Skip to content

reading-scan-results.md

Latest commit

 

History

History
272 lines (219 loc) · 8.26 KB

reading-scan-results.md

File metadata and controls

272 lines (219 loc) · 8.26 KB

StepSecurity Dev Machine Guard — Reading Scan Results

This guide explains how to interpret the output from Dev Machine Guard in all three output formats: pretty terminal output, JSON, and HTML.

Back to README | See also: Community Mode | MCP Audit

Pretty Terminal Output

The default output format is a styled, color-coded report printed to the terminal. It is organized into the following sections:

Banner

  ┌──────────────────────────────────────────────────────────┐
  │  StepSecurity Dev Machine Guard v1.8.1                   │
  │  https://github.com/step-security/dev-machine-guard      │
  └──────────────────────────────────────────────────────────┘
  Scanned at 2026-03-27 14:30:00

Shows the StepSecurity agent version and the timestamp of the scan.

DEVICE

  DEVICE
    Hostname         MacBook-Pro.local
    Serial           XXXXXXXXXXXX
    macOS            15.3
    User             user@example.com

Basic device identification. The user is determined from the currently logged-in console user.

SUMMARY

  SUMMARY
    AI Agents and Tools      5
    IDEs & Desktop Apps      3
    IDE Extensions           47

High-level counts for each detection category.

AI AGENTS AND TOOLS

  AI AGENTS AND TOOLS                                5 found
    claude-code              v1.0.12        [cli]         Anthropic
    codex                    v0.1.0         [cli]         OpenAI
    openclaw                 v0.5.2         [agent]       OpenSource
    ollama                   v0.5.4         [framework]   Unknown
    claude-cowork            v0.7.1         [agent]       Anthropic

Lists all detected AI tools, grouped by type:

  • cli -- Command-line AI coding assistants (Claude Code, Codex, Gemini CLI, etc.)
  • agent -- General-purpose AI agents (OpenClaw, GPT-Engineer, Claude Cowork, etc.)
  • framework -- AI frameworks and runtimes (Ollama, LM Studio, LocalAI, etc.)

IDE & AI DESKTOP APPS

  IDE & AI DESKTOP APPS                              3 found
    Visual Studio Code       v1.96.0        Microsoft
    Cursor                   v0.50.1        Cursor
    Claude                   v0.7.1         Anthropic

Lists installed IDEs and AI desktop applications found in /Applications/.

IDE EXTENSIONS

  IDE EXTENSIONS                                     47 found
    VSCode                                           35 found
      ms-python.python                v2024.22.0  ms-python
      esbenp.prettier-vscode          v10.4.0     esbenp
      ...
    Cursor                                           12 found
      ms-python.python                v2024.20.0  ms-python
      ...

Lists all installed extensions, grouped by IDE. Each entry shows the extension ID, version, and publisher.

JSON Output

When you run with --json, the scanner outputs a single JSON object to stdout. Here is the complete schema:

{
  "agent_version": "1.8.1",
  "scan_timestamp": 1709136000,
  "scan_timestamp_iso": "2026-02-28T14:00:00Z",
  "device": {
    "hostname": "MacBook-Pro.local",
    "serial_number": "XXXXXXXXXXXX",
    "os_version": "15.3",
    "platform": "darwin",
    "user_identity": "user@example.com"
  },
  "ai_agents_and_tools": [
    {
      "name": "claude-code",
      "vendor": "Anthropic",
      "type": "cli_tool",
      "version": "1.0.12",
      "binary_path": "/usr/local/bin/claude",
      "config_dir": "/Users/dev/.claude"
    },
    {
      "name": "openclaw",
      "vendor": "OpenSource",
      "type": "general_agent",
      "version": "0.5.2",
      "install_path": "/Users/dev/.openclaw"
    },
    {
      "name": "ollama",
      "vendor": "Unknown",
      "type": "framework",
      "version": "0.5.4",
      "binary_path": "/usr/local/bin/ollama",
      "is_running": true
    }
  ],
  "ide_installations": [
    {
      "ide_type": "vscode",
      "version": "1.96.0",
      "install_path": "/Applications/Visual Studio Code.app",
      "vendor": "Microsoft",
      "is_installed": true
    }
  ],
  "ide_extensions": [
    {
      "id": "ms-python.python",
      "name": "python",
      "version": "2024.22.0",
      "publisher": "ms-python",
      "install_date": 1709000000,
      "ide_type": "vscode"
    }
  ],
  "mcp_configs": [
    {
      "source": "claude_desktop",
      "vendor": "Anthropic",
      "config_path": "/Users/dev/Library/Application Support/Claude/claude_desktop_config.json",
      "servers": [
        {
          "name": "filesystem",
          "command": "npx",
          "args": ["-y", "@modelcontextprotocol/server-filesystem", "/Users/dev/projects"]
        }
      ]
    }
  ],
  "node_package_managers": [
    {
      "name": "npm",
      "version": "10.2.0",
      "path": "/usr/local/bin/npm"
    }
  ],
  "summary": {
    "ai_agents_and_tools_count": 5,
    "ide_installations_count": 3,
    "ide_extensions_count": 47,
    "mcp_configs_count": 1,
    "node_projects_count": 0
  }
}

Key Fields

Field Type Description
agent_version string Version of the scanner binary
scan_timestamp number Unix timestamp (seconds) of the scan
scan_timestamp_iso string ISO 8601 timestamp
device object Device identification information
ai_agents_and_tools array All detected AI tools, agents, and frameworks
ide_installations array Installed IDEs and AI desktop apps
ide_extensions array All installed IDE extensions
mcp_configs array MCP server configurations found across AI tools
node_package_managers array Detected Node.js package managers (npm, yarn, pnpm, bun)
summary object Count summaries

AI Tool Types

type value Description
cli_tool Command-line AI coding assistant
general_agent General-purpose AI agent
framework AI framework or runtime

IDE Types

ide_type value Description
vscode Visual Studio Code
cursor Cursor
windsurf Windsurf
antigravity Antigravity (Google)
zed Zed
claude_desktop Claude Desktop
microsoft_copilot_desktop Microsoft Copilot

HTML Output

The HTML report (--html report.html) generates a self-contained HTML file with the following layout:

  1. Header -- Purple gradient banner with "StepSecurity Dev Machine Guard Report" title
  2. Scan metadata -- Timestamp and agent version
  3. Summary cards -- Three cards showing counts for AI Agents and Tools, IDEs & Desktop Apps, and IDE Extensions
  4. Device grid -- Hostname, serial, macOS version, and user in a two-column grid
  5. AI Agents and Tools table -- Name, version, type (with a styled badge), and vendor
  6. IDE & AI Desktop Apps table -- Name, version, vendor, and install path
  7. IDE Extensions table -- Extension ID, version, publisher, and IDE

The HTML report is styled with StepSecurity branding (purple accent colors) and is fully responsive. It can be printed or shared as a standalone file.

Tips for Scripting with JSON Output

Extract all AI tool names

./stepsecurity-dev-machine-guard --json | jq -r '.ai_agents_and_tools[].name'

Count extensions per IDE

./stepsecurity-dev-machine-guard --json | jq '[.ide_extensions[] | .ide_type] | group_by(.) | map({(.[0]): length}) | add'

Check if a specific extension is installed

./stepsecurity-dev-machine-guard --json | jq '.ide_extensions[] | select(.id == "ms-python.python")'

Export extension list as CSV

./stepsecurity-dev-machine-guard --json | jq -r '.ide_extensions[] | [.id, .version, .publisher, .ide_type] | @csv'

Further Reading

StepSecurity -- securing the developer toolchain, from CI/CD to the dev machine.