15 entries
I've fallen behind on blog things. In the past few weeks, I've been experimenting with claude-agent-sdk, equipping the agent with MCP and prompting to use a CLI measure trade-offs and ergonomics using humanlayer's RPI (research, plan, implement) commands on large codebases to implement complex...
I built a Claude Code slash command that given a prompt, analyzes the ongoing chat and captures to insight or learning from the conversations. This seems like it's going to be useful for capturing and searching a domain-specific knowledge base using queries that come to mind when I remember I did...
I've been experimenting using Claude Code to learn Rust.
I tried out Claude Code's new planning mode recently and ran a demo with it at work today. I stumbled upon it last week.
I downgraded Claude Code to standard usage a little while ago and I have to say it's one of the best things I've done recently. Not to say that I think Claude Code is excellent and effective - it is - but when I (occasionally) max out my Claude Code quota, I'm motivated to try a different agent...
Currently watching this video a bit at a time.
<kbd>ctrl</kbd> + <kbd>v</kbd> on macOS pastes and image from the clipboard into Claude Code. Amazing to finally know how to do this.
Cute detail from the Claude Code update today
In a pinch, took a shot and found out claude --resume is a real thing
Claude Code supports custom slash commands. Custom slash commands allow you to define frequently-used prompts as Markdown files that Claude Code can execute. Commands are organized by scope (project-specific or personal) and support namespacing through directory structures.
A sad morning for Claude Code. Amp is mostly working though.
Played a bit more with Amp today. It's been a month or maybe more since I last used it.
I made another attempt at using Claude Code to build an LLM-based DJ app, integrated with Spotify. I figure this time, I would use a TUI (like Claude Code and Gemini CLI use) because I hoped it would be more testable. Agent coding tools seem to get much further when they can write tests for...
I've been trying to find a way to test Claude Code with Kimi K2 running on Groq. In my head, this could be a super fast, capable agent that would make trying larger tasks less time consuming to wait for.
Today, Anthropic entered the LLM code tools party with Claude Code.