Built auto-approve for Claude Code, instead of --dangerously-skip-permissions.
Checks local rules, lets you add patterns, sends the rest to Haiku to approve or leave for you. Caches decisions.
~80% fewer prompts. It's good.
https://t.co/YIDOuhmOus
@Shpigford@initialcommitco I think there's also something to be said for some cold outreach to potential buyers who for each site. For example, Clear Terms would probably have been valuable to a legal practice.
@Shpigford I use it to build an AI focus group of experts + ICP personas that I run feature ideas past. I've found that pretty handy. Agent deploys subagents and then they all interact via the agent.
Agreed. They aren't stopping people from using it just limiting access in third party apps to the API. It's how most platforms work. I was surprised a subscription worked in a third party tool already.
If OpenCode is so good, use the Claude API or another model.
An absolutely right decision. Claude models were fine-tuned with reinforcement learning to work with a specific interface. This is why Claude is so effective in coding.
When you withdraw the model that was finetuned to work in a certain environment and put it into a different one, the consequences are absolutely unpredictable. Furthermore, the user experience is arbitrarily worse compared to the UX with the environment for which the model was finetuned.
I'm 100% certain that people who still claim that Gemini is better for coding than Claude used Claude in a non-native environment.
Without signing a replacement of similar calibre to BK (top 10 in world at least), this likely signals the beginning of the end for LIV as a competitor the PGAT.
They'll become another feeder/second rung tour.
As the big names contracts expire they'll migrate away.