Dev Machine Guard CLI
A scanner finds AI agents, MCP servers, IDE extensions and risky packages and settings on a developer machine in seconds.
- 197 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The scanner reads the home directory, package lists and MCP and IDE configuration, and the telemetry mode uploads results to an external dashboard.
Install
Manual install
VERSION=$(curl -fsSL https://api.github.com/repos/step-security/dev-machine-guard/releases/latest | grep '"tag_name"' | cut -d'"' -f4 | sed 's/^v//')
curl -fsSL "https://github.com/step-security/dev-machine-guard/releases/download/v${VERSION}/stepsecurity-dev-machine-guard-${VERSION}-darwin" -o stepsecurity-dev-machine-guard
chmod +x stepsecurity-dev-machine-guard
./stepsecurity-dev-machine-guardThe macOS option from the README; Windows and Linux snippets sit there too. Build from source with `make build` and Go 1.26.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The scanner reads the home directory, package lists and MCP and IDE configuration, and the telemetry mode uploads results to an external dashboard.
This is third-party code. Review the repository files before installing.
What it does
The StepSecurity program inventories what EDR and MDM cannot see: IDE extensions, AI agents and tools, MCP configurations, Node.js, Homebrew and Python packages, package-manager settings and suspicious files. It is a single dependency-free Go binary for macOS, Windows and Linux. Output can be terminal, `--json` or a self-contained `--html FILE` report. In CI you can count MCP configurations from the JSON and fail when any are found. Organizations get a mode that sends telemetry to the StepSecurity dashboard and scheduled runs through launchd, systemd or schtasks.
Who it is for. Developers and security teams who want to see which agents, MCP servers and extensions are installed on work machines.
Good fit when
- Before an audit you need to know which MCP servers and AI agents run on a laptop.
- A suspicious VS Code extension is checked after news of credential theft.
- CI fails when an MCP config is found in an image.
- A fleet of machines is scanned through Jamf, Kandji or Intune.
Not a fit when
- You need virus protection: the tool does not detect malware.
- You need device posture and policy compliance, like an MDM.
- It will not change what it finds; it only reports.
Example request
Scan this laptop and return JSON: which AI agents, MCP servers and IDE extensions were found, with no data sent anywhere.Limitations
The tool complements EDR and MDM and does not replace them: it does not detect viruses or assess device posture. It scans `$HOME`, and other directories are set through `--search-dirs`. Package scanning is enabled with `--enable-npm-scan`, `--enable-brew-scan` and `--enable-python-scan`. The `send-telemetry` mode uploads data to the StepSecurity service and is meant only for organization customers. Enterprise customers are advised to use the loader rather than a manual install.
How to disable. Delete the binary. If scheduled scanning was installed, run `stepsecurity-dev-machine-guard uninstall` and clear stored settings and the API key from the configuration.
FAQ
How do I count MCP servers in CI?
Run `--json` and read `.summary.mcp_configs_count` through `jq`. The README shows a script that exits with code 1 if any config is found.
Which operating systems are supported?
macOS (universal binary), Windows and Linux; Linux also has `.deb` and `.rpm` packages on the releases page.
Can I narrow the scan area?
Yes, `--search-dirs /Volumes/code` searches only the given directories instead of `$HOME`.
Is a StepSecurity account needed?
Not for a local run. Credentials and an API key are required only for telemetry and organization scheduled reports.
Does it replace EDR or antivirus?
No. It complements EDR and MDM and shows the developer tooling layer. Malware and device posture stay on their side.
Discussion
No questions or reviews yet.
Related
CLI toolsSecurity scanner for coding agents: checks code, MCP servers, prompts and model-invented packages, and gives an A to F grade.
Security scanner for MCP servers that looks for hardcoded keys, tool poisoning and injection flaws in repository code.
Security scanner for agent skills and MCP servers: finds prompt injection, data exfiltration and supply chain risks before install
An AI security verification CLI: scans models and code, probes live endpoints and MCP servers, and analyzes recorded agent runs
Security Detections MCP: a detection rule database
Security Detections MCP
An MCP server that merges Sigma, Splunk, Elastic and KQL rules into one database for finding detections and ATT&CK gaps
A CLI that scans your machine for skills and MCP configs, finds poisoned tool descriptions and runs a system prompt through 225 attacks
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Dev Machine Guard (this card) | no | 197 | |
| MCP Scanner | yes | 1,117 | scans specific MCP servers for threats; Dev Machine Guard finds all servers, agents and extensions on the machine |
| Skill Scanner | no | 2,724 | checks skills for threats; Dev Machine Guard inventories agents, MCP servers and packages on a work machine |
README in short
The project explains that developer machines hold tokens and keys and run untrusted code through dependencies and AI tools. Scan coverage is described in `SCAN_COVERAGE.md`.