Advanced bug bounty reconnaissance and vulnerability scanning MCP server with integrated security tools (subfinder, httpx, nuclei, naabu, katana, dnsx, ffuf, gobuster, gowitness, etc.). Compatible with Claude Desktop, Claude CLI, and Gemini CLI.
24 tools with schemas present but significant quality gaps. Most tools have descriptions (good), but many lack parameter-level descriptions and output schemas are not documented. Naming is generally clear (verb_noun pattern), but several tools combine multiple concerns (e.g., full_recon orchestrates 5 operations). Parameter descriptions are sparse, many lack format/constraint guidance. No evidence of error handling patterns, recovery guidance, or actionable error messages. Security concerns: no visible secret injection for SHODAN_API_KEY, no permission gates on destructive tools (subdomain_brute, dir_fuzz marked WRITE but no auth checks visible). Output structure not documented for most tools.
Probe a list of hosts/URLs with httpx. Returns status codes, titles, web servers, and detected technologies. Identifies interesting endpoints.
Analyze and correlate security findings from multiple scans. Prioritizes issues by severity and business impact. Suggests remediation strategies.
Search for SSL/TLS certificates issued for a domain using crt.sh. Returns certificate details, SANs, and issuance dates.
Verify if a target is within the authorized scope. Checks against ALLOWED_SCOPE and BLOCKED_TARGETS from configuration.
Check for CORS misconfigurations. Tests if the target allows unauthorized cross-origin requests.
Crawl a website with katana to discover URLs, forms, JS files, and API endpoints. Great for finding hidden functionality.
Output schemas not documented. Tools like nuclei_scan, crawl, analyze_findings return complex objects but no schema is visible. LLMs cannot plan downstream calls or extract fields reliably.
Parameter descriptions incomplete. Many parameters lack format/constraint guidance. E.g., 'timeout' appears in 10+ tools with no min/max bounds; 'wordlist' path parameters lack validation hints; 'severity' in nuclei_scan lacks enum values.
No error handling or recovery guidance visible. Tools like shodan_search (requires API key) and nuclei_scan (external service) have no documented error cases, retry logic, or actionable error messages.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | <=2025-11-25 | v2 |
Fuzz directories and files using ffuf. Discovers hidden endpoints, backup files, and admin panels. Returns found paths with status codes.
Enumerate DNS records (A, MX, NS, TXT, CNAME, SOA) for a domain using dnsx.
Get detailed explanation of a security finding: what it means, why it's a problem, how to exploit it, and how to fix it.
Run a complete automated recon pipeline on a domain: subdomain_enum → alive_check → port_scan → tech_detect → screenshot. Best starting point for a new target.
Generate a comprehensive bug bounty report from all findings. Includes executive summary, detailed findings, severity breakdown, and remediation recommendations.
Get intelligent recommendations for next steps in the bug bounty hunt based on current findings and target profile.
List all available tools with descriptions, input parameters, and usage examples.
Run Nuclei vulnerability scanner on target(s). Detects known vulnerabilities using community templates. Returns findings with severity levels and remediation guidance.
Fast TCP port scan using naabu. Discovers open services. Returns port list with service guesses and investigation hints.
Take screenshots of web targets using gowitness. Screenshots saved to /app/screenshots/. Returns file paths + interesting observations.
Search Shodan for devices/services related to a target. Requires SHODAN_API_KEY in .env. Returns IP addresses, open ports, and service banners.
Scan for SQL injection vulnerabilities using Nuclei SQLi templates. Tests common injection points.
Actively brute-force subdomains using a wordlist (ffuf + DNS). More aggressive than passive enum — use only on targets you have permission to test.
Passively enumerate subdomains using subfinder (uses public APIs, crt.sh, etc). Returns subdomains + smart guidance on suspicious patterns and next steps.
Detect web technologies (frameworks, CDNs, CMS, servers) using httpx. Returns technology stack with known vulnerability hints.
Retrieve historical URLs from Wayback Machine + gau. Useful for finding old endpoints, backup files, and forgotten parameters.
Perform WHOIS lookup on a domain. Returns registrar, registrant, nameservers, and registration dates. Useful for identifying domain ownership and infrastructure.
Scan for XSS vulnerabilities using Nuclei XSS templates. Tests common injection points and payloads.
Security: SHODAN_API_KEY exposed via environment but no visible secret injection pattern in tool parameters. No permission gates on destructive tools (subdomain_brute, dir_fuzz marked WRITE). No audit logging visible.
Tool composition: full_recon combines 5 operations (subdomain_enum → alive_check → port_scan → tech_detect → screenshot). Violates single-responsibility principle. Agent cannot compose these independently or skip steps.