Zscaler Mcp Server— plugin

zscaler-mcp-server is a Model Context Protocol (MCP) server that connects AI agents with the Zscaler Zero Trust Exchange platform. By default, the server operates in read-only mode for security, requiring explicit opt-in to enable write operations.

Support Disclaimer

-> Disclaimer: Please refer to our General Support Statement before proceeding with the use of this provider. You can also refer to our troubleshooting guide for guidance on typical problems.

[!IMPORTANT] 🚧 Public Preview: This project is currently in public preview and under active development. Features and functionality may change before the stable 1.0 release. While we encourage exploration and testing, please avoid production deployments. We welcome your feedback through GitHub Issues to help shape the final release.

📄 Table of contents

• 📺 Overview
• 🔒 Security & Permissions
• 🔐 MCP Client Authentication
• Supported Tools
• Installation & Setup
• Prerequisites
• Environment Configuration
• Installation
• Usage
• Command Line
• Service Configuration
• Additional Command Line Options
• Zscaler API Credentials & Authentication
• OneAPI Authentication
• Authentication Troubleshooting
• MCP Server Configuration
• As a Library
• Container Usage
• Using Pre-built Image (Recommended)
• Building Locally (Development)
• Editor/Assistant Integration
• Using uvx (recommended)
• With Service Selection
• Using Individual Environment Variables
• Docker Version
• Additional Deployment Options
• Remote MCP Deployment (EC2, VM, etc.)
• Azure Container Apps / Virtual Machine / AKS (Preview)
• Google Cloud (Cloud Run / GKE / VM / ADK Agent)
• Kubernetes (Helm Chart)
• Amazon Bedrock AgentCore
• Using the MCP Server with Agents
• Claude Desktop
• Cursor
• Visual Studio Code + GitHub Copilot
• Platform Integrations
• Troubleshooting
• License

📺 Overview

The Zscaler Integrations MCP Server brings context to your agents. Try prompts like:

• "List my ZPA Application segments"
• "List my ZPA Segment Groups"
• "List my ZIA Rule Labels"

[!WARNING] 🔒 READ-ONLY BY DEFAULT: For security, this MCP server operates in read-only mode by default. Only list_* and get_* operations are available. To enable tools that can CREATE, UPDATE, or DELETE Zscaler resources, you must explicitly enable write mode using the --enable-write-tools flag or by setting ZSCALER_MCP_WRITE_ENABLED=true. See the Security & Permissions section for details.

<!-- markdownlint-disable MD028 -->

[!TIP] Writing effective prompts: This server exposes 402 tools across multiple Zscaler services. Most MCP clients (Claude Desktop, Cursor, etc.) use deferred tool loading and will search for relevant tools based on your prompt. For best results, be specific about the service and action in your prompts:

>

- Good: "List my ZPA application segments" — targets the right service and tool directly - Good: "Show ZIA firewall rules" — clear service (zia) and action (list) - Less effective: "Show me my devices" — ambiguous; multiple services expose device-related tools

>

When a service is disabled, its tools are fully removed from the server. However, the AI agent may still attempt to find related tools in other services. If you get unexpected results, refine your prompt with the specific service name (e.g. zpa, zia, zdx, zcc, zcell, zms).

🔒 Security & Permissions

The Zscaler MCP Server implements a security-first design with granular permission controls and safe defaults:

Read-Only Mode (Default - Always Available)

By default, the server operates in read-only mode, exposing only tools that list or retrieve information:

• ✅ ALWAYS AVAILABLE - Read-only tools are registered by the server
• ✅ Safe to use with AI agents autonomously
• ✅ No risk of accidental resource modification or deletion
• ✅ All list_* and get_* operations are available (110+ read-only tools)
• ❌ All create_*, update_*, and delete_* operations are disabled by default
• 💡 Note: You may need to enable read-only tools in your AI agent's UI settings

# Read-only mode (default - safe)
zscaler-mcp

When the server starts in read-only mode, you'll see:

🔒 Server running in READ-ONLY mode (safe default)
   Only list and get operations are available
   To enable write operations, use --enable-write-tools AND --write-tools flags

💡 Read-only tools are ALWAYS registered by the server regardless of any flags. You never need to enable them server-side. Note: Your AI agent UI (like Claude Desktop) may require you to enable individual tools before use.

Write Mode (Explicit Opt-In - Allowlist REQUIRED)

To enable tools that can create, modify, or delete Zscaler resources, you must provide BOTH flags:

1. ✅ --enable-write-tools - Global unlock for write operations
2. ✅ --write-tools "pattern" - MANDATORY explicit allowlist

🔐 SECURITY: Allowlist is MANDATORY - If you set --enable-write-tools without --write-tools, 0 write tools will be registered. This ensures you consciously choose which write operations to enable.

# ❌ WRONG: This will NOT enable any write tools (allowlist missing)
zscaler-mcp --enable-write-tools

# ✅ CORRECT: Explicit allowlist required
zscaler-mcp --enable-write-tools --write-tools "zpa_create_*,zpa_delete_*"

When you try to enable write mode without an allowlist:

⚠️  WRITE TOOLS MODE ENABLED
⚠️  NO allowlist provided - 0 write tools will be registered
⚠️  Read-only tools will still be available
⚠️  To enable write operations, add: --write-tools 'pattern'

#### Write Tools Allowlist (MANDATORY)

The allowlist provides two-tier security:

1. ✅ First Gate: --enable-write-tools must be set (global unlock)
2. ✅ Second Gate: Explicit allowlist determines which write tools are registered (MANDATORY)

Allowlist Examples:

# Enable ONLY specific write tools with wildcards
zscaler-mcp --enable-write-tools --write-tools "zpa_create_*,zpa_delete_*"

# Enable specific tools without wildcards
zscaler-mcp --enable-write-tools --write-tools "zpa_create_application_segment,zia_create_rule_label"

# Enable all ZPA write operations (but no ZIA/ZDX/ZTW)
zscaler-mcp --enable-write-tools --write-tools "zpa_*"

Or via environment variable:

export ZSCALER_MCP_WRITE_ENABLED=true
export ZSCALER_MCP_WRITE_TOOLS="zpa_create_*,zpa_delete_*"
zscaler-mcp

Wildcard patterns supported:

• zpa_create_* - Allow all ZPA creation tools
• zpa_delete_* - Allow all ZPA deletion tools
• zpa_* - Allow all ZPA write tools
• *_application_segment - Allow all operations on application segments
• zpa_create_application_segment - Exact match (no wildcard)

When using a valid allowlist, you'll see:

⚠️  WRITE TOOLS MODE ENABLED
⚠️  Explicit allowlist provided - only listed write tools will be registered
⚠️  Allowed patterns: zpa_create_*, zpa_delete_*
⚠️  Server can CREATE, MODIFY, and DELETE Zscaler resources
🔒 Security: 85 write tools blocked by allowlist, 8 allowed

Tool Design Philosophy

Each operation is a separate, single-purpose tool with explicit naming that makes its intent clear:

#### ✅ Good (Verb-Based - Current Design)

zpa_list_application_segments    ← Read-only, safe to allow-list
zpa_get_application_segment      ← Read-only, safe to allow-list
zpa_create_application_segment   ← Write operation, requires --enable-write-tools
zpa_update_application_segment   ← Write operation, requires --enable-write-tools
zpa_delete_application_segment   ← Destructive, requires --enable-write-tools

This design allows AI assistants (Claude, Cursor, GitHub Copilot) to:

• Allow-list read-only tools for autonomous exploration
• Require explicit user confirmation for write operations
• Clearly understand the intent of each tool from its name

#### Responses: the API record, verbatim

A read tool returns the Zscaler API record unchanged. The server does not trim, rename, or re-declare a resource's attributes — that attribute set belongs to the API, so a field Zscaler adds in future reaches you with no server upgrade.

To keep responses small, you control what comes back rather than the server guessing. Two levers:

• `--toolsets` — load only the slice of tools you need, so the tool catalog

stays small (see Toolsets).

• `query` — every list tool accepts an optional JMESPath

expression applied to the results, so the agent projects exactly what it wants:

  zcc_list_devices(query="[*].{user: user, policy: policyName}")   # just those two fields
  zcc_list_devices(query="[?registrationState=='Quarantined']")    # just quarantined devices
  zcc_list_devices(query="length(@)")                              # just the count

Field names are exactly what the Zscaler API returns. Omit query to get the full records.

Security Layers

The server implements multiple layers of security (defense-in-depth). The first nine apply on every transport, including stdio — they govern which tools are exposed and how dangerous calls are confirmed. The remaining HTTP-only layers (TLS, host-header validation, source-IP ACL, MCP client authentication) are described in the Network-Level Controls section further down.

1. Read-Only Tools Always Enabled: Safe list_* and get_* operations are always available (254 tools).
2. Default Write Mode Disabled: Write tools are disabled unless explicitly enabled via --enable-write-tools.
3. Mandatory Allowlist: Write operations require explicit --write-tools allowlist (wildcard support).
4. OneAPI Entitlement Filter: At startup, toolsets for products the OneAPI credentials cannot call are silently dropped (see OneAPI Entitlement Filter below).
5. Toolset Selection: Optionally narrow the registered tool surface to a specific slice (e.g. --toolsets zia_url_filtering,zpa_app_segments). See the Toolsets section below.
6. Verb-Based Tool Naming: Each tool clearly indicates its purpose (list, get, create, update, delete).
7. Tool Metadata Annotations: All tools are annotated with readOnlyHint or destructiveHint for AI agent frameworks.
8. AI Agent Confirmation: All write tools marked with destructiveHint=True trigger permission dialogs in AI assistants.
9. Human Confirmation for DELETE: Delete operations are gated by a server-initiated confirmation. On clients that support MCP elicitation (Claude Desktop, Cursor), the server asks the client to prompt a human and the answer arrives as a protocol field — so a hijacked agent cannot author the approval, and a failed round trip fails closed. Clients without that capability fall back to a cryptographic confirmation token (HMAC-SHA256, single-use, 5-minute TTL, bound to the exact tool and parameters), which stops an approval being spent on another resource, replayed or reused — but is not by itself an anti-prompt-injection control, since a hijacked agent receives and can redeem it. The fallback is also single-process; elicitation-capable clients are unaffected. See MCP protocol posture — confirmation threat model for the full analysis.
10. Environment Variable Control: ZSCALER_MCP_WRITE_ENABLED, ZSCALER_MCP_WRITE_TOOLS, ZSCALER_MCP_TOOLSETS, ZSCALER_MCP_DISABLE_ENTITLEMENT_FILTER, and the disable lists can all be managed centrally without code changes.
11. Output Sanitization: Every string in every tool result is run through a three-stage sanitizer before reaching the agent — invisible/control characters (BiDi overrides, zero-width chars, BOM, soft hyphen) are stripped, raw HTML and HTML comments are removed (via bleach), Markdown link/image syntax is neutralised so embedded URLs cannot be smuggled to the agent, and Markdown code-fence info-strings containing role-impersonation tokens (system, assistant, tool, ignore, …) are collapsed to a neutral text tag. This defends against prompt-injection payloads that an attacker — or a careless admin — might embed in editable Zscaler resources (rule descriptions, location names, label descriptions, etc.). On by default. Opt-out with ZSCALER_MCP_DISABLE_OUTPUT_SANITIZATION=true (use only for diagnostics).
12. Audit Logging: When --log-tool-calls / ZSCALER_MCP_LOG_TOOL_CALLS=true is set, every tool invocation is logged with its arguments (sensitive values redacted), duration, and a result summary.

This multi-layered approach ensures that even if one security control is bypassed, others remain in place to prevent unauthorized operations. Layers 1-12 above apply equally to stdio, sse, and streamable-http.

Toolsets

Tools are grouped into 63 named toolsets so you can load only the slice an agent actually needs (e.g. zia_url_filtering (5 tools) instead of every tool from every service (402)). Toolsets reduce the agent's context cost and improve tool-selection accuracy.

# Load just two slices
zscaler-mcp --toolsets zia_url_filtering,zpa_app_segments

# Or use the curated default-on subset
zscaler-mcp --toolsets default

# Or load every registered toolset explicitly
zscaler-mcp --toolsets all

# Equivalent via environment variable
export ZSCALER_MCP_TOOLSETS="zia_url_filtering,zpa_app_segments"

When --toolsets is unspecified, every toolset whose service is enabled is loaded (preserves the historical default).

The agent can also enable additional toolsets at runtime through the always-on zscaler_list_toolsets, zscaler_get_toolset_tools, and zscaler_enable_toolset tools.

For the full catalog (29 toolsets across all services), filter precedence rules, per-toolset agent guidance, and the complete reference, see docs/guides/toolsets.md.

OneAPI Entitlement Filter

After your toolset selection resolves, the server reads the product entitlements from the OneAPI bearer token issued for your ZSCALER_CLIENT_ID and silently drops toolsets for products the credentials cannot call. If your OneAPI client is only entitled to ZIA and ZPA, every zdx_* / zcc_* / ztw_* / zid_* / zeasm_* / zins_* / zms_* toolset is filtered out at startup — even with --toolsets all.

This prevents an agent from discovering tools whose first call would only ever return 401 Unauthorized. The filter applies on every transport, including stdio.

When the filter runs you'll see one log line at startup, for example:

entitlement filter applied: entitled services=['zia', 'zpa'], kept 12 toolset(s), removed 17 toolset(s)

The filter is non-fatal. If credentials are missing, the token endpoint is unreachable, the token doesn't decode, or the token has no recognizable product entitlements, the server logs a single WARN line and starts normally with the user-selected toolsets unchanged.

To bypass the filter (for example, while diagnosing an unusual token shape):

zscaler-mcp --no-entitlement-filter
# or
export ZSCALER_MCP_DISABLE_ENTITLEMENT_FILTER=true

Only product entitlement is honoured — not role names. The server defers per-action permission enforcement to the live API; the entitlement filter only ensures we don't advertise tools for products the client has zero access to.

Cryptographic Confirmation for Destructive Actions

Delete operations are never executed on the first call. What happens instead depends on the client:

• Clients that support MCP elicitation get an interactive prompt naming the resource, answered by a human. The AI agent never handles the approval.
• Every other client gets a cryptographic confirmation token (HMAC-SHA256) it must pass back to proceed. The token is bound to the specific operation and its parameters, is single-use, and expires after 5 minutes — so it can't be forged, replayed, or reused for a different resource.

There is no way to turn this off. Deletes are irreversible against a live tenant, so the server ships no flag or environment variable that skips the gate. If you don't want an agent deleting anything, don't allowlist the delete tools: write tools are off by default and --write-tools takes explicit patterns (see Write Mode).

Network-Level Controls (HTTP only)

The next four subsections — TLS, source-IP allowlist, host-header validation, and the .env plaintext-secret scanner — apply only to the HTTP transports (sse, streamable-http). They control who can reach the server over the network. They are independent of the tool-level controls listed in Security Layers above (read-only mode, write allowlist, toolsets, entitlement filter, HMAC confirmations), which apply on every transport including stdio.

The corresponding MCP client authentication (Bearer / Basic / OAuth 2.1) is a fifth network-level layer covered in detail in the MCP Client Authentication section further below.

#### HTTPS/TLS Support

HTTPS is required by default for non-localhost deployments. The server will refuse to start on a non-localhost interface without TLS certificates unless you explicitly set ZSCALER_MCP_ALLOW_HTTP=true.

When running with HTTP transports (sse or streamable-http), provide TLS certificates:

ZSCALER_MCP_TLS_CERTFILE=/path/to/cert.pem
ZSCALER_MCP_TLS_KEYFILE=/path/to/key.pem

# Optional: private key password and CA bundle
ZSCALER_MCP_TLS_KEYFILE_PASSWORD=your-key-password
ZSCALER_MCP_TLS_CA_CERTS=/path/to/ca-bundle.pem

When TLS is configured, the server automatically starts with HTTPS. This works with both public (CA-signed) and private (self-signed) certificates. Generate a self-signed certificate for testing:

openssl req -x509 -newkey rsa:2048 -keyout key.pem -out cert.pem -days 365 -nodes \
  -subj "/CN=localhost"

#### Source IP Access Control

You can restrict which client IPs are allowed to connect using ZSCALER_MCP_ALLOWED_SOURCE_IPS. When unset (the default), source IP filtering is disabled and deferred to upstream controls (firewall rules, AWS Security Groups, etc.).

# Allow only specific IPs/subnets
ZSCALER_MCP_ALLOWED_SOURCE_IPS=10.0.0.0/8,172.16.0.5

# Allow all (effectively disable — same as not setting the variable)
ZSCALER_MCP_ALLOWED_SOURCE_IPS=0.0.0.0/0

Supports individual IPv4/IPv6 addresses, CIDR notation, and the wildcard 0.0.0.0/0. Health-check endpoints (/health, /healthz, /ready) are exempt so load-balancer probes continue to work. Requests from disallowed IPs receive 403 Forbidden.

#### .env File Security Warning

When starting with HTTP transports, the server automatically scans any .env file in the working directory for plaintext secrets (values containing SECRET, PASSWORD, KEY, or TOKEN). If detected, a security warning is logged recommending the use of a secrets manager or environment variables instead.

Security Posture Banner

On startup, the server logs a consolidated Security Posture Banner summarizing the active security configuration — transport mode, host validation status, authentication mode, TLS status, and any active warnings. This makes it easy to verify the security state at a glance.

Key Security Principles:

• No "enable all write tools" backdoor exists - allowlist is mandatory
• AI agents must request permission before executing any write operation (destructiveHint)
• Every destructive action requires explicit user approval through the AI agent's permission framework
• Destructive confirmations are cryptographically bound to prevent prompt injection bypass

Best Practices

• Read-Only by Default: No configuration needed for safe operations - read-only tools are always available
• Mandatory Allowlist: Always provide explicit --write-tools allowlist when enabling write mode
• Development/Testing: Use narrow allowlists (e.g., --write-tools "zpa_create_application_segment")
• Production/Agents: Keep server in read-only mode (default) for AI agents performing autonomous operations
• CI/CD: Never set ZSCALER_MCP_WRITE_ENABLED=true without a corresponding ZSCALER_MCP_WRITE_TOOLS allowlist
• Least Privilege: Use narrowest possible allowlist patterns for your use case
• Wildcard Usage: Use wildcards for service-level control (e.g., zpa_create_*) or operation-level control (e.g., *_create_*)
• Audit Review: Regularly review which write tools are allowlisted and remove unnecessary ones
• Specific Prompts: With 402 tools and deferred loading, AI agents match prompts to tools by relevance. Use service-specific prompts (e.g., "List ZPA segments" instead of "Show my segments") for accurate tool selection

🔐 MCP Client Authentication

📖 Full Documentation: [Authentication & Deployment Guide](docs/deployment/authentication-and-deployment.md)

When running the MCP server over HTTP (sse or streamable-http transports), you can enable authentication to control who is allowed to connect to the server. This is independent from the Zscaler API credentials, which control how the server authenticates to Zscaler APIs.

For HTTP transports, the server auto-detects and enables authentication when auth-related environment variables are present. For stdio transport, authentication is not applicable (the operating system's process isolation provides security).

Authentication Modes

The server supports four authentication modes, configured via environment variables:

Quick Start

Enable authentication by setting these environment variables in your .env file:

# Enable authentication
ZSCALER_MCP_AUTH_ENABLED=true
ZSCALER_MCP_AUTH_MODE=api-key

# For api-key mode: set a shared secret
ZSCALER_MCP_AUTH_API_KEY=sk-your-secret-key-here

Then start the server with an HTTP transport:

zscaler-mcp --transport streamable-http

Clients must include the key in the Authorization header:

Authorization: Bearer sk-your-secret-key-here

How It Works

Authentication is implemented as ASGI middleware that wraps the HTTP transport layer:

MCP Client Request
      │
      ▼
┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Auth         │────▶│  MCP          │────▶│  Zscaler     │
│  Middleware   │     │  Server       │     │  APIs        │
└──────────────┘     └──────────────┘     └──────────────┘
 Layer 1: WHO          MCP Protocol        Layer 2: HOW
 can connect?          Processing          server talks
                                           to Zscaler

• Layer 1 (MCP Client Auth): Controlled by ZSCALER_MCP_AUTH_* variables — validates the incoming request
• Layer 2 (Zscaler API Auth): Controlled by ZSCALER_CLIENT_ID, ZSCALER_CLIENT_SECRET, etc. — authenticates the server to Zscaler APIs

These two layers are completely independent. You can enable one, both, or neither.

Configuration by Mode

#### API Key

ZSCALER_MCP_AUTH_ENABLED=true
ZSCALER_MCP_AUTH_MODE=api-key
ZSCALER_MCP_AUTH_API_KEY=sk-your-secret-key-here

#### JWT (External IdP via JWKS)

ZSCALER_MCP_AUTH_ENABLED=true
ZSCALER_MCP_AUTH_MODE=jwt
ZSCALER_MCP_AUTH_JWKS_URI=https://your-idp.com/.well-known/jwks.json
ZSCALER_MCP_AUTH_ISSUER=https://your-idp.com
ZSCALER_MCP_AUTH_AUDIENCE=zscaler-mcp-server
ZSCALER_MCP_AUTH_ALGORITHMS=RS256,ES256   # Optional (default: RS256,ES256)

#### Zscaler OneAPI Credentials

ZSCALER_MCP_AUTH_ENABLED=true
ZSCALER_MCP_AUTH_MODE=zscaler
# Uses ZSCALER_VANITY_DOMAIN and ZSCALER_CLOUD from your existing config

Clients authenticate with Basic Auth (client_id:client_secret) or custom headers (X-Zscaler-Client-ID / X-Zscaler-Client-Secret).

Authentication Defaults

For HTTP transports (sse, streamable-http), the server auto-detects and enables authentication if auth-related environment variables are present (e.g., ZSCALER_MCP_AUTH_JWKS_URI, ZSCALER_MCP_AUTH_API_KEY, or ZSCALER_VANITY_DOMAIN). If no auth configuration is detected and ZSCALER_MCP_AUTH_ENABLED is not explicitly set, the server logs a security warning but continues without authentication.

To explicitly disable authentication, set:

ZSCALER_MCP_AUTH_ENABLED=false

Authentication does not apply to stdio transport (process isolation provides security).

OAuth 2.1 (oidc mode)

oidc mode makes the server an OAuth 2.0 protected resource (RFC 9728). The server publishes /.well-known/oauth-protected-resource naming your Identity Provider; the client reads that, runs the OAuth flow against the IdP directly, and presents the resulting token. The server's only job is verifying the token signature against the IdP's published public keys.

It is configured entirely through environment variables — no code, no extra packages:

ZSCALER_MCP_AUTH_ENABLED=true
ZSCALER_MCP_AUTH_MODE=oidc

OIDCPROXY_CONFIG_URL=https://your-tenant.auth0.com/.well-known/openid-configuration
OIDCPROXY_CLIENT_ID=<your app registration's client id>
OIDCPROXY_BASE_URL=http://localhost:8000
OIDCPROXY_AUDIENCE=zscaler-mcp-server
# Optional: scopes a token must carry, comma-separated
# OIDCPROXY_REQUIRED_SCOPES=zscaler.read

Notes:

• No client secret. Verifying a signature needs the IdP's public keys, not a credential of ours. OIDCPROXY_CLIENT_SECRET is ignored if set.
• OIDCPROXY_BASE_URL is this server's public URL, which clients use as the resource identifier — not the IdP's.
• OIDCPROXY_AUDIENCE defaults to OIDCPROXY_CLIENT_ID. Entra ID puts the client ID in aud; Auth0 uses the API identifier.
• The issuer and JWKS URI are read from the IdP's discovery document at startup, so they always match what the IdP actually signs with.
• The server serves no /authorize, /token or /register — clients need a client ID issued by the IdP and cannot self-register.
• Works with any OIDC-compliant provider (Auth0, Okta, Microsoft Entra ID, Keycloak, Google, AWS Cognito, PingOne).
• All other security layers (TLS, source-IP ACL, host validation) remain active.

IdP requirements: an application with the callback URL your client uses (http://localhost:3334/oauth/callback for mcp-remote, with the port pinned) registered, and an API/resource identifier matching OIDCPROXY_AUDIENCE.

📖 For detailed setup instructions — including [Microsoft Entra ID step-by-step guide](docs/deployment/entra-id-oidcproxy.md), IdP-specific JWKS configuration, Docker deployment examples, client configuration for Claude/Cursor/VS Code, and troubleshooting — see the [Authentication & Deployment Guide](docs/deployment/authentication-and-deployment.md).

Supported Tools

<!-- The table below is auto-generated by zscaler-mcp --generate-docs. Edit descriptions in zscaler_mcp/services.py and re-run the generator. -->

<!-- generated:start service-summary -->

The Zscaler Integrations MCP Server provides 402 tools for all major Zscaler services:

<!-- generated:end service-summary -->

📖 [View Complete Tools Reference →](docs/guides/supported-tools.md)

Note: All write operations require the --enable-write-tools flag and an explicit --write-tools allowlist. See the Security & Permissions section for details.

Installation & Setup

Prerequisites

• Python 3.11 or higher
• uv or pip
• Zscaler API credentials (see below)

Environment Configuration

Copy the example environment file and configure your credentials:

cp .env.example .env

Then edit .env with your Zscaler API credentials:

Required Configuration (OneAPI):

• ZSCALER_CLIENT_ID: Your Zscaler OAuth client ID
• ZSCALER_CLIENT_SECRET: Your Zscaler OAuth client secret
• ZSCALER_CUSTOMER_ID: Your Zscaler customer ID
• ZSCALER_VANITY_DOMAIN: Your Zscaler vanity domain

Optional Configuration:

• ZSCALER_CLOUD: (Optional) Zscaler cloud environment (e.g., beta) - Required when interacting with Beta Tenant ONLY.
• ZSCALER_PRIVATE_KEY: (Optional) PEM-encoded private key for JWT-based OneAPI auth, used in place of ZSCALER_CLIENT_SECRET.
• ZSCALER_MCP_SERVICES: Comma-separated list of services to enable (default: all services)
• ZSCALER_MCP_TRANSPORT: Transport method - stdio, sse, or streamable-http (default: stdio)
• ZSCALER_MCP_DEBUG: Enable debug logging - true or false (default: false)
• ZSCALER_MCP_HOST: Host for HTTP transports (default: 127.0.0.1)
• ZSCALER_MCP_PORT: Port for HTTP transports (default: 8000)

Alternatively, you can set these as environment variables instead of using a `.env` file.

Important: Ensure your API client has the necessary permissions for the services you plan to use. You can always update permissions later in the Zscaler console.

Installation

#### Install with VS Code (Quick Setup)

Note: This will open VS Code and prompt you to configure the MCP server. You'll need to replace the placeholder values (<YOUR_CLIENT_ID>, etc.) with your actual Zscaler credentials.

#### Install using uv (recommended)

uv tool install zscaler-mcp

#### Install from source using uv (development)

uv pip install -e .

Remote deployment: When running on EC2/VM, activate the project venv before starting: source .venv/bin/activate. See Remote MCP Deployment.

#### Install from source using pip

pip install -e .

#### Install using make (convenience)

make install-dev

[!TIP] If zscaler-mcp-server isn't found, update your shell PATH.

For installation via code editors/assistants, see the Using the MCP Server with Agents section below.

Usage

[!NOTE] Default Security Mode: All examples below run in read-only mode by default (only list_* and get_* operations). To enable write operations (create_*, update_*, delete_*), add the --enable-write-tools flag to any command, or set ZSCALER_MCP_WRITE_ENABLED=true in your environment.

Command Line

Run the server with default settings (stdio transport, read-only mode):

zscaler-mcp

Run the server with write operations enabled:

zscaler-mcp --enable-write-tools

Run with SSE transport:

zscaler-mcp --transport sse

Run with streamable-http transport:

zscaler-mcp --transport streamable-http

Run with streamable-http transport on custom port:

zscaler-mcp --transport streamable-http --host 0.0.0.0 --port 8080

Service Configuration

The Zscaler Integrations MCP Server supports multiple ways to specify which services to enable:

#### 1. Command Line Arguments (highest priority)

Specify services using comma-separated lists:

# Enable specific services
zscaler-mcp --services zia,zpa,zdx

# Enable only one service
zscaler-mcp --services zia

#### 2. Environment Variable (fallback)

Set the ZSCALER_MCP_SERVICES environment variable:

# Export environment variable
export ZSCALER_MCP_SERVICES=zia,zpa,zdx
zscaler-mcp

# Or set inline
ZSCALER_MCP_SERVICES=zia,zpa,zdx zscaler-mcp

#### 3. Default Behavior (all services)

If no services are specified via command line or environment variable, all available services are enabled by default.

Service Priority Order:

1. Command line --services argument (overrides all)
2. ZSCALER_MCP_SERVICES environment variable (fallback)
3. All services (default when none specified)

Excluding Services and Tools

When you want to keep most tools available but exclude a few, use --disabled-tools or --disabled-services instead of listing every tool you want to include.

Both flags support wildcards via fnmatch patterns.

# Exclude a single tool
zscaler-mcp --disabled-tools zia_list_devices

# Exclude all tools from a service prefix
zscaler-mcp --disabled-tools "zcc_*"

# Exclude multiple patterns
zscaler-mcp --disabled-tools "zcc_*,zdx_list_devices"

# Exclude entire services
zscaler-mcp --disabled-services zcc,zdx

# Combine: keep all services but exclude specific tools
zscaler-mcp --disabled-tools "zia_list_devices,zdx_*_analysis"

Environment variables:

export ZSCALER_MCP_DISABLED_TOOLS="zia_list_devices,zdx_*"
export ZSCALER_MCP_DISABLED_SERVICES="zcc"

Precedence: --disabled-tools takes precedence over --tools (include list). A tool that matches both the include list and the exclude list will be excluded.

Additional Command Line Options

# Enable write operations (create, update, delete)
zscaler-mcp --enable-write-tools

# Enable debug logging
zscaler-mcp --debug

# Combine multiple options
zscaler-mcp --services zia,zpa --enable-write-tools --debug

For all available options:

zscaler-mcp --help

Available command-line flags:

• --transport: Transport protocol (stdio, sse, streamable-http)
• --services: Comma-separated list of services to enable
• --disabled-services: Comma-separated list of services to exclude (e.g., zcc,zdx)
• --tools: Comma-separated list of specific tools to enable
• --disabled-tools: Comma-separated list of tools to exclude, supports wildcards (e.g., zcc_*,zdx_list_devices)
• --toolsets: Comma-separated toolset ids to enable (e.g. zia_url_filtering,zpa_app_segments). Special values: default (curated default-on subset), all (every toolset). When unspecified, every toolset whose service is enabled is loaded. See docs/guides/toolsets.md.
• --no-entitlement-filter: Skip the OneAPI entitlement filter that trims toolsets to the products the configured ZSCALER_CLIENT_ID is entitled to. Emergency override only — the filter is non-fatal by default.
• --enable-write-tools: Enable write operations (disabled by default for safety)
• --write-tools: Mandatory allowlist of write tool patterns (e.g., "zpa_create_*,zpa_delete_*")
• --log-tool-calls: Enable per-tool-call audit logging (tool name, redacted arguments, duration, result summary)
• --debug: Enable debug logging
• --host: Host for HTTP transports (default: 127.0.0.1)
• --port: Port for HTTP transports (default: 8000)
• --user-agent-comment: Additional text appended to User-Agent header
• --generate-auth-token: Generate a client auth token snippet and exit
• --list-tools: List all available tools and exit
• --version: Show server version and exit

Supported Agents

• Claude
• Cursor
• VS Code or VS Code Insiders

Zscaler API Credentials & Authentication

The Zscaler Integrations MCP Server uses OneAPI authentication exclusively. A single set of credentials authenticates the server to every Zscaler product (ZIA, ZPA, ZCC, ZDX, Zscaler Cellular, ZTW, ZIdentity, ZMS, Z-Insights, EASM).

Zscaler Cellular (ZCell) needs one extra credential — your Zscaler Cellular customer ID via ZCELL_CUSTOMER_ID — which is separate from ZSCALER_CUSTOMER_ID (used by ZPA). See the environment-variable table below.

OneAPI Authentication

#### Prerequisites

1. Create an API Client in the ZIdentity platform.
2. Obtain your clientId, clientSecret (or privateKey for JWT), customerId, and vanityDomain.
3. Learn more: Understanding OneAPI.

#### Quick Setup

Create a .env file in your project root (or wherever you start the MCP server):

# OneAPI credentials (required)
ZSCALER_CLIENT_ID=your_client_id
ZSCALER_CLIENT_SECRET=your_client_secret
ZSCALER_CUSTOMER_ID=your_customer_id
ZSCALER_VANITY_DOMAIN=your_vanity_domain

# Required only for Zscaler Cellular (ZCell) tools
ZCELL_CUSTOMER_ID=your_zscaler_cellular_customer_id

# Optional: only required when targeting the Beta tenant
ZSCALER_CLOUD=beta

⚠️ Security: Do not commit .env to source control. Add it to your .gitignore.

#### OneAPI Environment Variables

#### Verification

After populating .env, start the server:

zscaler-mcp

If credentials are valid, the server starts cleanly. The Zscaler SDK client is created lazily on the first tool call, so missing or rotating credentials surface as a clear error message at call time rather than blocking server startup.

Authentication Troubleshooting

MCP Server Configuration

The following environment variables control MCP server behavior (not authentication):

#### User-Agent Header

The MCP server automatically includes a custom User-Agent header in all API requests to Zscaler services. The format is:

User-Agent: zscaler-mcp-server/<version> python/<python_version> <os>/<architecture>

Example:

User-Agent: zscaler-mcp-server/0.3.1 python/3.11.8 darwin/arm64

With Custom Comment:

You can append additional information (such as the AI agent details) using the ZSCALER_MCP_USER_AGENT_COMMENT environment variable or the --user-agent-comment CLI flag:

# Via environment variable
export ZSCALER_MCP_USER_AGENT_COMMENT="Claude Desktop 1.2024.10.23"

# Via CLI flag
zscaler-mcp --user-agent-comment "Claude Desktop 1.2024.10.23"

This results in:

User-Agent: zscaler-mcp-server/0.3.1 python/3.11.8 darwin/arm64 Claude Desktop 1.2024.10.23

The User-Agent helps Zscaler identify API traffic from the MCP server and can be useful for support, analytics, and debugging purposes.

As a Library

You can use the Zscaler Integrations MCP Server as a Python library in your own applications:

from zscaler_mcp.server import ZscalerMCPServer

# Create server with read-only mode (default - safe)
server = ZscalerMCPServer(
    debug=True,  # Optional, enable debug logging
    enabled_services={"zia", "zpa", "zdx"},  # Optional, defaults to all services
    enabled_tools={"zia_list_rule_labels", "zpa_list_application_segments"},  # Optional, defaults to all tools
    disabled_services={"zcc"},  # Optional, exclude entire services
    disabled_tools={"zcc_*", "zdx_list_devices"},  # Optional, exclude tools by name or wildcard
    user_agent_comment="My Custom App",  # Optional, additional User-Agent info
    enable_write_tools=False  # Optional, defaults to False (read-only mode)
)

# Run with stdio transport (default)
server.run()

# Or run with SSE transport
server.run("sse")

# Or run with streamable-http transport
server.run("streamable-http")

# Or run with streamable-http transport on custom host/port
server.run("streamable-http", host="0.0.0.0", port=8080)

Example with write operations enabled:

from zscaler_mcp.server import ZscalerMCPServer

# Create server with write operations enabled
server = ZscalerMCPServer(
    debug=True,
    enabled_services={"zia", "zpa"},
    enable_write_tools=True  # Enable create/update/delete operations
)

# Run the server
server.run("stdio")

Available Services: zcc, zdx, zcell, zia, zid, zeasm, zins, zms, zpa, ztw

Example with Environment Variables:

from zscaler_mcp.server import ZscalerMCPServer
import os

# Load from environment variables
server = ZscalerMCPServer(
    debug=True,
    enabled_services={"zia", "zpa"}
)

# Run the server
server.run("stdio")

Running Examples

# Run with stdio transport
python examples/basic_usage.py

# Run with SSE transport
python examples/sse_usage.py

# Run with streamable-http transport
python examples/streamable_http_usage.py

Container Usage

The Zscaler Integrations MCP Server is available as a pre-built container image for easy deployment:

Using Pre-built Image (Recommended)

# Pull the latest pre-built image
docker pull zscaler/zscaler-mcp-server:latest

# Run with .env file (recommended)
docker run --rm --env-file /path/to/.env zscaler/zscaler-mcp-server:latest

# Run with .env file and SSE transport
docker run --rm -p 8000:8000 --env-file /path/to/.env \
  zscaler/zscaler-mcp-server:latest --transport sse --host 0.0.0.0

# Run with .env file and streamable-http transport
docker run --rm -p 8000:8000 --env-file /path/to/.env \
  zscaler/zscaler-mcp-server:latest --transport streamable-http --host 0.0.0.0

# Run with .env file and custom port
docker run --rm -p 8080:8080 --env-file /path/to/.env \
  zscaler/zscaler-mcp-server:latest --transport streamable-http --host 0.0.0.0 --port 8080

# Run with .env file and specific services
docker run --rm --env-file /path/to/.env \
  zscaler/zscaler-mcp-server:latest --services zia,zpa,zdx

# Use a specific version instead of latest
docker run --rm --env-file /path/to/.env \
  zscaler/zscaler-mcp-server:1.2.3

# Alternative: Individual environment variables
docker run --rm -e ZSCALER_CLIENT_ID=your_client_id -e ZSCALER_CLIENT_SECRET=your_secret \
  -e ZSCALER_CUSTOMER_ID=your_customer_id -e ZSCALER_VANITY_DOMAIN=your_vanity_domain \
  zscaler/zscaler-mcp-server:latest

Building Locally (Development)

For development or customization purposes, you can build the image locally:

# Build the Docker image
docker build -t zscaler-mcp-server .

# Run the locally built image
docker run --rm -e ZSCALER_CLIENT_ID=your_client_id -e ZSCALER_CLIENT_SECRET=your_secret \
  -e ZSCALER_CUSTOMER_ID=your_customer_id -e ZSCALER_VANITY_DOMAIN=your_vanity_domain zscaler-mcp-server

Note: When using HTTP transports in Docker, always set --host 0.0.0.0 to allow external connections to the container.

Editor/Assistant Integration

You can integrate the Zscaler Integrations MCP server with your editor or AI assistant. Here are configuration examples for popular MCP clients:

Using uvx (recommended)

{
  "mcpServers": {
    "zscaler-mcp-server": {
      "command": "uvx",
      "args": ["--env-file", "/absolute/path/to/.env", "zscaler-mcp"]
    }
  }
}

Note: The published PyPI package is zscaler-mcp (not zscaler-mcp-server). When integrating as a Claude Code plugin, use ${CLAUDE_PLUGIN_ROOT}/.env instead of an absolute path; for Gemini extensions, use ${extensionPath}${pathSeparator}.env.

Additional Deployment Options

Remote MCP Deployment (EC2, VM, etc.)

When deploying the MCP server on a remote host (EC2, VM, internal server) so clients connect over HTTP from another machine:

Server setup:

1. Install and configure credentials (see Installation and Environment Configuration).
2. If using an editable install (uv pip install -e .), you must activate the project venv before running—otherwise an older or different installation may run:

   cd /path/to/zscaler-mcp-server
   source .venv/bin/activate
   zscaler-mcp --transport streamable-http --host 0.0.0.0 --port 8000

1. Use --host 0.0.0.0 to bind on all interfaces. This automatically disables Host header validation (required when clients send the server's public IP in the Host header). For production, consider ZSCALER_MCP_ALLOWED_HOSTS in .env to restrict to known hostnames.
2. Ensure the firewall allows inbound traffic on the chosen port (e.g. 8000).

Client configuration (Claude Desktop):

Claude Desktop expects a command that spawns a process. For remote HTTP, use mcp-remote which supports custom authentication headers.

macOS / Linux:

{
  "mcpServers": {
    "zscaler-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "http://YOUR_SERVER_IP:8000/mcp",
        "--allow-http",
        "--header",
        "Authorization: Bearer sk-your-api-key"
      ]
    }
  }
}

Windows:

On Windows, paths with spaces (e.g., C:\Program Files\...) cause npx to fail when invoked directly. Wrap the call through cmd /c:

{
  "mcpServers": {
    "zscaler-mcp-server": {
      "command": "cmd",
      "args": [
        "/c",
        "npx",
        "-y",
        "mcp-remote",
        "http://YOUR_SERVER_IP:8000/mcp",
        "--allow-http",
        "--header",
        "Authorization: Bearer sk-your-api-key"
      ]
    }
  }
}

`--allow-http`: Required when connecting to a non-localhost HTTP endpoint. mcp-remote enforces HTTPS by default for non-localhost URLs. Omit this flag when connecting over HTTPS or to localhost.

Using Zscaler auth mode (Basic Auth):

Replace the Authorization header with Basic Auth credentials. The value is the Base64 encoding of client_id:client_secret:

# Generate the Base64 value
echo -n "your-client-id:your-client-secret" | base64

Then use "Authorization: Basic <base64_value>" in place of the Bearer header above.

Prerequisites on the client: Node.js (for npx) must be installed.

📖 Full remote deployment details (venv usage, 421 troubleshooting, security, TLS): Remote Deployment · 421 Misdirected Request · Troubleshooting

Azure Container Apps / Virtual Machine / AKS (Preview)

Deploy the Zscaler MCP Server to Azure with your choice of deployment target:

# Interactive guided deployment — no .env file required
cd integrations/azure
python azure_mcp_operations.py deploy

# MCP Server management
python azure_mcp_operations.py status     # check health
python azure_mcp_operations.py logs       # stream logs
python azure_mcp_operations.py ssh        # SSH into VM (VM only)
python azure_mcp_operations.py destroy    # tear down all resources

# Azure AI Foundry Agent (optional — wraps MCP server as an AI agent)
pip install azure-ai-projects azure-identity
python azure_mcp_operations.py agent_create    # create Foundry agent
python azure_mcp_operations.py agent_chat      # interactive CLI chat
python azure_mcp_operations.py agent_status    # show agent info
python azure_mcp_operations.py agent_destroy   # delete agent

The script will prompt you for:

• Deployment target: Container Apps, Virtual Machine, or Azure Kubernetes Service (Preview)
• Credential source: .env file path or manual entry
• Auth mode: OIDC (OAuth 2.1), JWT, API Key, Zscaler, or None (OIDC not yet supported on AKS)
• Azure options: resource group, region, Key Vault (new or existing); for AKS: cluster lifecycle (create new or use existing), node count/size, namespace

Container Apps and VM store all secrets in Azure Key Vault (mandatory) and auto-configure Claude Desktop / Cursor. AKS Preview injects credentials as Kubernetes environment variables on the Deployment — Workload Identity Federation + Key Vault CSI driver integration is planned.

Foundry Agent: Optionally create an Azure-hosted AI agent that wraps the MCP server. The agent is accessible via CLI chat (with spinner, token tracking, and timing), the Azure AI Foundry portal playground, REST APIs, or Microsoft 365 Copilot integrations.

📖 Full Azure deployment guide: integrations/azure/README.md

Google Cloud (Cloud Run / GKE / VM / ADK Agent)

Google Cloud deployments cover both the standalone MCP Server and the Gemini-powered ADK Agent. A complete video walkthrough is available here:

[Zscaler Integration MCP Server in GCP — Video Demo](https://zscaler.wistia.com/medias/13jxjizk3r)

Two interactive Python scripts manage all five Google Cloud deployment targets:

Standalone MCP Server — Automated Deployment:

cd integrations/google/gcp
python gcp_mcp_operations.py deploy      # guided deployment (prompts for target)
python gcp_mcp_operations.py status      # check health
python gcp_mcp_operations.py logs        # stream logs
python gcp_mcp_operations.py ssh         # SSH into VM (VM target only)
python gcp_mcp_operations.py destroy     # tear down

The script prompts for deployment target, credentials, and auth mode. It:

• Prompts for GCP project/region (or reads from .env)
• Optionally stores credentials in GCP Secret Manager
• Deploys the container to Cloud Run with zscaler authentication mode
• Generates Authorization: Basic headers from your Zscaler OneAPI credentials
• Auto-configures Claude Desktop (claude_desktop_config.json) and Cursor (~/.cursor/mcp.json)
• Writes a deployment state file (.gcp-deploy-state.json) for subsequent status / logs / destroy operations

ADK Agent — Interactive Deployment:

cd integrations/google/adk
python adk_agent_operations.py deploy      # guided deployment (local, Cloud Run, Agent Engine, Agentspace)
python adk_agent_operations.py status
python adk_agent_operations.py logs
python adk_agent_operations.py destroy

Manual Deployment:

# Deploy with credentials as env vars
gcloud run deploy zscaler-mcp-server \
  --image=zscaler/zscaler-mcp-server:latest \
  --set-env-vars="ZSCALER_CLIENT_ID=...,ZSCALER_CLIENT_SECRET=...,ZSCALER_VANITY_DOMAIN=...,ZSCALER_CUSTOMER_ID=...,ZSCALER_CLOUD=production,ZSCALER_MCP_ALLOW_HTTP=true,ZSCALER_MCP_DISABLE_HOST_VALIDATION=true,ZSCALER_MCP_AUTH_ENABLED=true,ZSCALER_MCP_AUTH_MODE=zscaler" \
  --args="--transport,streamable-http,--host,0.0.0.0,--port,8000" \
  --port=8000 --region=us-central1 --allow-unauthenticated

With GCP Secret Manager (recommended for production):

The Docker image includes a built-in GCP Secret Manager loader. Store credentials as individual secrets (e.g., zscaler-client-id, zscaler-client-secret) and enable with a single env var:

gcloud run deploy zscaler-mcp-server \
  --image=zscaler/zscaler-mcp-server:latest \
  --set-env-vars="ZSCALER_MCP_GCP_SECRET_MANAGER=true,GCP_PROJECT_ID=your-project,ZSCALER_MCP_ALLOW_HTTP=true,ZSCALER_MCP_DISABLE_HOST_VALIDATION=true,ZSCALER_MCP_AUTH_ENABLED=true,ZSCALER_MCP_AUTH_MODE=zscaler" \
  --args="--transport,streamable-http,--host,0.0.0.0,--port,8000" \
  --port=8000 --region=us-central1 --allow-unauthenticated

The loader also works on GKE and Compute Engine — anywhere GCP Application Default Credentials are available.

📖 Full Google Cloud deployment guide (all five targets, IAM roles, authentication modes, enterprise patterns): integrations/google/README.md

>

📖 Secret Manager deep-dive (GKE manifests, credential rotation, loader internals): GCP Secret Manager Integration

Kubernetes (Helm Chart)

Deploy the Zscaler MCP Server to any Kubernetes cluster via Helm — EKS, GKE, AKS, OpenShift, Rancher, k3s, Talos, or kind / minikube for local dev. The chart is cluster-vendor-agnostic and never calls aws, az, or gcloud. Use this when the cluster is already a fact and your operating model treats every workload as a Helm release; if you instead need to stand up brand-new cloud infra, use the Azure / GCP / AWS deployment sections above.

# Interactive guided install — same pattern as the Azure / GCP scripts
python integrations/helm-chart/helm_mcp_operations.py deploy

# Lifecycle subcommands
python integrations/helm-chart/helm_mcp_operations.py status      # release + pods + Service + port-forward
python integrations/helm-chart/helm_mcp_operations.py logs        # tail Deployment logs
python integrations/helm-chart/helm_mcp_operations.py configure   # re-write Cursor / Claude configs
python integrations/helm-chart/helm_mcp_operations.py test        # run `helm test` smoke probe
python integrations/helm-chart/helm_mcp_operations.py destroy     # uninstall + optional ns deletion

The deployment script materialises a Kubernetes Secret from your existing .env (no translation into values.yaml), runs helm upgrade --install, waits for the rollout with live per-pod feedback (detects ImagePullBackOff / CreateContainerConfigError / CrashLoopBackOff and exits with a tailored recovery hint), starts a background kubectl port-forward when no Ingress is configured, and auto-configures Claude Desktop + Cursor with the right Authorization: Basic header.

Five credential-setup paths are supported (interactive script, kubectl create secret --from-env-file, inline --set, pre-existing Secret for GitOps, and External Secrets Operator). All five converge on the same chart contract — the Deployment does envFrom: secretRef: so every ZSCALER_MCP_* / ZSCALER_* key in your secret flows into the container untouched.

For the full chart reference (values.yaml keys, Ingress / HTTPRoute / cert-manager / HPA / PDB toggles, ExternalSecret examples), see integrations/helm-chart/README.md.

Amazon Bedrock AgentCore

[!IMPORTANT] One image, no AWS fork. Bedrock AgentCore used to be served by a separate build. It no longer is — the standard image runs on AgentCore as-is, configured entirely through environment variables. AgentCore's containerConfiguration accepts only an ECR URI (Docker Hub is not a legal value) and requires `linux/arm64`, so the published multi-arch image still needs to be copied into ECR — or pulled from the AWS Marketplace listing, which is the same image already in ECR for you.

🚀 Quick Start with AWS Marketplace:

The easiest way to deploy the Zscaler Integrations MCP Server to Amazon Bedrock AgentCore is through the AWS Marketplace listing. What you