Installation

From the Marketplace

1. Open VS Code, Cursor, Windsurf, or Antigravity
2. Go to Extensions (Cmd/Ctrl+Shift+X)
3. Search for "Cisco AI Security Scanner"
4. Click Install

From VSIX

1. Download the latest .vsix file from GitHub Releases
2. Open the Command Palette (Cmd/Ctrl+Shift+P)
3. Run "Extensions: Install from VSIX..."
4. Select the downloaded file and reload

Requirements

Required:

• Python 3.10+ — If not available, a portable Python 3.11 runtime is downloaded automatically on first use. You can also point to a specific Python binary via the mcp-scanner.python.customPath setting.

Optional (for enhanced analysis):

• LLM API Key — For AI-driven analysis (see provider setup below)
• Cisco AI Defense API Key — For cloud-based threat classification
• VirusTotal API Key — For binary file hash lookup and scanning

Post-Install Setup

1. Look for the shield icon in the Activity Bar (left sidebar) to open the Security Scanner panel.
2. The Setup Wizard opens automatically on first launch. Follow its steps, or configure manually:
   • Run "Configure LLM Provider" from the Command Palette to set up AI analysis
   • Run "Configure Cisco AI Defense" for cloud-based threat classification
   • Run "Configure VirusTotal API" for binary scanning
3. Run "Scan All (MCP + Skills)" to start your first scan.

LLM Provider Setup

The scanner supports nine LLM providers for AI-driven analysis. You only need one. YARA and behavioral analysis run locally without any provider.

OpenAI

1. Run "Configure LLM Provider" from the Command Palette
2. Select OpenAI as the provider
3. Paste your API key
4. Ensure your account has API credits available

Anthropic

1. Run "Configure LLM Provider" from the Command Palette
2. Select Anthropic as the provider
3. Paste your API key

Azure OpenAI

1. Run "Configure LLM Provider" and select Azure OpenAI
2. Paste your API key
3. Set mcp-scanner.llm.azureEndpoint to your resource endpoint
4. Set mcp-scanner.llm.azureDeployment to your deployment name

Azure AI Services (Claude/Llama)

Setup is the same as Azure OpenAI — configure the API key, endpoint, and deployment name. Uses the Anthropic API format with Azure authentication.

AWS Bedrock

1. Create IAM credentials in AWS Console → IAM → Users
2. Grant bedrock:InvokeModel permission to your IAM user
3. Request model access in the Bedrock console before use
4. Run "Configure LLM Provider", select AWS Bedrock, and paste your secret key
5. Set the AWS region in settings (must have Bedrock access enabled, e.g. us-east-1)

GCP Vertex AI

1. Create a service account in GCP Console → IAM → Service Accounts
2. Grant the Vertex AI User role
3. Download the JSON key file
4. Run "Configure LLM Provider", select GCP Vertex, and provide the path to the JSON file
5. Set your project ID and region in settings

Alternatively, leave the key empty to use Application Default Credentials (gcloud auth application-default login).

Ollama (Local / Offline)

1. Install Ollama from ollama.ai
2. Pull a model: ollama pull llama3.2
3. Ensure Ollama is running: ollama serve
4. Run "Configure LLM Provider", select Ollama
5. Adjust the endpoint in settings if not using the default

OpenRouter

1. Run "Configure LLM Provider", select OpenRouter
2. Paste your API key
3. Model names include a provider prefix (e.g. openai/gpt-4o, anthropic/claude-3.5-sonnet)

Google AI Studio

1. Run "Configure LLM Provider", select Google AI Studio
2. Paste your API key
3. Free tier is available with usage limits

Supported Platforms

IDEs

• VS Code
• Cursor
• Windsurf
• Antigravity

MCP Configuration Sources

The scanner auto-discovers MCP server configurations from these locations:

Toggle individual sources on or off in the mcp-scanner.globalConfigs.* settings.

Skill Sources

LLM Providers