MCP Server Reference

HailBytes ASM

16 tools for reconnaissance, scan lifecycle, vulnerability intelligence, and OSINT. Endpoint https://<asm-host>/mcp/, auth via X-API-Key.

HailBytes SAT

14 tools for campaigns, templates, landing pages, target groups, and sending profiles. Endpoint https://<sat-host>:3333/api/mcp/, auth via Authorization: Bearer.

1. Generate an API key

In the ASM admin UI: Administration › API Keys › Add Key. The key is shown only once and is SHA-256 hashed at rest.

2. Add the server to your client

Drop the JSON snippet below into your client’s MCP config. Transport is Streamable HTTP; the endpoint is /mcp/ on your ASM host.

3. Talk to it

Ask the agent to scan, triage, or build a target list. It will pick the right tools and chain calls without you naming them.

Claude Desktop

Edit claude_desktop_config.json via Settings › Developer › Edit Config.

{
  "mcpServers": {
    "hailbytes-asm": {
      "type": "streamable-http",
      "url":  "https://hailbytes-asm.yourcompany.com/mcp/",
      "headers": { "X-API-Key": "YOUR_ASM_KEY" }
    }
  }
}

Claude Code (CLI)

Add to .claude/settings.json in a project, or ~/.claude/settings.json globally.

{
  "mcpServers": {
    "hailbytes-asm": {
      "type": "streamable-http",
      "url":  "https://hailbytes-asm.yourcompany.com/mcp/",
      "headers": { "X-API-Key": "YOUR_ASM_KEY" }
    }
  }
}

Cursor / Windsurf

Both editors accept the same Streamable HTTP config shape. Set transport to Streamable HTTP, URL to https://<host>/mcp/, and the X-API-Key header.

Transport: streamable-http
URL:       https://<host>/mcp/
Header:    X-API-Key: <api_key>

Anthropic SDK (Python)

Pass the MCP server inline on a messages.create call.

import anthropic

client = anthropic.Anthropic()
client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=4096,
    mcp_servers=[{
        "type": "url",
        "url":  "https://hailbytes-asm.yourcompany.com/mcp/",
        "name": "hailbytes-asm",
        "headers": {"X-API-Key": "YOUR_ASM_KEY"},
    }],
    messages=[{"role": "user",
               "content": "Scan example.com and tell me what you find."}],
)

Cold-start reconnaissance

You: Add example.com to the "client-pentest" project
     and run a full scan.

Agent:
  add_target(example.com, client-pentest)
  initiate_scan(target_id, "Full Scan")
  → "Scan 42 started. I’ll check status in a few
       minutes."

Triage critical findings

You: Show me critical and high vulns from the latest
     scan of acme.com and suggest where to start.

Agent:
  list_vulnerabilities(severity=4, target=acme.com)
  list_vulnerabilities(severity=3, target=acme.com)
  → "2 SQLi, 1 exposed .git, 1 stored XSS in admin.
       Start with the SQLi on api.acme.com/search."

Build an attack-surface report

You: Give me subdomains, open ports, and tech stack
     for acme.com.

Agent:
  list_subdomains(acme.com)
  list_ports(acme.com)
  list_technologies(acme.com)
  → structured narrative with risky surface flagged.

OSINT → phishing target list

You: Build me a phishing target list for acme.com
     based on what HailBytes found.

Agent:
  list_emails(acme.com)
  list_employees(acme.com)
  → CSV ready to import into a SAT campaign.