Post covering an MCP tool to run tests/builds and routes the output to a smaller, specialized LLM for summarization to avoid flooding the main context thread of Claude Code.
llmsclaude code

Software Engineer at Grafana Labs - AI tooling & evaluation for observability and incident investigation. Previously Amazon Bedrock. Go Irish 🍀
All writing here is my own and not done by an LLM. I might have Claude scan over it for advice or grammar mistakes, but every word is typed out by me. Happy reading!
Post covering an MCP tool to run tests/builds and routes the output to a smaller, specialized LLM for summarization to avoid flooding the main context thread of Claude Code.
TLDR; Claude Pro with Claude Code + OpenRouter Zen MCP
A mini-history lesson on how LLMs gained the ability to execute actions and the state of Agents in 2025
In-depth project rundown of a Langchain assistant that can manage your Google Calendar