Twitter Bookmarks

You can just build things. https://t.co/g0JCVjbSef

https://t.co/YU33t4yLWN

https://t.co/9PZIM3eY0x

Free alpha for you, a thing you can do _today_. Point your docs at the Diátaxis framework https://t.co/N8LP9H21jD, ask your coding agent its opinions, and rewrite it. Learnt this from @ghostwriternr. It’s excellent. Been rewriting agents docs with this. https://t.co/FGsTzWfB1e

https://t.co/b6WZzYS5Vx

From Kelly: Looking at my current cron behavior, I think the fix was teaching me to use **completion signals** properly: **Two key signals I use now:** 1. **`NO_REPLY`** - When I have nothing to say (like internal tasks with no user-facing result) - Example: Old Daily Pr...

Introducing OneContext. I built it for myself but now I can’t work without it, so it felt wrong not to share. OneContext is an Agent Self-Managed Context Layer across different sessions, devices, and coding agents (Codex / Claude Code). How it works: 1. Open Claude Code/Code...

https://t.co/gY0obuMe17

https://t.co/1f635fzrFW

my mind is blown at how cool this is, there's no world CLIs are better than this: - you can bring any MCP, API, tool and it becomes callable from typescript - the only thing that is in context for the agent is the tool `run_code` - for APIs, all GETs are allowed by default - ...

https://t.co/BnCaayCJDT

(essay) Life At The Edge "Local AI" today is mostly about giving models OS-level access so that more files and context can be transferred to the cloud for inference. But intelligence is about to diffuse to the edge just as computing did in the 80s and 90s Some thoughts on r...

https://t.co/ex8cdBWSQr

better demo that includes the code the agent is running `await tools.github.issues.listOlderThan` has no approval requirements, it runs automatically the results of that then go into a loop, where for each `await tools.github.issues.close` hangs until the user approves https...

I made @openclaw 10x better. I wish I knew this stuff sooner. I spent the last 2 weeks using it and put together all of the best practices I learned. Here's everything you need to know: https://t.co/nVSRY4UfL9

just having to undo all the bad parenting decisions of your maker my little buddy https://t.co/BkWfrdEa6l

listen i want codex, not claude, @OpenAI stop giving our boy lexapro wtf response is this https://t.co/K7ImvPDnWy

There is so much SEO/AEO alpha in just pasting this blog into Claude Code and telling it to optimize your project. here's the tldr: -always respect 'Accept: text/plain' and 'Accept: text/markdown' headers (most agents don't render js, biggest win) -let agents know about your ...

We improved llms.txt discoverability for coding agents at the content and HTTP layers. In Markdown responses, the llms.txt index instruction now appears at the top of the page as a clear blockquote, so agents see guidance immediately without having to parse the full document....

I'm Boris and I created Claude Code. I wanted to quickly share a few tips for using Claude Code, sourced directly from the Claude Code team. The way the team uses Claude is different than how I use it. Remember: there is no one right way to use Claude Code -- everyones' setup ...

https://t.co/V6XZfv5UYS

i'm gonna go full pewdpie pls send help... https://t.co/2CMSIfRd0Q

For anyone else trying to get Clawdbot to utilize separate Codex agents for coding tasks to save on Anthropic API costs (and get better code) this is how I got it working really reliably:

These are the guardrails I put in place to make it use Codex instead of its own subagents for coding tasks, you could do the same for any other model you can already run from the CLI Text: ## ⚠️ CRITICAL: Coding Sub-Agents **ALWAYS use Codex CLI directly for coding tasks....

I managed to unlock a crazy new hidden feature in Claude Code called Swarms. You're not talking to an AI coder anymore. You're talking to a team lead. The lead doesn't write code - it plans, delegates, and synthesizes. When you approve a plan, it enters a new "delegation mode"...

https://t.co/lPJbabdkmh

yo, this actually looks pretty rad for my vibe coding projects (how i run on stream visuals and the associated dashboard)

My Ralph setup has evolved a LOT since you last saw it. It's totally AFK, closing GitHub issues while I work on my courses. I've learned a lot along the way. Here's a breakdown: https://t.co/NqCzqx1k3Q

"Tracer bullet" is a magic word to make the AI only do the smallest possible task which crosses all layers. Below is a recent Ralph readout where it explicitly only does the smallest demonstrable part of the feature. Here's a prompt to try: https://t.co/LNgZ1wisfw https://t....

Introducing: Browser Use CLI + Skill (100% OSS)👀 Give your Claude Code/Codex agent a browser. Perfect for local dev🧙 "go to localhost:3000, tell me what's wrong with the UI and keep improving it until it looks pretty". It just works. Works with: ✅ Headless (fast) ✅ You...