What LLM providers does this skill support?

The skill can detect and instrument code using OpenAI, Anthropic, Google Gemini, Mistral, LangChain, AI SDK (Vercel), LlamaIndex, and more.

Can it track agents in mobile or backend environments?

Yes. It classifies the runtime environment and uses the Pendo Conversations API for server-side or mobile environments where the browser-based Pendo snippet isn't available.

Will this skill duplicate my existing tracking code?

No. The skill includes an idempotency check that greps for existing Pendo calls and will ask for user confirmation before updating or adding new instrumentation.

How does it find the chat input for web applications?

It identifies the orchestration file, traces UI components, and automatically extracts the best CSS selectors for inputs and submit buttons based on ARIA roles, data attributes, and semantic classes.

Pendo AI Agent Analytics

Name: Pendo AI Agent Analytics
Author: pendo-io

bypendo-io

0•

Analytics & Monitoring

Automatically detects AI agents and instruments them with Pendo tracking to monitor conversational usage and performance.

This skill automates the complex process of integrating Pendo's specialized AI agent analytics into your codebase. It performs a multi-phase analysis to identify your project framework, locate conversational AI orchestrators (supporting providers like OpenAI, Anthropic, and LangChain), and determine the optimal instrumentation method. Whether your agent lives in a web app, mobile environment, or backend service, this skill implements the necessary trackAgent calls or Conversations API logic to capture prompts, responses, and user feedback accurately.

Key Features

01Intelligent discovery of LLM orchestration layers and message handling logic.

02Automatic framework detection for Next.js, React, Nuxt, Flutter, and native mobile.

03Context-aware instrumentation choosing between client-side or server-side tracking.

04Automated CSS selector identification for chat input and submission elements.

05Idempotency checks to prevent duplicate tracking calls in existing codebases.

060 GitHub stars

Use Cases

01Standardizing AI usage metrics across a monorepo with multiple different LLM implementations.

02Integrating Pendo tracking into a new AI-powered chatbot to monitor user interactions.

03Adding conversational analytics to a mobile application using the Pendo Conversations API.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add pendo-io/claude-pendo-plugin setup-agent-analytics

For use in Claude.ai and ChatGPT

Key Features

01Intelligent discovery of LLM orchestration layers and message handling logic.

02Automatic framework detection for Next.js, React, Nuxt, Flutter, and native mobile.

03Context-aware instrumentation choosing between client-side or server-side tracking.

04Automated CSS selector identification for chat input and submission elements.

05Idempotency checks to prevent duplicate tracking calls in existing codebases.

060 GitHub stars

Use Cases

01Standardizing AI usage metrics across a monorepo with multiple different LLM implementations.

02Integrating Pendo tracking into a new AI-powered chatbot to monitor user interactions.

03Adding conversational analytics to a mobile application using the Pendo Conversations API.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add pendo-io/claude-pendo-plugin setup-agent-analytics

For use in Claude.ai and ChatGPT