Llm

LangGraph のスレッドを永続化する checkpoint と HumanInTheLoopMiddleware によるユーザーへの確認の挙動を確認する

Installing the herdr Agent Multiplexer and Using the hunk Plugin to Inspect Diffs and Send Fix Instructions

Splitting Tasks Fed into the Hermes Agent Kanban Across Multiple Agents

Running the A2A (Agent2Agent) helloworld Agent and Calling JSON-RPC with curl to Inspect the Interface

Working Efficiently and Reproducibly with Claude Code's Memory, Subagent, and Skills

Blocking dangerous operations with Claude Code Hooks when using --dangerously-skip-permissions

Trying out bubblewrap used in Claude Code's Sandbox Runtime and exploring its network restriction mechanism

Running browser operations by Claude Code & Playwright MCP in parallel using tmux and working it on each branch with git worktree

MCP で Todoist にタスクを登録する

Handle Actions that take too long or are too heavy for Agents for Bedrock by using RETURN_CONTROL and processing them in a Go client

Create Agents for Bedrock using CDK and check that Lambda functions are called based on input

Generate SQL from natural language and access databases with AgentExecutor of Langchain's SQL Database Toolkits

Create an LLM agent from a graph containing a cycle using the LangChain ecosystem LangGraph

Fine-tuning OpenAI's GPT with Japanese Prime Minister's speech in the Diet