Works with the AI Agents you already use
ast url detonate
Self-hosted URL detonation sandbox for safe browser-based malware analysis and redirect tracking.
$10
ast url detonate
Example session with this skill installed
Run a detonation for http://suspicious-malware-site.com/login and tell me where it redirects.
- Read your context and instructions
- Compiled the ast url detonate
I have initiated a dry run for the requested URL detonation. To proceed, I need your explicit approval to contact this infrastructure from your sandbox host.
Dry Run Details:
- URL:
http://suspicious-malware-site.com/login - Host:
sandbox-host(192.168.1.50) - Subnet:
10.0.5.0/24(Egress Isolated) - Action: Remote browser execution and traffic capture.
Detonation Proposal
| Parameter | Value |
|---|---|
| Target URL | http://suspicious-malware-site.com/login |
| Exec Token | tok_8f2d931a4e5c |
| Egress Policy | Allowlisted (DNS, HTTP/S) |
| Duration Limit | 60 seconds |
| Artifacts | Screenshot, HAR, DOM, pcap |
Next steps
- Say "yes" to execute this detonation.
- After execution, I will poll the status until the redirect chain is captured.
- I will then provide a summary of the final landing page and any captured payloads.
Connects securely to your tools. The creator never sees your data.
About this skill
Suspecting a URL is malicious is one thing, but proving it requires a safe way to watch it behave. AST URL Detonate provides a self-hosted sandbox environment for security engineers to safely trigger, observe, and capture URL behavior without risking their primary workstation or local network.
What it does
- Remote Detonation triggers suspicious URLs inside a dedicated headless Chromium instance on a user-controlled Linux host.
- Evidence Capture retrieves the full redirect chain, DOM snapshots, HAR files, pcap network traffic, and screenshots.
- Secure File Retrieval downloads payloads and malware samples from the sandbox to the workstation only with explicit user flags.
- Infrastructure Isolation uses iptables egress allowlisting and dedicated subnets to prevent malware from scanning the local network.
- Verified Bundles generates a cryptographically signed artifact bundle for every detonation run for chain-of-custody.
How it works
- Dry Run the
url-detonate runcommand to generate a unique execution token for a specific URL. - Execute the detonation by replaying the token, which triggers the remote browser on the sandbox host.
- Poll Status via the CLI or MCP tool to monitor progress and capture completion events.
- Fetch Artifacts to pull the HAR, screenshots, and logs back to the local environment for analysis.
Frameworks & tools
The client requires Python 3.12+ and connects via SSH to a Linux sandbox host. The server kit runs on Docker with iptables for network isolation. It integrates with Claude Code and Cursor via the included MCP server.
Why this beats prompting it yourself
Generic AI agents often refuse to visit suspicious links or, worse, visit them directly from their own IP, alerting attackers. This skill enforces a strict air-gapped workflow, ensuring all risky network activity happens on an isolated host while providing the agent with the specific CLI and MCP tools to parse security artifacts accurately.
Use cases
- Phishing Analysis to follow shortened or obscured redirect chains to the final destination.
- Malware Triage to capture and download files served by suspicious domains in a controlled environment.
- Threat Hunting to record the network traffic (pcap) generated by a suspicious landing page.
- Evidence Collection for security reports requiring visual proof and DOM content of a live threat.
Known limitations
Requires a dedicated Linux host with Docker and root access for initial server setup. Does not support macOS or Windows for the sandbox host. Does not provide automated reputation verdicts.
Requires dedicated Linux host with Docker. No support for macOS/Windows sandboxes. No automated reputation scoring. Single detonation at a time.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 month ago
Requires
Permissions
Allowed Hosts
File Scopes
Terminal/Shell runs `install.sh`, `url-detonate`, `sectools` and `ssh`; Read/Write Files cover the venv, your `.env`, pending exec tokens and fetched bundles; Environment Variables carry the SSH host/key settings and the gate; Network Access is PyPI at install and SSH to your own sandbox host at run time (no vendor API, no API key). Over MCP every `url_detonate_*` tool, reads included, is registered only with `SECTOOLS_URL_DETONATE=1`. `run` is a dry run that returns a single-use exec token; detonating, `purge` (irreversible) and `fetch --include-payloads` (copies live malware) each need your explicit approval. Installing the server kit needs root on the sandbox host and is never run by the agent unprompted.