io.github.cyanheads/attack-surface-mcp-server

@cyanheads/attack-surface-mcp-server

Passive external attack-surface mapping: CT subdomains, DNS, TLS, HTTP posture, RDAP/WHOIS, Shodan via MCP. STDIO or Streamable HTTP.

8 Tools • 1 Resource

[!IMPORTANT] Authorized, defensive use only. Point this server only at assets you own or are explicitly authorized to assess. It performs passive, non-intrusive reconnaissance — it reads public records (Certificate Transparency logs, DNS, RDAP/WHOIS) and each target's own published surface (one TLS handshake and one HTTP GET per host). It does not port-scan, exploit, brute-force, fuzz, or probe for vulnerabilities; that capability is excluded from the surface by design, not gated behind a flag. Output is descriptive — what exists and what the security posture is — never an exploitation plan. Every outbound connection passes an SSRF guard that refuses private, loopback, link-local, and cloud-metadata targets.

Overview

Passive external attack-surface mapping (EASM) over Certificate Transparency logs, DNS, TLS, HTTP, and RDAP/WHOIS registries, with optional Shodan host intelligence. Discover subdomains, resolve DNS records, and inspect TLS and HTTP security posture across a domain's live hosts.

Tools

Resources

All resource data is also reachable via tools — tool-only clients lose nothing, since attacksurface_map_domain covers the same ground.

Capability reference

attacksurface_map_domain tool

• depth: quick discovers subdomains and liveness; standard adds DNS records, TLS, and HTTP posture; thorough adds per-IP Shodan data when a key is present, otherwise a note.
• Returns a structured surface map and a defensive assessment of observable facts. Failed sources and unreachable hosts become notes, and subdomain resolution is capped by ATTACKSURFACE_MAX_SUBDOMAINS (default 200), with truncation disclosed.
• includeRegistration adds RDAP/WHOIS data for the apex at standard or thorough depth. Standard and thorough also return DNS records from the first configured resolver plus propagation mismatches for live hosts; null posture at quick depth means the work was skipped.
• coverage lists what was left unexamined or without host detail (unexamined, detailOmitted), and retrieval gives up to three ready-to-run follow-up calls (DNS/TLS batches of at most 50 hosts, one HTTP URL). nextToolSuggestions prefills an attacksurface_recon_guidance call for up to 50 hosts, with guidanceCoverage naming any omitted.

attacksurface_enumerate_subdomains tool

• Discovers names through crt.sh → Certspotter → the apex TLS certificate SANs, then resolves DNS liveness. includeUnresolved: false keeps only live hosts; filteredAddressErrors retains DNS failures for excluded names.
• Returns each name's source provenance and per-source status. complete, pagesFetched, and incompleteReason disclose partial coverage, and addressError separates DNS failures from ENODATA/ENOTFOUND absence.
• offset and limit (default 50, larger requests clamped) page through candidates, bounded by min(50, ATTACKSURFACE_MAX_SUBDOMAINS). Use nextOffset or the prefilled continuation for the next page, and remainingCount to see what is left. CT logs are re-read per page, so changes can shift offsets.

attacksurface_resolve_dns tool

• Accepts 1–1000 hosts and processes the first 50 in input order; processedCount and remainingHosts show what to resubmit. Queries A, AAAA, CNAME, MX, NS, TXT, and CAA across public resolvers (default 8.8.8.8, 1.1.1.1, 9.9.9.9); reverse DNS (PTR) is optional.
• Returns per-resolver answers to expose propagation gaps, with canonical records from the first configured resolver. Resolver failures (including SERVFAIL) appear in queryError and hostError next to any successful records.
• Private or loopback resolver IPs produce blocked_resolver; an individual host failure becomes a per-host error.

attacksurface_inspect_tls tool

• Accepts 1–1000 hosts and processes the first 50 in input order; processedCount and remainingHosts show what to resubmit. Port defaults to 443 and handshake timeout to 8000ms (1000–30000ms).
• Returns negotiated protocol, cipher, leaf certificate and chain depth, SANs, issuer, validity window, days to expiry, and validation status. Invalid, expired, and self-signed certificates are reported; one host failure becomes a per-host error.
• Chain trust (validationAuthorized/validationError) and hostname match (identityAuthorized/identityError) are separate checks. daysUntilExpiry is null only when the expiry date cannot be parsed.

attacksurface_probe_http tool

• Accepts one HTTP(S) URL; timeout defaults to 10000ms (1000–30000ms) and covers the whole probe. Follows up to ten redirects with an SSRF check at every hop; a refused target produces blocked_target, and exceeding the redirect limit returns finalStatus: 0 and a transportError.
• Returns status, redirect chain, raw headers, security-header findings (HSTS, CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, Permissions-Policy, cookie flags, CORS reflection), and technology detections with their triggering evidence. CSP analysis covers frame-ancestors only.
• bodyInspection reports complete, limited (64,000-byte prefix cap), failed, or not_read; a body failure keeps the status and headers.

attacksurface_lookup_registration tool

• Accepts a domain, IP, or CIDR (prefix IPv4 0–32, IPv6 0–128); type (auto / domain / ip) controls interpretation. Uses RDAP first, then WHOIS when needed.
• Domain results include registrar, EPP statuses, lifecycle events, nameservers, and DNSSEC; IP/CIDR results include netblock, allocation CIDRs, ASN, and country. Sparse or redacted fields stay unknown, and the raw WHOIS response is kept in rawWhois.
• An explicit WHOIS no-match returns no_match; a failed referral or an empty or unrecognized WHOIS response returns source_unavailable.

attacksurface_lookup_host tool

• mode: "host" (default) looks up one bare IPv4/IPv6 address and returns invalid_target for anything else; mode: "search" takes an internet-wide query with optional facets and may consume Shodan query credits. Requires SHODAN_API_KEY.
• Search returns one 100-match page per call (page defaults to 1, ignored in host mode), with hasMore, nextPage, and a prefilled continuation. Later pages are opt-in and may consume credits.
• Returns ports, banners, software versions, hostnames, ASN, and geography from Shodan's last scan. Missing credentials produce source_unavailable; no host data produces no_data. The server performs no port scans.

attacksurface_recon_guidance tool

• Takes prior findings (hosts, certificates, missing headers, software versions, ports). Topic triage covers all findings ordered by urgency, posture covers TLS/header hardening, and coverage covers inspection and version gaps. Open ports are inventory evidence, never proof of a vulnerability.
• Returns priorities plus suggestions as {toolName, reason, args} that target local tools only; no external calls or active scans are emitted.
• priorityOffset and suggestionOffset page independently with limit (default 50, max 50). Resubmit the original findings and topic with the next offsets.

attacksurface://surface/{domain} resource

• Takes a domain and returns an application/json snapshot equivalent to a standard-depth attacksurface_map_domain call.
• Includes subdomain and live-host counts plus per-host DNS/TLS/HTTP posture and errors. Resolution uses the same cap as mapping, and posture detail covers at most 50 live hosts; coverage lists what was omitted with follow-up calls.

Features

Built on @cyanheads/mcp-ts-core: stdio and Streamable HTTP transports, pluggable auth (none / jwt / oauth), swappable storage (in-memory, filesystem, Supabase, Cloudflare KV/R2/D1), structured logging with optional OpenTelemetry tracing.

Attack-surface-specific:

• Passive and non-intrusive by mandate — public records plus each target's own single published response; active scanning, exploitation, and brute-forcing are excluded from the surface, not toggled by a flag
• Keyless core — CT subdomain enumeration, DNS, TLS, HTTP/tech, and RDAP/WHOIS all work with zero API keys; Shodan is strictly additive depth
• SSRF guard on every outbound connection — rejects private, loopback, link-local, cloud-metadata, and reserved IPv4/IPv6 ranges before connecting (opt out for trusted internal assessment via ATTACKSURFACE_ALLOW_PRIVATE_TARGETS)
• Multi-source aggregation with fallback chains — CT discovery falls through crt.sh → Certspotter → TLS-SAN; registration falls through RDAP → WHOIS

Agent-friendly output:

• Provenance on every result — source labels (source: crt.sh | certspotter | tls-san, source: rdap | whois) and per-source status so agents can assess completeness and trust
• Graceful partial failure — multi-target and multi-source tools return per-item/per-source error fields and operational notes instead of failing the whole call; only malformed input throws
• Discriminated, typed contracts — typed error reasons (source_unavailable, blocked_target, all_sources_failed) and union output (kind: domain | ip) let callers branch on data, not string parsing
• No fabricated signal — technology detections carry their triggering evidence; absent CT/DNS/RDAP fields are reported as unknown, never inferred

Getting started

Add the following to your MCP client configuration file. Every tool except attacksurface_lookup_host works with no configuration — the keyless core boots on an empty environment.

{
  "mcpServers": {
    "attack-surface-mcp-server": {
      "type": "stdio",
      "command": "bunx",
      "args": ["@cyanheads/attack-surface-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info"
      }
    }
  }
}

Or with npx (no Bun required):

{
  "mcpServers": {
    "attack-surface-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@cyanheads/attack-surface-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info"
      }
    }
  }
}

Or with Docker:

{
  "mcpServers": {
    "attack-surface-mcp-server": {
      "type": "stdio",
      "command": "docker",
      "args": ["run", "-i", "--rm", "-e", "MCP_TRANSPORT_TYPE=stdio", "ghcr.io/cyanheads/attack-surface-mcp-server:latest"]
    }
  }
}

To enable Shodan host intelligence (attacksurface_lookup_host), add SHODAN_API_KEY to the env block.

For Streamable HTTP, set the transport and start the server:

MCP_TRANSPORT_TYPE=http MCP_HTTP_PORT=3010 bun run start:http
# Server listens at http://localhost:3010/mcp

Prerequisites

• Bun v1.4.0 or higher (or Node.js v24+).
• No API key required for the core tools. Optional: a Shodan API key for attacksurface_lookup_host, and a Certspotter API key to raise CT-fallback rate limits.

Installation

1. Clone the repository:

git clone https://github.com/cyanheads/attack-surface-mcp-server.git

1. Navigate into the directory:

cd attack-surface-mcp-server

1. Install dependencies:

bun install

1. Configure environment (optional):

cp .env.example .env
# edit .env only if you want Shodan, a Certspotter key, or non-default behavior

Configuration

All variables are optional — the server boots and delivers its keyless core with an empty environment.

See .env.example for the full list of optional overrides.

Running the server

Local development

• Build and run:

  # One-time build
  bun run rebuild
  
  # Run the built server
  bun run start:stdio
  # or
  bun run start:http

• Run checks and tests:

  bun run devcheck   # Lint, format, typecheck, security
  bun run test       # Vitest test suite
  bun run lint:mcp   # Validate MCP definitions against spec

Docker

docker build -t attack-surface-mcp-server .
docker run --rm -e MCP_TRANSPORT_TYPE=http -p 3010:3010 attack-surface-mcp-server

The Dockerfile defaults to HTTP transport, stateless session mode, and logs to /var/log/attack-surface-mcp-server. OpenTelemetry peer dependencies are installed by default — build with --build-arg OTEL_ENABLED=false to omit them.

Project structure

Development guide

See CLAUDE.md/AGENTS.md for development guidelines and architectural rules. The short version:

• Handlers throw, framework catches — no try/catch in tool logic
• Use ctx.log for request-scoped logging, ctx.state for tenant-scoped storage
• Register new tools and resources via the barrels in src/mcp-server/*/definitions/index.ts
• Every outbound connection to a user-supplied target must pass the SSRF guard (assertSafeDomain / assertSafeUrl / assertSafeResolverIp) before connecting
• Wrap external API calls: validate raw → normalize to domain type → return output schema; never fabricate missing fields

Contributing

Issues are welcome. Run checks and tests before submitting:

bun run devcheck
bun run test

License

Apache-2.0 — see LICENSE for details.