mwiki
Re:Claude Code burns ~15–20% of your tokens on a 3-word status line
on Friday October 30, 2026 @01:57AM (#50300332)
https://news.ycombinator.com/item?id=50300097 - I had something similar. Subagent behaviors are concerning.
Re:A Claude Code agent's test escaped into my Mac. I only noticed beca...
on Friday October 30, 2026 @01:54AM (#50300321)
Yes. I spinned up 7 parallel agents, there was no visibility into what subagents are doing, which made it worse.
Re:A Claude Code agent's test escaped into my Mac. I only noticed beca...
on Friday October 30, 2026 @01:36AM (#50300223)
Yep. It just makes a good safety case for why to always use sandboxes and how serious the impacts are.
Steer a coding agent on your own machine from Claude/ChatGPT (via MCP)
Re:Steer a coding agent on your own machine from Claude/ChatGPT (via MCP)
on Saturday October 03, 2026 @02:55PM (#49946734)
I started working on this project to solve my own problem of not able to access my locally hosted agents remotely, It also came up with some infrastructure overhead.Now, This helps me in spawn agents via simple wake, able to work on my projects via any interface of my choice with little overhead. https://github.com/phyra-research/roamux
Re:
on Saturday October 03, 2026 @02:51PM (#49946706)
I started working on this project to solve my own problem of not able to access my locally hosted agents remotely, It also came up with some infrastructure overhead.Now, This helps me in spawn agents via simple wake, able to work on my projects via any interface of my choice with little overhead.
We Are the Bootloader: Transitioning to Intelligent Systems
Show HN: Run coding agents on your own machine, steer them from anywhere
Do LLMs hold the opinions they give you?
Re:OpenClaw Is a Security Nightmare Dressed Up as a Daydream
on Monday March 23, 2026 @03:49PM (#47494260)
https://github.com/pandyamarut/AgentBPF something adjacent to this.