What happened
Injective Labs shipped a Model Context Protocol server that maps the chain's core smart contract functions to MCP tool calls, CryptoBriefing reported Sunday. MCP, the open standard Anthropic published in late 2024 and OpenAI adopted in early 2025, is the plumbing that lets an AI agent call external tools with structured inputs and outputs. Plug the Injective endpoint into a Claude Desktop config or an OpenAI Agents SDK client and the model gets a menu of on-chain actions it can invoke, from reading balances to compiling and deploying CosmWasm contracts.
The prompt does the work a Rust developer used to do by hand. According to CryptoBriefing, the release covers contract deployment, token creation and basic DeFi actions on the exchange module, with more coverage promised in follow-up drops.
Why it matters
This is the first time a major layer-1 has shipped a first-party MCP endpoint, and it collapses the distance between an idea and an on-chain artifact from days to minutes. A trader who wants a bespoke perp market, a token launcher who wants a vesting contract, a market maker who wants a custom order router, all of them can now describe the thing in English and have an agent produce the deployed address.
That is a real change in who can ship code. It's also a real change in the attack surface. An agent with signing authority and a prompt window is a new class of key-holder, and the security literature on prompt injection against tool-using models is still thin.