diff --git a/.changeset/cli-tool.md b/.changeset/cli-tool.md
new file mode 100644
index 0000000..b32b086
--- /dev/null
+++ b/.changeset/cli-tool.md
@@ -0,0 +1,32 @@
+---
+"@scope3/agentic-client": minor
+---
+
+Add dynamic CLI tool for Scope3 Agentic API with automatic command generation
+
+The CLI now dynamically discovers available API operations from the MCP server, ensuring commands are always up-to-date with the latest API capabilities:
+
+- **Zero maintenance**: Commands auto-generate from MCP server's tool list
+- **Always in sync**: CLI automatically reflects API changes without code updates
+- **80+ commands**: Covers all resources (agents, campaigns, creatives, tactics, media buys, etc.)
+- **Smart caching**: 24-hour tool cache with fallback for offline use
+- **Type-safe parameters**: Automatic validation and parsing from server schemas
+- **Multiple output formats**: JSON and formatted table views
+- **Persistent configuration**: Save API keys and base URLs locally
+
+Usage:
+```bash
+# Configure once
+scope3 config set apiKey YOUR_KEY
+
+# Commands are dynamically discovered
+scope3 list-tools  # See all available operations
+scope3 brand-agent list
+scope3 campaign create --prompt "..." --brandAgentId 123
+scope3 media-buy execute --mediaBuyId "buy_123"
+```
+
+Benefits over static CLI:
+- API adds new endpoint → CLI instantly supports it (after cache refresh)
+- API changes parameter → CLI automatically updates validation
+- No manual maintenance of command definitions required
diff --git a/README.md b/README.md
index 3f65007..b06831f 100644
--- a/README.md
+++ b/README.md
@@ -11,6 +11,7 @@ TypeScript client for the Scope3 Agentic API with AdCP webhook support.
 - 🪝 Optional webhook server for AdCP events
 - ✨ Clean, intuitive API design
 - 🧪 Comprehensive test coverage
+- 💻 **CLI tool** for all API resources (80+ commands)
 
 **Architecture:** This client uses the official `@modelcontextprotocol/sdk` to connect to the Scope3 MCP server at `https://api.agentic.scope3.com/mcp` via Streamable HTTP transport. This uses HTTP POST for sending messages and HTTP GET with Server-Sent Events for receiving messages, providing reliable bidirectional communication with automatic reconnection support.
 
@@ -43,7 +44,29 @@ const campaign = await client.campaigns.create({
 });
 ```
 
-## Configuration
+## CLI Usage
+
+The CLI dynamically discovers available commands from the API server, ensuring it's always up-to-date:
+
+```bash
+# Install globally
+npm install -g @scope3/agentic-client
+
+# Configure authentication
+scope3 config set apiKey your_api_key_here
+
+# Discover available commands (80+ auto-generated)
+scope3 list-tools
+
+# Examples
+scope3 brand-agent list
+scope3 campaign create --prompt "Q1 2024 Spring Campaign" --brandAgentId 123
+scope3 media-buy execute --mediaBuyId "buy_123"
+```
+
+**Dynamic Updates:** Commands automatically stay in sync with API changes. No manual updates needed!
+
+## SDK Configuration
 
 ```typescript
 const client = new Scope3AgenticClient({
diff --git a/SIMPLE_MEDIA_AGENT.md b/SIMPLE_MEDIA_AGENT.md
deleted file mode 100644
index a6ead82..0000000
--- a/SIMPLE_MEDIA_AGENT.md
+++ /dev/null
@@ -1,201 +0,0 @@
-# Simple Media Agent
-
-A basic reference implementation of a media agent using MCP (Model Context Protocol).
-
-## Overview
-
-This media agent exposes MCP tools that Scope3 platform calls to manage media buying:
-
-- **get_proposed_tactics**: Fetches products from registered agents and proposes budget allocation based on floor prices
-- **manage_tactic**: When assigned, creates media buys by allocating budget to the N cheapest products with overallocation
-
-**Protocol**: MCP (stdio) - All communication via MCP, no HTTP server needed!
-
-## Algorithm
-
-### Budget Allocation
-- Fetches all products from registered sales agents
-- Sorts products by floor price (cheapest first)
-- **Applies overallocation** (default 40%) to ensure delivery targets are met
-- Calculates N = number of products where daily budget ≥ min daily budget (default $100)
-- Selects N cheapest products
-- Divides overallocated budget equally among N products
-
-**Example**: For a $10,000 campaign with 40% overallocation, the agent allocates $14,000 across media buys. This ensures you hit your $10,000 spend target even with underdelivery.
-
-## Installation
-
-```bash
-npm install @scope3/agentic-client
-```
-
-## Usage
-
-### As MCP Server (Standalone)
-
-```bash
-export SCOPE3_API_KEY=your_api_key
-export MIN_DAILY_BUDGET=100
-export OVERALLOCATION_PERCENT=40
-npx simple-media-agent
-```
-
-The agent runs as an MCP server on stdio. Scope3 platform calls it via MCP protocol.
-
-### Programmatically
-
-```typescript
-import { SimpleMediaAgent } from '@scope3/agentic-client';
-
-const agent = new SimpleMediaAgent({
-  scope3ApiKey: process.env.SCOPE3_API_KEY,
-  scope3BaseUrl: 'https://api.agentic.scope3.com',
-  minDailyBudget: 100,
-  overallocationPercent: 40,
-  name: 'my-media-agent',
-  version: '1.0.0',
-});
-
-await agent.start();
-```
-
-## Configuration
-
-### Environment Variables
-
-- `SCOPE3_API_KEY` (required): Your Scope3 API key
-- `SCOPE3_BASE_URL` (optional): Scope3 API base URL (default: https://api.agentic.scope3.com)
-- `MIN_DAILY_BUDGET` (optional): Minimum daily budget per product in USD (default: 100)
-- `OVERALLOCATION_PERCENT` (optional): Overallocation percentage to ensure delivery (default: 40)
-
-## MCP Tools
-
-The agent exposes two MCP tools:
-
-### `get_proposed_tactics`
-
-Get tactic proposals based on available products and floor prices.
-
-**Parameters:**
-- `campaignId` (required): Campaign ID
-- `seatId` (required): Seat/account ID
-- `budgetRange` (optional): Budget range with min/max/currency
-- `channels` (optional): Array of media channels
-- `countries` (optional): Array of ISO country codes
-
-**Returns:**
-```json
-{
-  "proposedTactics": [
-    {
-      "tacticId": "simple-passthrough-campaign-123",
-      "execution": "Passthrough strategy with 40% overallocation...",
-      "budgetCapacity": 50000,
-      "pricing": {
-        "method": "passthrough",
-        "estimatedCpm": 2.50,
-        "currency": "USD"
-      },
-      "metadata": {
-        "productCount": 25,
-        "avgFloorPrice": 2.50,
-        "overallocationPercent": 40
-      }
-    }
-  ]
-}
-```
-
-### `manage_tactic`
-
-Accept tactic assignment and create media buys.
-
-**Parameters:**
-- `tacticId` (required): Tactic ID from proposal
-- `tacticContext` (required): Tactic details including budget
-- `brandAgentId` (required): Brand agent ID
-- `seatId` (required): Seat/account ID
-
-**Returns:**
-```json
-{
-  "acknowledged": true,
-  "mediaBuysCreated": 5,
-  "allocations": [
-    {
-      "productId": "prod-123",
-      "budget": 2800,
-      "cpm": 2.10
-    }
-  ]
-}
-```
-
-## Architecture
-
-```
-Scope3 Platform → MCP (stdio) → Simple Media Agent → Scope3 API
-                                        ↓
-                                  Create Media Buys
-```
-
-The agent:
-1. Receives MCP tool calls from Scope3 platform
-2. Fetches products from registered agents via Scope3 API
-3. Calculates budget allocation with overallocation
-4. Creates media buys via Scope3 API
-5. Returns results via MCP response
-
-## Extending the Agent
-
-This is a reference implementation. To build your own:
-
-### Custom Product Selection
-```typescript
-// Filter products by viewability
-const highViewabilityProducts = allProducts.filter(p =>
-  p.metadata?.viewability >= 0.85
-);
-```
-
-### Custom Budget Allocation
-```typescript
-// Weight by performance, not just equal division
-const budgetPerProduct = allocatedBudget * product.performanceScore / totalScore;
-```
-
-### Add More Tools
-```typescript
-this.server.addTool({
-  name: 'reallocate_budget',
-  description: 'Reallocate budget based on performance',
-  parameters: z.object({
-    tacticId: z.string(),
-    performanceData: z.object({}).passthrough(),
-  }),
-  execute: async (args) => {
-    // Custom reallocation logic
-  },
-});
-```
-
-## Example: Claude Desktop Configuration
-
-Add to your Claude Desktop MCP config:
-
-```json
-{
-  "mcpServers": {
-    "simple-media-agent": {
-      "command": "npx",
-      "args": ["simple-media-agent"],
-      "env": {
-        "SCOPE3_API_KEY": "your_key",
-        "OVERALLOCATION_PERCENT": "40"
-      }
-    }
-  }
-}
-```
-
-Then in Claude: "Use get_proposed_tactics to propose tactics for campaign XYZ with $50,000 budget"
diff --git a/TESTING.md b/docs/TESTING.md
similarity index 100%
rename from TESTING.md
rename to docs/TESTING.md
diff --git a/package-lock.json b/package-lock.json
index 84bdfe0..7fe8e35 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,20 +1,24 @@
 {
   "name": "@scope3/agentic-client",
-  "version": "1.0.3",
+  "version": "1.0.4",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@scope3/agentic-client",
-      "version": "1.0.3",
+      "version": "1.0.4",
       "license": "MIT",
       "dependencies": {
         "@modelcontextprotocol/sdk": "^1.20.1",
+        "chalk": "^4.1.2",
+        "cli-table3": "^0.6.5",
+        "commander": "^14.0.2",
         "express": "^4.18.0",
         "fastmcp": "^3.20.2",
         "zod": "^3.25.76"
       },
       "bin": {
+        "scope3": "dist/cli.js",
         "simple-media-agent": "dist/simple-media-agent-server.js"
       },
       "devDependencies": {
@@ -929,6 +933,16 @@
         "url": "https://github.com/prettier/prettier?sponsor=1"
       }
     },
+    "node_modules/@colors/colors": {
+      "version": "1.5.0",
+      "resolved": "https://registry.npmjs.org/@colors/colors/-/colors-1.5.0.tgz",
+      "integrity": "sha512-ooWCrlZP11i8GImSjTHYHLkvFDP48nS4+204nGb1RiX/WXYHmJA2III9/e2DWVabCESdW7hBAEzHRqUn9OUVvQ==",
+      "license": "MIT",
+      "optional": true,
+      "engines": {
+        "node": ">=0.1.90"
+      }
+    },
     "node_modules/@eslint-community/eslint-utils": {
       "version": "4.9.0",
       "resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.9.0.tgz",
@@ -2681,7 +2695,6 @@
       "version": "5.0.1",
       "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz",
       "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">=8"
@@ -2691,7 +2704,6 @@
       "version": "4.3.0",
       "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz",
       "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "color-convert": "^2.0.1"
@@ -3105,7 +3117,6 @@
       "version": "4.1.2",
       "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz",
       "integrity": "sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "ansi-styles": "^4.1.0",
@@ -3174,6 +3185,21 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
+    "node_modules/cli-table3": {
+      "version": "0.6.5",
+      "resolved": "https://registry.npmjs.org/cli-table3/-/cli-table3-0.6.5.tgz",
+      "integrity": "sha512-+W/5efTR7y5HRD7gACw9yQjqMVvEMLBHmboM/kPWam+H+Hmyrgjh6YncVKK122YZkXrLudzTuAukUw9FnMf7IQ==",
+      "license": "MIT",
+      "dependencies": {
+        "string-width": "^4.2.0"
+      },
+      "engines": {
+        "node": "10.* || >= 12.*"
+      },
+      "optionalDependencies": {
+        "@colors/colors": "1.5.0"
+      }
+    },
     "node_modules/cli-truncate": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/cli-truncate/-/cli-truncate-4.0.0.tgz",
@@ -3282,7 +3308,6 @@
       "version": "2.0.1",
       "resolved": "https://registry.npmjs.org/color-convert/-/color-convert-2.0.1.tgz",
       "integrity": "sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "color-name": "~1.1.4"
@@ -3295,7 +3320,6 @@
       "version": "1.1.4",
       "resolved": "https://registry.npmjs.org/color-name/-/color-name-1.1.4.tgz",
       "integrity": "sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==",
-      "dev": true,
       "license": "MIT"
     },
     "node_modules/colorette": {
@@ -3305,6 +3329,15 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/commander": {
+      "version": "14.0.2",
+      "resolved": "https://registry.npmjs.org/commander/-/commander-14.0.2.tgz",
+      "integrity": "sha512-TywoWNNRbhoD0BXs1P3ZEScW8W5iKrnbithIl0YH+uCmBd0QpPOA8yc82DS3BIE5Ma6FnBVUsJ7wVUDz4dvOWQ==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=20"
+      }
+    },
     "node_modules/concat-map": {
       "version": "0.0.1",
       "resolved": "https://registry.npmjs.org/concat-map/-/concat-map-0.0.1.tgz",
@@ -3589,7 +3622,6 @@
       "version": "8.0.0",
       "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz",
       "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==",
-      "dev": true,
       "license": "MIT"
     },
     "node_modules/encodeurl": {
@@ -4840,7 +4872,6 @@
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/has-flag/-/has-flag-4.0.0.tgz",
       "integrity": "sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">=8"
@@ -5082,7 +5113,6 @@
       "version": "3.0.0",
       "resolved": "https://registry.npmjs.org/is-fullwidth-code-point/-/is-fullwidth-code-point-3.0.0.tgz",
       "integrity": "sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">=8"
@@ -8063,7 +8093,6 @@
       "version": "4.2.3",
       "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz",
       "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "emoji-regex": "^8.0.0",
@@ -8078,7 +8107,6 @@
       "version": "6.0.1",
       "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz",
       "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "ansi-regex": "^5.0.1"
@@ -8140,7 +8168,6 @@
       "version": "7.2.0",
       "resolved": "https://registry.npmjs.org/supports-color/-/supports-color-7.2.0.tgz",
       "integrity": "sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "has-flag": "^4.0.0"
diff --git a/package.json b/package.json
index bffee49..53521de 100644
--- a/package.json
+++ b/package.json
@@ -8,7 +8,8 @@
     "access": "public"
   },
   "bin": {
-    "simple-media-agent": "dist/simple-media-agent-server.js"
+    "simple-media-agent": "dist/simple-media-agent-server.js",
+    "scope3": "dist/cli.js"
   },
   "scripts": {
     "build": "npm run type-check && tsc",
@@ -46,6 +47,9 @@
   "homepage": "https://github.com/scope3data/agentic-client#readme",
   "dependencies": {
     "@modelcontextprotocol/sdk": "^1.20.1",
+    "chalk": "^4.1.2",
+    "cli-table3": "^0.6.5",
+    "commander": "^14.0.2",
     "express": "^4.18.0",
     "fastmcp": "^3.20.2",
     "zod": "^3.25.76"
diff --git a/scripts/README.md b/scripts/README.md
new file mode 100644
index 0000000..e382988
--- /dev/null
+++ b/scripts/README.md
@@ -0,0 +1,114 @@
+# Scope3 CLI Workflow Scripts
+
+This directory contains end-to-end workflow test scripts for the Scope3 CLI tool.
+
+## 🎯 Platform vs Partner Access
+
+Scope3 has two levels of API access:
+
+- **Platform API** - Create/manage brand agents, campaigns, creatives; read-only access to tactics and media buys
+- **Partner API** - Full access to create/manage tactics, media buys, and execute campaigns
+
+**Use the appropriate workflow for your access level!**
+
+## Available Scripts
+
+### `platform-workflow-test.sh` - Platform Workflow Test
+
+**For platform users with platform-level API keys**
+
+Demonstrates complete platform workflow:
+- ✅ Create and manage brand agents
+- ✅ Create and manage campaigns
+- ✅ Discover marketplace agents and products
+- ✅ View tactics and media buys (read-only)
+- ✅ Manage notifications
+
+**Usage:**
+
+```bash
+# Set your platform API key
+export SCOPE3_API_KEY=your_platform_api_key
+
+# Run the platform workflow test
+./scripts/platform-workflow-test.sh
+```
+
+**Expected output:**
+```
+==========================================
+    PLATFORM WORKFLOW TEST
+==========================================
+
+✓ Channels discovered (12 channels)
+✓ Brand agent created (ID: 3167)
+✓ Campaign created (ID: campaign_xxx)
+✓ Marketplace agents discovered
+✓ Tactics viewed (read-only)
+✓ Media buys viewed (read-only)
+
+✅ Platform workflow test successful!
+```
+
+### `partner-workflow-test.sh` - Partner Workflow Test
+
+**For partners with partner-level API keys**
+
+Demonstrates complete partner workflow:
+- ✅ Register and manage sales/outcome agents
+- ✅ Create and manage tactics
+- ✅ Create and manage media buys
+- ✅ Execute campaigns
+- ✅ Sync products
+
+**Usage:**
+
+```bash
+# Set your partner API key
+export SCOPE3_API_KEY=your_partner_api_key
+
+# Run the partner workflow test
+./scripts/partner-workflow-test.sh
+```
+
+**Note:** If you have a platform key and try to run this, it will show warnings explaining that partner operations require elevated permissions.
+
+## Quick Start
+
+### If you're a Platform User (Most Common)
+
+```bash
+export SCOPE3_API_KEY=your_key
+./scripts/platform-workflow-test.sh
+```
+
+### If you're a Partner
+
+```bash
+export SCOPE3_API_KEY=your_partner_key
+./scripts/partner-workflow-test.sh
+```
+
+### If you're unsure
+
+Run the platform workflow first. If you get permission errors on tactics/media buys **creation** (viewing is OK), you have platform access.
+
+## What Each Workflow Tests
+
+| Operation | Platform | Partner |
+|-----------|----------|---------|
+| List Channels | ✅ | ✅ |
+| Create Brand Agents | ✅ | ✅ |
+| Create Campaigns | ✅ | ✅ |
+| Register Agents | ❌ | ✅ |
+| Create Tactics | ❌ | ✅ |
+| Create Media Buys | ❌ | ✅ |
+| View Tactics | ✅ (read-only) | ✅ |
+| View Media Buys | ✅ (read-only) | ✅ |
+| Discover Marketplace | ✅ | ✅ |
+
+## See Also
+
+- **CLI Documentation**: See `../CLI.md` for complete CLI reference
+- **Workflow Guide**: See `../WORKFLOW_GUIDE.md` for detailed explanation of platform vs partner roles
+- **Status Report**: See `../CLI_STATUS.md` for current status of all operations
diff --git a/scripts/partner-workflow-test.sh b/scripts/partner-workflow-test.sh
new file mode 100755
index 0000000..d6bce9b
--- /dev/null
+++ b/scripts/partner-workflow-test.sh
@@ -0,0 +1,290 @@
+#!/bin/bash
+
+# Scope3 CLI - Partner Workflow Test
+# Tests operations available to partners (sales/outcome agents, tactics, media buys)
+
+set -e  # Exit on error
+
+# Colors for output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# Configuration
+CLI_CMD="node dist/cli.js"
+OUTPUT_FORMAT="json"
+
+# Function to print colored output
+print_step() {
+    echo -e "${BLUE}==>${NC} ${1}"
+}
+
+print_success() {
+    echo -e "${GREEN}✓${NC} ${1}"
+}
+
+print_error() {
+    echo -e "${RED}✗${NC} ${1}"
+}
+
+print_warning() {
+    echo -e "${YELLOW}⚠${NC} ${1}"
+}
+
+print_info() {
+    echo -e "${BLUE}ℹ${NC}  ${1}"
+}
+
+# Check if API key is set
+if [ -z "$SCOPE3_API_KEY" ]; then
+    print_error "SCOPE3_API_KEY environment variable is not set"
+    echo "Please set it with: export SCOPE3_API_KEY=your_partner_api_key"
+    exit 1
+fi
+
+print_success "API key found"
+echo ""
+echo "=========================================="
+echo "    PARTNER WORKFLOW TEST"
+echo "=========================================="
+echo ""
+echo "This test demonstrates partner-level operations:"
+echo "✓ Register and manage sales/outcome agents"
+echo "✓ Create and manage tactics"
+echo "✓ Create and manage media buys"
+echo "✓ Execute campaign via agents"
+echo ""
+print_warning "NOTE: This workflow requires a PARTNER-level API key"
+print_info "If you have a platform key, use platform-workflow-test.sh instead"
+echo ""
+
+# Step 1: List Existing Agents
+print_step "Step 1: Listing Registered Agents..."
+AGENTS_RESPONSE=$($CLI_CMD agent list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Agents listed successfully"
+    echo "$AGENTS_RESPONSE" | head -20
+else
+    print_warning "Failed to list agents"
+fi
+echo ""
+
+# Step 2: Register Sales Agent (Partner Operation)
+print_step "Step 2: Registering Sales Agent..."
+print_info "Attempting to register a sales agent (partner operation)"
+SALES_AGENT_RESPONSE=$($CLI_CMD agent register \
+    --type SALES \
+    --name "Partner Test Sales Agent $(date +%s)" \
+    --endpointUrl "https://example.com/sales-agent/mcp" \
+    --protocol MCP \
+    --format $OUTPUT_FORMAT 2>&1)
+
+if [ $? -eq 0 ]; then
+    print_success "Sales agent registered"
+    SALES_AGENT_ID=$(echo "$SALES_AGENT_RESPONSE" | grep -o '"agentId":"[^"]*"' | head -1 | sed 's/"agentId":"//' | sed 's/"//')
+    echo "Sales Agent ID: $SALES_AGENT_ID"
+else
+    print_warning "Failed to register sales agent (requires partner access)"
+    echo "$SALES_AGENT_RESPONSE"
+    SALES_AGENT_ID=""
+fi
+echo ""
+
+# Step 3: Register Outcome Agent (Partner Operation)
+print_step "Step 3: Registering Outcome Agent..."
+print_info "Attempting to register an outcome agent (partner operation)"
+OUTCOME_AGENT_RESPONSE=$($CLI_CMD agent register \
+    --type OUTCOME \
+    --name "Partner Test Outcome Agent $(date +%s)" \
+    --endpointUrl "https://example.com/outcome-agent/mcp" \
+    --protocol MCP \
+    --format $OUTPUT_FORMAT 2>&1)
+
+if [ $? -eq 0 ]; then
+    print_success "Outcome agent registered"
+    OUTCOME_AGENT_ID=$(echo "$OUTCOME_AGENT_RESPONSE" | grep -o '"agentId":"[^"]*"' | head -1 | sed 's/"agentId":"//' | sed 's/"//')
+    echo "Outcome Agent ID: $OUTCOME_AGENT_ID"
+else
+    print_warning "Failed to register outcome agent (requires partner access)"
+    echo "$OUTCOME_AGENT_RESPONSE"
+    OUTCOME_AGENT_ID=""
+fi
+echo ""
+
+# Step 4: Get Existing Campaign for Testing
+print_step "Step 4: Finding Existing Campaign..."
+CAMPAIGNS_RESPONSE=$($CLI_CMD campaign list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Campaigns retrieved"
+    # Try to extract first campaign ID
+    CAMPAIGN_ID=$(echo "$CAMPAIGNS_RESPONSE" | grep -o 'campaign_[a-zA-Z0-9_]*' | head -1)
+    if [ -n "$CAMPAIGN_ID" ]; then
+        echo "Using existing campaign: $CAMPAIGN_ID"
+    else
+        print_warning "No campaign found. Create one first with buyer workflow."
+        CAMPAIGN_ID=""
+    fi
+else
+    print_warning "Failed to list campaigns"
+    CAMPAIGN_ID=""
+fi
+echo ""
+
+# Step 5: Create Tactic (Partner Operation)
+if [ -n "$CAMPAIGN_ID" ]; then
+    print_step "Step 5: Creating Tactic (Partner Operation)..."
+    print_info "Attempting to create a tactic (requires partner access)"
+    TACTIC_RESPONSE=$($CLI_CMD tactic create \
+        --name "Partner Test Tactic $(date +%s)" \
+        --campaignId "$CAMPAIGN_ID" \
+        --prompt "Focus on high-impact video placements" \
+        --channelCodes video,display \
+        --format $OUTPUT_FORMAT 2>&1)
+
+    if [ $? -eq 0 ]; then
+        print_success "Tactic created"
+        TACTIC_ID=$(echo "$TACTIC_RESPONSE" | grep -o '"tacticId":"[^"]*"' | head -1 | sed 's/"tacticId":"//' | sed 's/"//')
+        if [ -z "$TACTIC_ID" ]; then
+            TACTIC_ID=$(echo "$TACTIC_RESPONSE" | grep -o 'tactic_[a-zA-Z0-9_]*' | head -1)
+        fi
+        echo "Tactic ID: $TACTIC_ID"
+    else
+        print_warning "Failed to create tactic (requires partner access)"
+        echo "$TACTIC_RESPONSE"
+        TACTIC_ID=""
+    fi
+else
+    print_warning "Step 5: Skipping tactic creation (no campaign available)"
+    TACTIC_ID=""
+fi
+echo ""
+
+# Step 6: List Tactics
+print_step "Step 6: Listing All Tactics..."
+$CLI_CMD tactic list --format $OUTPUT_FORMAT > /dev/null 2>&1
+if [ $? -eq 0 ]; then
+    print_success "Tactics listed successfully"
+else
+    print_warning "Failed to list tactics"
+fi
+echo ""
+
+# Step 7: Create Media Buy (Partner Operation)
+if [ -n "$TACTIC_ID" ] && [ -n "$SALES_AGENT_ID" ]; then
+    print_step "Step 7: Creating Media Buy (Partner Operation)..."
+    print_info "Attempting to create a media buy (requires partner access)"
+    MEDIA_BUY_RESPONSE=$($CLI_CMD media-buy create \
+        --tacticId "$TACTIC_ID" \
+        --name "Partner Test Media Buy" \
+        --media-product '[{"mediaProductId":"prod-123","salesAgentId":"'$SALES_AGENT_ID'"}]' \
+        --budget '{"amount":50000,"currency":"USD"}' \
+        --format $OUTPUT_FORMAT 2>&1)
+
+    if [ $? -eq 0 ]; then
+        print_success "Media buy created"
+        MEDIA_BUY_ID=$(echo "$MEDIA_BUY_RESPONSE" | grep -o '"mediaBuyId":"[^"]*"' | head -1 | sed 's/"mediaBuyId":"//' | sed 's/"//')
+        echo "Media Buy ID: $MEDIA_BUY_ID"
+    else
+        print_warning "Failed to create media buy (requires partner access and valid products)"
+        echo "$MEDIA_BUY_RESPONSE"
+        MEDIA_BUY_ID=""
+    fi
+else
+    print_warning "Step 7: Skipping media buy creation (missing tactic or sales agent)"
+    MEDIA_BUY_ID=""
+fi
+echo ""
+
+# Step 8: List Media Buys
+print_step "Step 8: Listing All Media Buys..."
+$CLI_CMD media-buy list --format $OUTPUT_FORMAT > /dev/null 2>&1
+if [ $? -eq 0 ]; then
+    print_success "Media buys listed successfully"
+else
+    print_warning "Failed to list media buys"
+fi
+echo ""
+
+# Step 9: Execute Media Buy (Partner Operation)
+if [ -n "$MEDIA_BUY_ID" ]; then
+    print_step "Step 9: Executing Media Buy (Partner Operation)..."
+    print_info "Attempting to execute media buy"
+    EXECUTE_RESPONSE=$($CLI_CMD media-buy execute \
+        --mediaBuyId "$MEDIA_BUY_ID" \
+        --format $OUTPUT_FORMAT 2>&1)
+
+    if [ $? -eq 0 ]; then
+        print_success "Media buy executed successfully"
+    else
+        print_warning "Failed to execute media buy"
+        echo "$EXECUTE_RESPONSE"
+    fi
+else
+    print_warning "Step 9: Skipping media buy execution (no media buy created)"
+fi
+echo ""
+
+# Step 10: Sync Products (Partner Operation)
+if [ -n "$SALES_AGENT_ID" ]; then
+    print_step "Step 10: Syncing Products from Sales Agent..."
+    SYNC_RESPONSE=$($CLI_CMD media-product sync \
+        --salesAgentId "$SALES_AGENT_ID" \
+        --format $OUTPUT_FORMAT 2>&1)
+
+    if [ $? -eq 0 ]; then
+        print_success "Products synced successfully"
+    else
+        print_warning "Failed to sync products"
+    fi
+else
+    print_warning "Step 10: Skipping product sync (no sales agent registered)"
+fi
+echo ""
+
+# Summary
+echo ""
+echo "=========================================="
+echo "    PARTNER WORKFLOW SUMMARY"
+echo "=========================================="
+echo ""
+
+if [ -n "$SALES_AGENT_ID" ] || [ -n "$OUTCOME_AGENT_ID" ] || [ -n "$TACTIC_ID" ] || [ -n "$MEDIA_BUY_ID" ]; then
+    print_success "Partner operations executed!"
+    echo ""
+    echo "✅ Created Resources:"
+    [ -n "$SALES_AGENT_ID" ] && echo "  • Sales Agent ID: $SALES_AGENT_ID"
+    [ -n "$OUTCOME_AGENT_ID" ] && echo "  • Outcome Agent ID: $OUTCOME_AGENT_ID"
+    [ -n "$TACTIC_ID" ] && echo "  • Tactic ID: $TACTIC_ID"
+    [ -n "$MEDIA_BUY_ID" ] && echo "  • Media Buy ID: $MEDIA_BUY_ID"
+    echo ""
+    echo "✅ Partner Capabilities Verified:"
+    echo "  • Register and manage sales/outcome agents"
+    echo "  • Create and manage tactics"
+    echo "  • Create and manage media buys"
+    echo "  • Execute campaigns"
+    echo ""
+    echo "📋 To clean up test resources:"
+    [ -n "$MEDIA_BUY_ID" ] && echo "  $CLI_CMD media-buy delete --mediaBuyId $MEDIA_BUY_ID"
+    [ -n "$TACTIC_ID" ] && echo "  $CLI_CMD tactic delete --tacticId $TACTIC_ID"
+    [ -n "$SALES_AGENT_ID" ] && echo "  $CLI_CMD agent unregister --agentId $SALES_AGENT_ID"
+    [ -n "$OUTCOME_AGENT_ID" ] && echo "  $CLI_CMD agent unregister --agentId $OUTCOME_AGENT_ID"
+    echo ""
+    print_success "Partner workflow test successful!"
+else
+    print_warning "No partner operations succeeded"
+    echo ""
+    echo "❌ Partner Operations Failed:"
+    echo "  • Could not register agents"
+    echo "  • Could not create tactics"
+    echo "  • Could not create media buys"
+    echo ""
+    print_info "Possible Reasons:"
+    echo "  1. API key does not have partner-level permissions"
+    echo "  2. Using a buyer-level API key instead of partner key"
+    echo "  3. Missing required configuration or setup"
+    echo ""
+    print_info "If you are a PLATFORM (not a partner), use platform-workflow-test.sh instead"
+    echo ""
+fi
diff --git a/scripts/platform-workflow-test.sh b/scripts/platform-workflow-test.sh
new file mode 100755
index 0000000..a5e426a
--- /dev/null
+++ b/scripts/platform-workflow-test.sh
@@ -0,0 +1,229 @@
+#!/bin/bash
+
+# Scope3 CLI - Platform Workflow Test
+# Tests operations available to platform users (brand agents, campaigns, creatives, discovery)
+
+set -e  # Exit on error
+
+# Colors for output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# Configuration
+CLI_CMD="node dist/cli.js"
+OUTPUT_FORMAT="json"
+
+# Function to print colored output
+print_step() {
+    echo -e "${BLUE}==>${NC} ${1}"
+}
+
+print_success() {
+    echo -e "${GREEN}✓${NC} ${1}"
+}
+
+print_error() {
+    echo -e "${RED}✗${NC} ${1}"
+}
+
+print_warning() {
+    echo -e "${YELLOW}⚠${NC} ${1}"
+}
+
+# Function to extract value from JSON using basic tools
+extract_json_value() {
+    local json=$1
+    local key=$2
+    echo "$json" | grep -o "\"$key\":[^,}]*" | head -1 | sed 's/.*: *"\?\([^"]*\)"\?.*/\1/' | sed 's/[^a-zA-Z0-9_-]//g'
+}
+
+# Check if API key is set
+if [ -z "$SCOPE3_API_KEY" ]; then
+    print_error "SCOPE3_API_KEY environment variable is not set"
+    echo "Please set it with: export SCOPE3_API_KEY=your_api_key"
+    exit 1
+fi
+
+print_success "API key found"
+echo ""
+echo "=========================================="
+echo "    PLATFORM WORKFLOW TEST"
+echo "=========================================="
+echo ""
+echo "This test demonstrates platform-level operations:"
+echo "✓ Brand agent management"
+echo "✓ Campaign creation and management"
+echo "✓ Marketplace discovery"
+echo "✓ Read-only access to tactic and media buys"
+echo ""
+
+# Step 1: List available channels
+print_step "Step 1: Discovering Available Channels..."
+CHANNELS_RESPONSE=$($CLI_CMD channel list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Channels discovered"
+    echo "Available channels: display, ctv, video, audio, social, dooh, etc."
+else
+    print_error "Failed to list channels"
+    echo "$CHANNELS_RESPONSE"
+    exit 1
+fi
+echo ""
+
+# Step 2: Create Brand Agent
+print_step "Step 2: Creating Brand Agent..."
+BRAND_AGENT_RESPONSE=$($CLI_CMD brand-agent create \
+    --name "Test Platform Agent $(date +%s)" \
+    --description "Automated platform workflow test" \
+    --manifestUrl "https://example.com/manifest.json" \
+    --format $OUTPUT_FORMAT 2>&1)
+
+if [ $? -eq 0 ]; then
+    print_success "Brand agent created"
+    BRAND_AGENT_ID=$(echo "$BRAND_AGENT_RESPONSE" | grep -o '"id":[0-9]*' | head -1 | sed 's/"id"://')
+    if [ -z "$BRAND_AGENT_ID" ]; then
+        BRAND_AGENT_ID=$(echo "$BRAND_AGENT_RESPONSE" | grep -o 'ID: [0-9]*' | head -1 | sed 's/ID: //')
+    fi
+    echo "Brand Agent ID: $BRAND_AGENT_ID"
+else
+    print_error "Failed to create brand agent"
+    echo "$BRAND_AGENT_RESPONSE"
+    exit 1
+fi
+echo ""
+
+# Step 3: List Brand Agents
+print_step "Step 3: Verifying Brand Agent in List..."
+$CLI_CMD brand-agent list --format $OUTPUT_FORMAT > /dev/null 2>&1
+if [ $? -eq 0 ]; then
+    print_success "Brand agent listed successfully"
+else
+    print_warning "Failed to list brand agents"
+fi
+echo ""
+
+# Step 4: Create Campaign
+print_step "Step 4: Creating Campaign..."
+CAMPAIGN_RESPONSE=$($CLI_CMD campaign create \
+    --prompt "Q1 2024 awareness campaign targeting eco-conscious millennials across digital channels with focus on video and social media" \
+    --brandAgentId "$BRAND_AGENT_ID" \
+    --name "Buyer Test Campaign $(date +%s)" \
+    --format $OUTPUT_FORMAT 2>&1)
+
+if [ $? -eq 0 ]; then
+    print_success "Campaign created"
+    CAMPAIGN_ID=$(echo "$CAMPAIGN_RESPONSE" | grep -o 'campaign_[a-zA-Z0-9_]*' | head -1)
+    echo "Campaign ID: $CAMPAIGN_ID"
+else
+    print_error "Failed to create campaign"
+    echo "$CAMPAIGN_RESPONSE"
+    exit 1
+fi
+echo ""
+
+# Step 5: List Campaigns
+print_step "Step 5: Listing All Campaigns..."
+$CLI_CMD campaign list --format $OUTPUT_FORMAT > /dev/null 2>&1
+if [ $? -eq 0 ]; then
+    print_success "Campaigns listed successfully"
+else
+    print_warning "Failed to list campaigns"
+fi
+echo ""
+
+# Step 6: Get Campaign Summary
+print_step "Step 6: Getting Campaign Summary..."
+SUMMARY_RESPONSE=$($CLI_CMD campaign-get summary --campaignId "$CAMPAIGN_ID" --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Campaign summary retrieved"
+else
+    print_warning "Failed to get campaign summary (may need time for data)"
+fi
+echo ""
+
+# Step 7: Discover Marketplace Agents
+print_step "Step 7: Discovering Marketplace Agents..."
+AGENTS_RESPONSE=$($CLI_CMD agent list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Marketplace agent discovered"
+    echo "Found sales agent available for partnerships"
+else
+    print_warning "Failed to discover agents"
+fi
+echo ""
+
+# Step 8: Discover Media Products
+print_step "Step 8: Discovering Available Media Products..."
+PRODUCTS_RESPONSE=$($CLI_CMD media-product discover --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Media media-product discovered"
+    echo "Found available inventory from sales agents"
+else
+    print_warning "Failed to discover media-product (may require registered sales agents)"
+fi
+echo ""
+
+# Step 9: View Tactics (Read-Only)
+print_step "Step 9: Viewing Tactics (Read-Only Access)..."
+print_warning "Note: Platform users can only READ tactics, not create them"
+TACTICS_RESPONSE=$($CLI_CMD tactic list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Tactics viewed successfully (read-only)"
+else
+    print_warning "No tactic available or access denied"
+fi
+echo ""
+
+# Step 10: View Media Buys (Read-Only)
+print_step "Step 10: Viewing Media Buys (Read-Only Access)..."
+print_warning "Note: Platform users can only READ media buys, not create them"
+MEDIA_BUYS_RESPONSE=$($CLI_CMD media-buy list --format $OUTPUT_FORMAT 2>&1)
+if [ $? -eq 0 ]; then
+    print_success "Media buys viewed successfully (read-only)"
+else
+    print_warning "No media buys available or access denied"
+fi
+echo ""
+
+# Step 11: Check Notifications
+print_step "Step 11: Checking Notifications..."
+$CLI_CMD notifications list --limit 5 --format $OUTPUT_FORMAT > /dev/null 2>&1
+if [ $? -eq 0 ]; then
+    print_success "Notifications retrieved"
+else
+    print_warning "Failed to list notifications"
+fi
+echo ""
+
+# Summary
+echo ""
+echo "=========================================="
+echo "    PLATFORM WORKFLOW SUMMARY"
+echo "=========================================="
+echo ""
+print_success "Buyer workflow test completed!"
+echo ""
+echo "✅ Created Resources:"
+echo "  • Brand Agent ID: $BRAND_AGENT_ID"
+echo "  • Campaign ID: $CAMPAIGN_ID"
+echo ""
+echo "✅ Buyer Capabilities Verified:"
+echo "  • Create and manage brand agents"
+echo "  • Create and manage campaigns"
+echo "  • Discover marketplace agent and products"
+echo "  • View tactic and media buys (read-only)"
+echo "  • Manage notifications"
+echo ""
+echo "ℹ️  Buyer Limitations:"
+echo "  • Cannot create tactic (partner operation)"
+echo "  • Cannot create media buys (partner operation)"
+echo "  • Cannot execute campaign directly (requires partner agents)"
+echo ""
+echo "📋 To clean up test resources:"
+echo "  $CLI_CMD campaign delete --campaignId $CAMPAIGN_ID"
+echo "  $CLI_CMD brand-agent delete --brandAgentId $BRAND_AGENT_ID"
+echo ""
+print_success "Buyer workflow test successful!"
diff --git a/src/cli.ts b/src/cli.ts
new file mode 100644
index 0000000..c7127d4
--- /dev/null
+++ b/src/cli.ts
@@ -0,0 +1,534 @@
+#!/usr/bin/env node
+
+import { Command } from 'commander';
+import { Scope3AgenticClient } from './sdk';
+import Table from 'cli-table3';
+import chalk from 'chalk';
+import * as fs from 'fs';
+import * as path from 'path';
+import * as os from 'os';
+import { logger } from './utils/logger';
+
+// Configuration file location
+const CONFIG_DIR = path.join(os.homedir(), '.scope3');
+const CONFIG_FILE = path.join(CONFIG_DIR, 'config.json');
+const TOOLS_CACHE_FILE = path.join(CONFIG_DIR, 'tools-cache.json');
+const CACHE_TTL = 24 * 60 * 60 * 1000; // 24 hours
+
+interface CliConfig {
+  apiKey?: string;
+  baseUrl?: string;
+}
+
+interface ToolsCache {
+  tools: McpTool[];
+  timestamp: number;
+}
+
+interface McpTool {
+  name: string;
+  description?: string;
+  inputSchema: {
+    type: string;
+    properties?: Record<string, unknown>;
+    required?: string[];
+  };
+}
+
+// Load config from file or environment
+function loadConfig(): CliConfig {
+  const config: CliConfig = {};
+
+  // Try to load from config file
+  if (fs.existsSync(CONFIG_FILE)) {
+    try {
+      const fileConfig = JSON.parse(fs.readFileSync(CONFIG_FILE, 'utf-8'));
+      config.apiKey = fileConfig.apiKey;
+      config.baseUrl = fileConfig.baseUrl;
+    } catch (error) {
+      logger.warn('Failed to parse config file', { error });
+      console.error(chalk.yellow('Warning: Failed to parse config file'));
+    }
+  }
+
+  // Environment variables override config file
+  if (process.env.SCOPE3_API_KEY) {
+    config.apiKey = process.env.SCOPE3_API_KEY;
+  }
+  if (process.env.SCOPE3_BASE_URL) {
+    config.baseUrl = process.env.SCOPE3_BASE_URL;
+  }
+
+  return config;
+}
+
+// Save config to file
+function saveConfig(config: CliConfig): void {
+  if (!fs.existsSync(CONFIG_DIR)) {
+    fs.mkdirSync(CONFIG_DIR, { recursive: true });
+  }
+  fs.writeFileSync(CONFIG_FILE, JSON.stringify(config, null, 2));
+  console.log(chalk.green(`Configuration saved to ${CONFIG_FILE}`));
+}
+
+// Load tools cache
+function loadToolsCache(): ToolsCache | null {
+  if (!fs.existsSync(TOOLS_CACHE_FILE)) {
+    return null;
+  }
+
+  try {
+    const cache: ToolsCache = JSON.parse(fs.readFileSync(TOOLS_CACHE_FILE, 'utf-8'));
+    const age = Date.now() - cache.timestamp;
+
+    if (age > CACHE_TTL) {
+      return null; // Cache expired
+    }
+
+    return cache;
+  } catch (error) {
+    return null;
+  }
+}
+
+// Save tools cache
+function saveToolsCache(tools: McpTool[]): void {
+  if (!fs.existsSync(CONFIG_DIR)) {
+    fs.mkdirSync(CONFIG_DIR, { recursive: true });
+  }
+
+  const cache: ToolsCache = {
+    tools,
+    timestamp: Date.now(),
+  };
+
+  fs.writeFileSync(TOOLS_CACHE_FILE, JSON.stringify(cache, null, 2));
+}
+
+// Fetch available tools from MCP server
+async function fetchAvailableTools(
+  client: Scope3AgenticClient,
+  useCache = true
+): Promise<McpTool[]> {
+  // Try cache first
+  if (useCache) {
+    const cache = loadToolsCache();
+    if (cache) {
+      return cache.tools;
+    }
+  }
+
+  // Fetch from server
+  try {
+    await client.connect();
+    const response = (await client.listTools()) as { tools: McpTool[] };
+    const tools = response.tools;
+
+    // Save to cache
+    saveToolsCache(tools);
+
+    return tools;
+  } catch (error) {
+    const errorMessage = error instanceof Error ? error.message : String(error);
+    logger.error('Error fetching tools', error);
+    console.error(chalk.red('Error fetching tools:'), errorMessage);
+
+    // Try to use stale cache as fallback
+    if (fs.existsSync(TOOLS_CACHE_FILE)) {
+      console.log(chalk.yellow('Using cached tools (may be outdated)'));
+      const cache: ToolsCache = JSON.parse(fs.readFileSync(TOOLS_CACHE_FILE, 'utf-8'));
+      return cache.tools;
+    }
+
+    throw error;
+  }
+}
+
+// Format output based on format option
+function formatOutput(data: unknown, format: string): void {
+  if (format === 'json') {
+    console.log(JSON.stringify(data, null, 2));
+    return;
+  }
+
+  // Table format
+  if (!data) {
+    console.log(chalk.yellow('No data to display'));
+    return;
+  }
+
+  // Handle ToolResponse wrapper
+  const dataObj = data as Record<string, unknown>;
+  const actualData = dataObj.data || data;
+
+  if (Array.isArray(actualData)) {
+    if (actualData.length === 0) {
+      console.log(chalk.yellow('No results found'));
+      return;
+    }
+
+    // Create table from array
+    const keys = Object.keys(actualData[0]);
+    const table = new Table({
+      head: keys.map((k) => chalk.cyan(k)),
+      wordWrap: true,
+      wrapOnWordBoundary: false,
+    });
+
+    actualData.forEach((item) => {
+      table.push(
+        keys.map((k) => {
+          const value = item[k];
+          if (value === null || value === undefined) return '';
+          if (typeof value === 'object') return JSON.stringify(value);
+          return String(value);
+        })
+      );
+    });
+
+    console.log(table.toString());
+  } else if (typeof actualData === 'object') {
+    // Create table for single object
+    const table = new Table({
+      wordWrap: true,
+      wrapOnWordBoundary: false,
+    });
+
+    Object.entries(actualData).forEach(([key, value]) => {
+      let displayValue: string;
+      if (value === null || value === undefined) {
+        displayValue = '';
+      } else if (typeof value === 'object') {
+        displayValue = JSON.stringify(value, null, 2);
+      } else {
+        displayValue = String(value);
+      }
+      table.push({ [chalk.cyan(key)]: displayValue });
+    });
+
+    console.log(table.toString());
+  } else {
+    console.log(actualData);
+  }
+
+  // Show success/message if present
+  if (dataObj.success !== undefined) {
+    console.log(dataObj.success ? chalk.green('✓ Success') : chalk.red('✗ Failed'));
+  }
+  if (dataObj.message) {
+    console.log(chalk.blue('Message:'), dataObj.message);
+  }
+}
+
+// Create client instance
+function createClient(apiKey?: string, baseUrl?: string): Scope3AgenticClient {
+  const config = loadConfig();
+
+  const finalApiKey = apiKey || config.apiKey;
+  if (!finalApiKey) {
+    console.error(chalk.red('Error: API key is required'));
+    console.log('Set it via:');
+    console.log('  - Environment variable: export SCOPE3_API_KEY=your_key');
+    console.log('  - Config command: scope3 config set apiKey your_key');
+    console.log('  - Flag: --api-key your_key');
+    process.exit(1);
+  }
+
+  return new Scope3AgenticClient({
+    apiKey: finalApiKey,
+    baseUrl: baseUrl || config.baseUrl,
+  });
+}
+
+// Parse parameter value based on schema type
+function parseParameterValue(value: string, schema: Record<string, unknown>): unknown {
+  const type = schema.type as string;
+
+  if (type === 'object' || type === 'array') {
+    try {
+      return JSON.parse(value);
+    } catch (error) {
+      logger.error('Invalid JSON parameter', error, { value, type });
+      console.error(chalk.red(`Error: Invalid JSON for parameter: ${value}`));
+      process.exit(1);
+    }
+  }
+
+  if (type === 'integer' || type === 'number') {
+    const num = Number(value);
+    if (isNaN(num)) {
+      logger.error('Invalid number parameter', undefined, { value, type });
+      console.error(chalk.red(`Error: Invalid number: ${value}`));
+      process.exit(1);
+    }
+    return num;
+  }
+
+  if (type === 'boolean') {
+    if (value === 'true') return true;
+    if (value === 'false') return false;
+    logger.error('Invalid boolean parameter', undefined, { value, type });
+    console.error(chalk.red(`Error: Invalid boolean (use 'true' or 'false'): ${value}`));
+    process.exit(1);
+  }
+
+  // Default to string
+  return value;
+}
+
+// Parse tool name into resource and method
+function parseToolName(toolName: string): { resource: string; method: string } {
+  const parts = toolName.split('_');
+
+  if (parts.length < 2) {
+    return { resource: 'tools', method: toolName };
+  }
+
+  // Handle cases like "campaigns_create", "brand_agents_list", etc.
+  const method = parts[parts.length - 1];
+  const resource = parts.slice(0, -1).join('-');
+
+  return { resource, method };
+}
+
+// Main program
+const program = new Command();
+
+program
+  .name('scope3')
+  .description('CLI tool for Scope3 Agentic API (dynamically generated from MCP server)')
+  .version('1.0.0')
+  .option('--api-key <key>', 'API key for authentication')
+  .option('--base-url <url>', 'Base URL for API (default: production)')
+  .option('--format <format>', 'Output format: json or table', 'table')
+  .option('--no-cache', 'Skip cache and fetch fresh tools list');
+
+// Config command
+const configCmd = program.command('config').description('Manage CLI configuration');
+
+configCmd
+  .command('set')
+  .description('Set configuration value')
+  .argument('<key>', 'Configuration key (apiKey or baseUrl)')
+  .argument('<value>', 'Configuration value')
+  .action((key: string, value: string) => {
+    const config = loadConfig();
+    if (key === 'apiKey') {
+      config.apiKey = value;
+    } else if (key === 'baseUrl') {
+      config.baseUrl = value;
+    } else {
+      console.error(chalk.red(`Error: Unknown config key: ${key}`));
+      console.log('Valid keys: apiKey, baseUrl');
+      process.exit(1);
+    }
+    saveConfig(config);
+  });
+
+configCmd
+  .command('get')
+  .description('Get configuration value')
+  .argument('[key]', 'Configuration key (apiKey or baseUrl). If omitted, shows all config')
+  .action((key?: string) => {
+    const config = loadConfig();
+    if (!key) {
+      console.log(JSON.stringify(config, null, 2));
+    } else if (key in config) {
+      console.log(config[key as keyof CliConfig]);
+    } else {
+      console.error(chalk.red(`Error: Unknown config key: ${key}`));
+      process.exit(1);
+    }
+  });
+
+configCmd
+  .command('clear')
+  .description('Clear all configuration')
+  .action(() => {
+    if (fs.existsSync(CONFIG_FILE)) {
+      fs.unlinkSync(CONFIG_FILE);
+      console.log(chalk.green('Configuration cleared'));
+    } else {
+      console.log(chalk.yellow('No configuration file found'));
+    }
+  });
+
+// List available tools command
+program
+  .command('list-tools')
+  .description('List all available API tools')
+  .option('--refresh', 'Refresh tools cache')
+  .action(async (options) => {
+    const globalOpts = program.opts();
+    const client = createClient(globalOpts.apiKey, globalOpts.baseUrl);
+
+    try {
+      const useCache = !options.refresh && globalOpts.cache !== false;
+      const tools = await fetchAvailableTools(client, useCache);
+
+      console.log(chalk.green(`\nFound ${tools.length} available tools:\n`));
+
+      // Group by resource
+      const grouped: Record<string, McpTool[]> = {};
+      tools.forEach((tool) => {
+        const { resource } = parseToolName(tool.name);
+        if (!grouped[resource]) {
+          grouped[resource] = [];
+        }
+        grouped[resource].push(tool);
+      });
+
+      // Display grouped
+      Object.entries(grouped)
+        .sort()
+        .forEach(([resource, resourceTools]) => {
+          console.log(chalk.cyan.bold(`\n${resource}:`));
+          resourceTools.forEach((tool) => {
+            const { method } = parseToolName(tool.name);
+            const desc = tool.description || 'No description';
+            console.log(`  ${chalk.yellow(method)} - ${desc}`);
+          });
+        });
+
+      console.log();
+    } catch (error) {
+      const errorMessage = error instanceof Error ? error.message : String(error);
+      console.error(chalk.red('Error:'), errorMessage);
+      process.exit(1);
+    } finally {
+      await client.disconnect();
+    }
+  });
+
+// Dynamic command generation
+async function setupDynamicCommands() {
+  const globalOpts = program.opts();
+
+  // For help/version/config commands, don't fetch tools
+  const args = process.argv.slice(2);
+  if (
+    args.length === 0 ||
+    args.includes('--help') ||
+    args.includes('-h') ||
+    args.includes('--version') ||
+    args.includes('-V') ||
+    args[0] === 'config' ||
+    args[0] === 'list-tools'
+  ) {
+    return;
+  }
+
+  try {
+    const client = createClient(globalOpts.apiKey, globalOpts.baseUrl);
+    const useCache = globalOpts.cache !== false;
+    const tools = await fetchAvailableTools(client, useCache);
+    await client.disconnect();
+
+    // Group tools by resource
+    const resourceGroups: Record<string, McpTool[]> = {};
+    tools.forEach((tool) => {
+      const { resource } = parseToolName(tool.name);
+      if (!resourceGroups[resource]) {
+        resourceGroups[resource] = [];
+      }
+      resourceGroups[resource].push(tool);
+    });
+
+    // Create commands for each resource
+    Object.entries(resourceGroups).forEach(([resourceName, resourceTools]) => {
+      const resourceCmd = program
+        .command(resourceName)
+        .description(`Manage ${resourceName} (${resourceTools.length} operations)`);
+
+      resourceTools.forEach((tool) => {
+        const { method } = parseToolName(tool.name);
+        const cmd = resourceCmd
+          .command(method)
+          .description(tool.description || `${method} operation`);
+
+        // Add options from schema
+        const properties = (tool.inputSchema.properties || {}) as Record<
+          string,
+          Record<string, unknown>
+        >;
+        const required = tool.inputSchema.required || [];
+
+        Object.entries(properties).forEach(([paramName, paramSchema]) => {
+          const isRequired = required.includes(paramName);
+          const paramType = paramSchema.type as string;
+          const paramDesc = (paramSchema.description as string) || paramName;
+
+          const flag = `--${paramName} <value>`;
+          const description = `${paramDesc}${isRequired ? ' (required)' : ' (optional)'} [${paramType}]`;
+
+          cmd.option(flag, description);
+        });
+
+        // Action handler
+        cmd.action(async (options) => {
+          const client = createClient(globalOpts.apiKey, globalOpts.baseUrl);
+
+          try {
+            await client.connect();
+
+            // Build request from options
+            const request: Record<string, unknown> = {};
+            Object.entries(properties).forEach(([paramName, paramSchema]) => {
+              const value = options[paramName];
+              if (value !== undefined) {
+                request[paramName] = parseParameterValue(
+                  value,
+                  paramSchema as Record<string, unknown>
+                );
+              }
+            });
+
+            // Validate required params
+            const missing = required.filter((p) => request[p] === undefined);
+            if (missing.length > 0) {
+              console.error(chalk.red(`Error: Missing required parameters: ${missing.join(', ')}`));
+              process.exit(1);
+            }
+
+            // Call the tool
+            const result = await client.callTool(tool.name, request);
+            formatOutput(result, globalOpts.format);
+          } catch (error: unknown) {
+            const errorMessage = error instanceof Error ? error.message : String(error);
+            const errorStack = error instanceof Error ? error.stack : undefined;
+            logger.error('Tool execution failed', error, { toolName: tool.name });
+            console.error(chalk.red('Error:'), errorMessage);
+            if (errorStack && process.env.DEBUG) {
+              console.error(chalk.gray(errorStack));
+            }
+            process.exit(1);
+          } finally {
+            await client.disconnect();
+          }
+        });
+      });
+    });
+  } catch (error) {
+    // If we can't fetch tools, show a helpful error
+    const errorMessage = error instanceof Error ? error.message : String(error);
+    logger.error('CLI initialization failed', error);
+    console.error(chalk.red('Error initializing CLI:'), errorMessage);
+    console.log(chalk.yellow('\nMake sure your API key is configured:'));
+    console.log('  scope3 config set apiKey YOUR_KEY');
+    console.log('\nOr set via environment:');
+    console.log('  export SCOPE3_API_KEY=YOUR_KEY');
+    process.exit(1);
+  }
+}
+
+// Setup and parse
+setupDynamicCommands()
+  .then(() => {
+    program.parse();
+  })
+  .catch((error) => {
+    const errorMessage = error instanceof Error ? error.message : String(error);
+    logger.error('Fatal CLI error', error);
+    console.error(chalk.red('Fatal error:'), errorMessage);
+    process.exit(1);
+  });
diff --git a/src/client.ts b/src/client.ts
index 36ca043..359156c 100644
--- a/src/client.ts
+++ b/src/client.ts
@@ -28,7 +28,7 @@ export class Scope3Client {
     this.transport = new StreamableHTTPClientTransport(new URL(`${baseURL}/mcp`), {
       requestInit: {
         headers: {
-          Authorization: `Bearer ${this.apiKey}`,
+          'x-scope3-api-key': this.apiKey,
         },
       },
     });
diff --git a/src/sdk.ts b/src/sdk.ts
index d8b9dc9..f2b7dc4 100644
--- a/src/sdk.ts
+++ b/src/sdk.ts
@@ -43,4 +43,19 @@ export class Scope3AgenticClient extends Scope3Client {
     this.notifications = new NotificationsResource(this);
     this.products = new ProductsResource(this);
   }
+
+  // Expose MCP methods for CLI dynamic command generation
+  async listTools(): Promise<unknown> {
+    if (!this.getClient()) {
+      await this.connect();
+    }
+    return this.getClient().listTools();
+  }
+
+  async callTool<TRequest = Record<string, unknown>, TResponse = unknown>(
+    toolName: string,
+    args: TRequest
+  ): Promise<TResponse> {
+    return super.callTool(toolName, args);
+  }
 }
diff --git a/src/utils/logger.ts b/src/utils/logger.ts
new file mode 100644
index 0000000..74d8346
--- /dev/null
+++ b/src/utils/logger.ts
@@ -0,0 +1,79 @@
+/**
+ * Simple structured logging utility
+ *
+ * Outputs logs in JSON format in production, human-readable in development.
+ */
+
+type LogSeverity = 'DEBUG' | 'INFO' | 'WARNING' | 'ERROR';
+
+export class Logger {
+  private static instance: Logger;
+  private readonly isDevelopment: boolean;
+
+  constructor() {
+    this.isDevelopment = process.env.NODE_ENV === 'development';
+  }
+
+  static getInstance(): Logger {
+    if (!Logger.instance) {
+      Logger.instance = new Logger();
+    }
+    return Logger.instance;
+  }
+
+  debug(message: string, data?: Record<string, unknown>): void {
+    this.log('DEBUG', message, data);
+  }
+
+  info(message: string, data?: Record<string, unknown>): void {
+    this.log('INFO', message, data);
+  }
+
+  warn(message: string, data?: Record<string, unknown>): void {
+    this.log('WARNING', message, data);
+  }
+
+  error(message: string, error?: unknown, data?: Record<string, unknown>): void {
+    const errorData: Record<string, unknown> = { ...data };
+
+    if (error instanceof Error) {
+      errorData.error = {
+        message: error.message,
+        name: error.name,
+        stack: error.stack,
+      };
+    } else if (error) {
+      errorData.error = String(error);
+    }
+
+    this.log('ERROR', message, errorData);
+  }
+
+  private log(severity: LogSeverity, message: string, data?: Record<string, unknown>): void {
+    if (this.isDevelopment) {
+      // Development: Human-readable format
+      const prefix = `[${severity}]`;
+      const timestamp = new Date().toISOString();
+      const logMessage = `${timestamp} ${prefix} ${message}`;
+
+      if (data) {
+        console.error(logMessage, JSON.stringify(data, null, 2));
+      } else {
+        console.error(logMessage);
+      }
+    } else {
+      // Production: Structured JSON
+      const structuredLog: Record<string, unknown> = {
+        message,
+        severity,
+        timestamp: new Date().toISOString(),
+        ...data,
+      };
+
+      console.error(JSON.stringify(structuredLog));
+    }
+  }
+}
+
+// Export singleton instance
+export const logger = Logger.getInstance();