ARL-Next MCP
MCP server for the ARL-Next platform: returns ports, fingerprints, certificates, vulnerabilities and leaks for a domain, IP or URL from your asset database
- 473 ★ on the repository
- Last commit
- GPL-3.0
Broad access
Why this level
- The ARL token opens the company's whole database of assets, vulnerabilities and found leaks, and replies land in the model's context
- The platform itself scans external targets with Nmap and Nuclei, and unauthorized scanning of other people's systems is illegal
- The ARL_TLS_VERIFY=false variable disables certificate checks and allows connection interception
Install
- Works in
Manual install
git clone --depth 1 https://github.com/owl234/ARL-Next.git && cd ARL-Next && bash start-prod.shDeploys the ARL-Next platform itself in Docker (the git clone route). Root is required: the script installs Docker and tunes the kernel. Open https://<server IP>:5173 once it finishes.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The ARL token opens the company's whole database of assets, vulnerabilities and found leaks, and replies land in the model's context
- The platform itself scans external targets with Nmap and Nuclei, and unauthorized scanning of other people's systems is illegal
- The ARL_TLS_VERIFY=false variable disables certificate checks and allows connection interception
This is third-party code. Review the repository files before installing.
What it does
ARL-Next is an attack-surface reconnaissance and asset monitoring platform. It finds subdomains, scans ports with Nmap, fingerprints web applications, runs Nuclei and tracks new CVEs and credential leaks on GitHub. Tasks run through Celery and RabbitMQ, data lives in MongoDB, and the start-prod.sh script deploys the stack in Docker. For agents the platform ships an MCP server in the mcp-server folder. It is written on the Python 3.8 or newer standard library, so the mcp, pydantic and requests packages are not needed. It exposes one tool, get_asset_profile, which takes a domain, an IP or a full URL and an optional 24-character ARL task ID. The reply is a compact JSON with IP, geolocation and ASN, open ports and services, the site's title and fingerprints, certificate SANs, found vulnerabilities and exposed files such as .git/config. Empty fields and redundant headers are pruned so the reply does not eat the agent's context. The server calls the API of your deployed ARL over HTTPS with a token. Starting scans and managing policies through MCP is not available yet: the author says those capabilities are still being integrated.
Who it is for. For security engineers and administrators who already inventory external assets in ARL and want to query them from Claude Code, Cursor or Codex.
Good fit when
- ARL already scanned the company's infrastructure, and for oa.example.com you need ports, stack and found vulnerabilities fast
- Before an SRC check or an internal pentest you want a host profile by IP: ASN, Nmap services, certificate and Nuclei hits
- You want to know whether a site exposes .git/config or another file listed in the leaks section of the profile
- An agent drafts a report on a group of assets and pulls profile data from the database instead of from memory
Not a fit when
- ARL-Next is not deployed: the MCP server scans nothing itself and works only on top of the platform API
- You need to start scans and change policies from chat: only profile reading is available through MCP today
- The target is not yours and you have no written permission from its owner: the author allows the platform only for authorized assessments
- You need an MCP for manual web traffic testing: the catalog has a Burp Suite extension for that
Example request
Use arl-next to get the profile of admin.example.com: ports, web stack, certificate and found vulnerabilities, and list exposed files separatelyMCP
- Transport
- stdio
- Authentication
- API key
Environment variables
ARL_HOSTrequired- Address of your deployed ARL, for example https://arl.example.com:5173
ARL_TOKENrequired, secret- API token from the ARL_auth.md file on the platform server
ARL_CA_BUNDLE- Absolute path to a trusted PEM certificate for a remote ARL with a self-signed certificate
ARL_STRICT_HOSTNAME- true makes host name matching in the certificate mandatory
ARL_TLS_VERIFY- false turns TLS verification off entirely, for temporary tests only
Limitations
- A running ARL-Next installation is required: start-prod.sh brings up Docker, MongoDB, RabbitMQ and the other services, and on a bare machine it installs Docker and tunes kernel parameters.
- The only tool, get_asset_profile, reads profiles. Creating tasks and managing policies through MCP is not connected yet.
- The first-run certificate is self-signed with CN=localhost and no SAN. When connecting by IP from another machine, set ARL_CA_BUNDLE to the server certificate path. ARL_TLS_VERIFY=false turns verification off entirely, and the author allows it only for temporary tests.
- Licensed under GPL-3.0. The author requires using the platform only for authorized security assessments and agreeing targets with their owners in writing.
How to disable. Run claude mcp remove arl-next or delete the arl-next block from the client's mcp.json. The ARL token stays in ARL_auth.md on the platform server, so reissue it in ARL if needed. The platform itself can be stopped with Docker Compose in the deployment folder.
FAQ
How do I connect ARL-Next to Claude Code?
Clone the repository and from its root run claude mcp add -s user arl-next python3 "$(pwd)/mcp-server/server.py" with ARL_HOST and ARL_TOKEN passed through --env. The token sits in ARL_auth.md on the platform server. If the variables are unset, the script looks for ARL_auth.md in the current folder and the project root.
What if ARL uses a self-signed certificate?
When it runs on the same machine as ARL, the server trusts certs/arl.crt on its own. For a remote ARL reached by IP, copy the certificate to the client and set ARL_CA_BUNDLE to its absolute path. ARL_STRICT_HOSTNAME=true turns on strict host name matching.
What does get_asset_profile return?
For a domain, IP or URL it returns resolved IPs, geolocation, ASN, ports with services and products, the site's title and status code, fingerprints, certificate issuer and SANs, a list of vulnerabilities with plugin and severity and a list of exposed files found. If nothing matches, the reply is not_found.
Do I need to install the mcp or requests package?
No. The server uses only sys, json, urllib and ssl from the Python 3.8 or newer standard library, so the system python3 runs it without a virtual environment.
How does ARL-Next differ from the classic ARL?
The author rebuilt the architecture: OSINT and screenshot capture became separate services, queues split into light and heavy, asset groups with corporate links, CVE sync and GitHub leak monitoring were added, along with this MCP server. The stack moved to Python 3.13, Vue 3.5, MongoDB 7.0 and Nmap 7.95.
Discussion
No questions or reviews yet.
Related
MCP serversAn MCP server that gives the agent vulnerability triage: NVD data, EPSS, the CISA KEV catalog, MITRE ATT&CK mapping and one-call CVE risk scoring
MCP Scanner
mcp-scanner
A scanner from Cisco AI Defense: checks MCP servers and their tools for threats with three engines, YARA rules, a model and the Cisco API
A bridge between an MCP client and a Kali Linux terminal so the agent can run nmap, sqlmap, hydra and other tools for authorized pentesting and CTF
An offensive security MCP server: the agent runs 150+ pentest and recon tools and drives automated target assessments
MCP server and CLI for reverse engineering: an agent inspects binaries, JavaScript apps, APKs and websites on your machine and returns the evidence
Self-hosted AI pentesting framework: attack surface graph, Kali sandbox, human approval gates and an MCP server you can drive from Claude Code
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| ARL-Next (this entry) | no | API key | 473 | |
| Burp Suite MCP Server | yes | not required | 1,247 | a Burp Suite extension for manual web traffic testing, no asset database with ports and certificates |
| RedAmon | yes | API key | 3,198 | a self-contained pentest framework with Kali in a container and manual step approval |
README in short
The front page in Chinese and English opens by comparing the old ARL with ARL-Next: decoupled queues, asset groups with corporate links, native MCP, Docker images and a refreshed runtime. Next come the architecture diagram, a deployment guide with three routes (Alibaba Cloud mirror, local build and git clone), HTTPS and Basic Auth setup, updates and links to Swagger UI. At the end the author places a legal notice: the platform is for authorized assessments only.