Linked Firefly MCP

Connect Claude Code to Firefly AI services for intelligent asset discovery, infrastructure analysis, and cloud resource management through the Model Context Protocol.

When to Use This MCP Server

Connect this server when you need to:

• Discover and analyze cloud infrastructure assets across AWS, Azure, GCP, and Kubernetes environments directly from your development workflow
• Get AI-powered recommendations for infrastructure optimization, cost reduction, and security improvements while reviewing infrastructure code
• Map dependencies between cloud resources, understand blast radius of changes, and validate infrastructure configurations before deployment

Consider alternatives when:

• You need to manage a single cloud provider exclusively, where the provider-specific MCP server (AWS, Azure, GCP) offers deeper native integration
• Your infrastructure management relies entirely on Terraform or Pulumi CLI tools that handle resource state management directly

Quick Start

Configuration

Copy
name: linked-firefly-mcp
type: mcp
category: devtools

Example Connection

Copy
claude mcp:connect linked-firefly-mcp

Available Tools

discover-assets:     Scan and catalog cloud resources across connected cloud accounts
analyze-resource:    Get detailed analysis of a specific cloud resource with recommendations
find-dependencies:   Map resource dependencies and identify connected infrastructure
cost-analysis:       Analyze resource costs and identify optimization opportunities
security-scan:       Evaluate resource configurations against security best practices

Core Concepts

Firefly MCP Integration Overview

Aspect	Details
Transport	stdio-based process via npx running @fireflyai/firefly-mcp package
Authentication	Access key and secret key pair for Firefly API authentication
Multi-Cloud	Unified view across AWS, Azure, GCP, and Kubernetes environments
AI Analysis	Intelligent recommendations for cost, security, and operational improvements
Asset Catalog	Complete inventory of cloud resources with relationship mapping

Firefly Cloud Intelligence Architecture

Claude Code IDE
    |
    v
[MCP Client] --stdio--> [Firefly MCP Server]
                              |
                              v
                     [Firefly AI Platform]
                              |
                    +---------+---------+
                    |         |         |
                    v         v         v
              [AWS        [Azure    [GCP/K8s
              Scanner]    Scanner]  Scanner]
                    \         |         /
                     [Unified Asset Catalog]
                      Dependencies | Costs

Configuration

Parameter	Type	Default	Description
FIREFLY_ACCESS_KEY	string	required	Firefly platform access key for API authentication
FIREFLY_SECRET_KEY	string	required	Firefly platform secret key paired with the access key
default_account	string	all	Default cloud account to scope queries; use "all" for cross-account
include_unmanaged	boolean	true	Include resources not managed by IaC tools in discovery results
cost_period	string	last30d	Default time period for cost analysis: last7d, last30d, or last90d

Best Practices

1. Run Discovery Before Infrastructure Changes - Use discover-assets to get a current inventory of your cloud resources before making infrastructure modifications. This baseline helps you understand what exists and prevents accidental conflicts.

2. Check Dependencies Before Deleting Resources - Use find-dependencies to map what other resources depend on the one you plan to remove. Deleting a resource with active dependents can cause cascading failures across your infrastructure.

3. Review Security Scan Results Regularly - Run security-scan periodically to catch configuration drift and newly introduced security issues. Cloud resource configurations can change through console access, API calls, or automated processes.

4. Use Cost Analysis to Justify Optimization - Before proposing infrastructure changes, use cost-analysis to quantify the financial impact. Data-driven cost optimization proposals are more likely to gain approval from stakeholders.

5. Store Credentials in Environment Variables - Keep FIREFLY_ACCESS_KEY and FIREFLY_SECRET_KEY in environment variables, never in configuration files. These credentials provide broad access to your cloud infrastructure data.

Common Issues

1. Discovery Returns Partial Results - The Firefly account may not have all cloud accounts connected. Verify in the Firefly dashboard that all relevant AWS, Azure, and GCP accounts are integrated and their scan schedules are active.

2. Authentication Fails with Invalid Credentials - Access key and secret key must be used together and match. Regenerate both credentials from the Firefly platform settings if authentication fails repeatedly.

3. Cost Data Not Available for Some Resources - Cost attribution requires cloud billing integration. Resources in accounts without billing data connected to Firefly will show zero or missing cost information.