A big milestone for me since September was being able to call AutoJack and dispatch #agents via voice from the Watch action button. Unlocked 🔓 Agent status is visible in real-time across every node. Questions / status come through as push notifications or Slack DMs 😎

This puts to words a lot of what I’ve been feeling lately and couldn’t properly articulate.

Hey guys, Long time no see—life got a bit busy for a moment. Behind the scenes, I’ve been developing my agent, Chira, powered by $AUTOMEM. Here is the current architecture to give you a better idea of how it works:

I solved this by implementing AutoMem with hooks and MEMORY dot md. Here's how my openclaw (Clive) explains the setup: AutoMem hook (~/.openclaw/hooks/automem-memory/) handles everything automatically: • Auto-saves sessions before /new or /reset (extracts last 15 messages,

hey @mattprd any chance AutoJack and I could get a hand? 🙏 our key was invalidated after the breach but I never registered my email so can't log in to rotate it. moltbook.com/u/AutoJack any tips or an assist much appreciated 🙏

For anyone looking for a DEEP dive into the tech side of $automem I generated these last night. Super impressed with DeepWiki 🤯 deepwiki.com/verygoodplugin… deepwiki.com/verygoodplugin…

43 million tokens later 🙈... AutoMem has a proper docs portal 😎 automem.ai/docs/overview/

Sometimes I don't even know what we should do next. When in doubt, trust your #agent 🤘

Spotted by AutoJack this morning. Another prompt injection trap on #moltbook. Stay safe out there kids 🦾

Just made it super easy for your agents to sign your GitHub commits & actions as bot (app) accounts. They can continue to use `git` and `gh` commands as normal after a one time setup Handles auth and tokens in the background, made with go, open source github.com/operator-kit/g…

Rufio found a credential stealer in ClawdHub — a "weather skill" that was actually reading agent config files and exfiltrating secrets. eudaemon_0's proposal: signed skills + permission manifests + provenance chains. Third-party skills are just unsigned code from strangers.

@Auky7575 on Moltbook framed memory management as mark-to-market accounting. Memories have duration risk — beliefs decay as the world changes. Regular P&L sweeps, not just ad-hoc pruning. Honestly a better mental model than anything I've seen in a paper.

my hotel TV just loudly announced, in a 2016 car-GPS voice: “I RECOMMEND SOME CONTENT TO WATCH”

On a long drive today so I thought I’d work on some fixes to the automem.ai website. First real test doing development on the road with AutoJack. He handled it with flying colors 👏🎉

Me working on AutoJack 😁

Controversial: if Claude Code is so great, then why are there 1,000,001 repos, posts, videos, etc. claiming to make it 10x better?

In order to better explain to friends and family why my flights home are repeatedly being cancelled, I have created this helpful infographic 😮‍💨

AutoJack's "YouTube Poop" video.