-$500

Session Example

You: Hunt for bugs on target.com bug bounty program
Agent: I'll perform comprehensive reconnaissance and vulnerability hunting on target.com.
## Phase 1: Scope Analysis
[Parsing program scope...]
In Scope:
- *.target.com
- api.target.com
- app.target.com
Out of Scope:
- blog.target.com (Wordpress)
- support.target.com (Zendesk)
## Phase 2: Passive Reconnaissance
[Certificate Transparency...]
Found 234 certificates → 156 unique subdomains
[Archive Mining...]
Wayback URLs: 12,847
Interesting params: redirect, file, url, callback
[GitHub Recon...]
⚠ FINDING - Exposed API key in public repo
FINDING: Exposed API Key in GitHub
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Repository: target/mobile-app (commit abc123)
File: src/config/api.js
Secret: STRIPE_API_KEY=sk_live_...
Severity: Critical (CVSS 9.8)
Evidence: Active Stripe key found in public commit
CWE: CWE-798
Bounty Estimate: $5,000-$15,000
## Phase 3: Active Enumeration
[Subdomain brute forcing...]
New subdomains: 23
[Live host probing...]
Live hosts: 89/179
[Content discovery on api.target.com...]
⚠ FINDING - Unauthenticated GraphQL introspection
FINDING: GraphQL Introspection Enabled
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
URL: https://api.target.com/graphql
Issue: Full schema accessible via introspection
Severity: Medium (CVSS 5.3)
Evidence: __schema query returns all types/fields
CWE: CWE-200
Bounty Estimate: $300-$800
## Phase 4: Subdomain Takeover Check
[Checking 179 subdomains...]
⚠ FINDING - Takeover possible on staging
FINDING: Subdomain Takeover (Heroku)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Subdomain: staging.target.com
CNAME: target-staging.herokuapp.com
Status: No such app
Severity: High (CVSS 8.6)
Evidence: NXDOMAIN for Heroku app
Bounty Estimate: $1,000-$3,000
[Continuing with remaining phases...]

Report Template

## Summary
[One sentence describing the vulnerability]
## Affected Asset
- **URL:** https://target.com/vulnerable-endpoint
- **Parameter:** redirect_url
- **Method:** GET
## Vulnerability Type
Open Redirect (CWE-601)
## Severity
Medium (CVSS 5.4)
## Steps to Reproduce
1. Navigate to https://target.com/login
2. Append ?redirect_url=//evil.com to the URL
3. Complete login process
4. Observe redirect to evil.com
## Proof of Concept
[Screenshot or video]
## Impact
An attacker can craft a malicious link that appears to be from target.com
but redirects users to a phishing page. This can be used for:
- Credential harvesting
- Malware distribution
- OAuth token theft
## Remediation
1. Validate redirect URLs against a whitelist
2. Use relative URLs only
3. Implement destination verification
## References
- https://cwe.mitre.org/data/definitions/601.html
- https://owasp.org/www-project-web-security-testing-guide/

Best Practices

  1. Read Program Policy - Understand scope, rules, and exclusions
  2. Start Passive - Begin with non-invasive reconnaissance
  3. Track Progress - Use Memory tool to avoid duplicate work
  4. Verify Impact - Demonstrate real security impact
  5. Write Clear Reports - Help triagers reproduce issues
  6. Focus on Unique Issues - Automated findings are often duplicates

Caution

Always respect program scope and rate limits. Aggressive scanning can result in program bans and legal issues.