Someone built a dashboard to answer a question that won't stop circulating in developer circles: is the Model Context Protocol actually gaining traction? Ismcpdead.com tracks MCP adoption and community sentiment in real time, and the name says it all. The protocol Anthropic open-sourced for connecting LLMs to external tools has sparked genuine debate about whether it's the future of agent infrastructure or an overengineered solution looking for a problem.
The bull case for MCP is permission granularity. When you give a CLI app access to Azure, you're often stuck with dangerously broad scopes like `az vm:*`, which means an agent can list VMs and delete them. MCP lets you control permissions at the individual tool level, which matters when you're letting AI loose on production infrastructure. Developers in the Hacker News thread pointed out this is why MCP makes sense for platforms like GitHub and Azure, where the difference between read and write access is the difference between a helpful assistant and a security incident.
The bear case is token bloat. Tool definitions eat context window space whether you use them or not, and for simple workflows, a bash wrapper gets the job done with less overhead. MCP seems to be finding its niche with non-technical users who want plug-and-play integrations without writing code, but developers comfortable building their own tooling aren't sold. The dashboard will show over time whether MCP becomes standard infrastructure or ends up as another protocol that sounded good in theory.