setup-mcp-agent-analytics Claude Code Plugin
Claude Code PluginClaude CodeDevelopmentMonitoring & ObservabilityAI Agents & AI App DevelopmentThe explanation below is AI-generated. Please verify it against the sources.
This plugin provides a single skill, setup-mcp-agent-analytics, that scans an MCP server's codebase to identify its implementation language and then adds the corresponding Pendo SDK call needed for agent analytics tracking. According to the README, it recognizes Python, TypeScript, Go, and (in one table entry) Java projects, wiring in language-specific instrumentation such as PendoMCPServer, initMcp(), a Go SDK instrument call, or a wrapAll() method depending on the detected stack. Once instrumented, calls made to the MCP server's tools are reported into Pendo's agent analytics so usage can be observed within Pendo. The skill is invoked with a dedicated slash command inside Claude Code. It is distributed as part of a larger marketplace of Pendo-related plugins alongside analytics, guide-creation, and journey-building tools.
About the service
Pendo is the product analytics and engagement platform this plugin connects to. Instrumenting an MCP server routes its tool-call activity to Pendo's agent analytics feature, and using Pendo this way requires authenticating a Pendo MCP connector inside Claude Code, as described in the README.
What you can do with setup-mcp-agent-analytics
- Run the setup-mcp-agent-analytics skill to scan an MCP server project and detect its programming language
- Automatically insert the matching Pendo instrumentation call for Python, TypeScript, Go, or Java MCP servers
- Send MCP tool-call activity into Pendo's agent analytics once instrumentation is added
- Authenticate the Pendo MCP connector via the /mcp command before analytics data can flow
- Use the skill alongside other plugins in the same marketplace, such as those for Pendo analytics reporting or in-app guide creation
Sources
History of setup-mcp-agent-analytics
- Claude Code Plugin Added setup-mcp-agent-analytics