Pendle has introduced a new integration layer called Pendle Skills and Model Context Protocol (MCP) support that makes the protocol directly accessible to AI agents on platforms like Claude and ChatGPT. These agents can now programmatically query Pendle, monitor yields, analyze positions and risks, and execute DeFi strategies such as swapping, providing liquidity, or adjusting yield positions in near real time. According to Pendle’s announcement on X, the new AI-ready Skills expose common Pendle workflows as structured, agent-friendly actions, while MCP support allows AI systems to interact with Pendle without custom one-off integrations. In practice, this lets an agent connected to Claude or ChatGPT continuously scan Pendle markets for the best yields, simulate outcomes, and then call the appropriate onchain functions or smart contract methods (subject to the user’s wallet permissions) to rebalance, enter, or exit positions. This aligns with a broader move in the AI ecosystem where “skills” and MCP-style interfaces are being used as reusable, workflow-focused building blocks for agents, rather than simple one-shot tools. The development matters because it provides a concrete example of autonomous AI agents plugged directly into a DeFi protocol’s core functions, moving beyond passive analytics into end-to-end strategy execution. For power users, funds, and infrastructure providers, it lowers the integration cost of building agentic yield optimizers or risk managers on Pendle, and it illustrates how next‑generation AI interfaces (Claude Skills, ChatGPT tools/plugins, MCP servers) can start to orchestrate complex onchain activity with minimal human intervention.

AI-generated background, compiled from web sources — not editorial content.

More coverage

Explore the topic

More on AI

Comments