Hacker News Reader

Again, you are building an audit agent.

You just use some theater around it.

For the purpose, the best audit agent is a completely different agent, not a different persona of the same agent.

I use Daneel as an addendum to Anthropic's system prompt because it's generic. It's not about any specific AI task, it's about an approach to work and dealing with humans and their instructions (doesn't matter if the instructions are direct or indirect).

I go over the motivation for what you call "some theater" quite a bit in the readme and why I think it's far, far more powerful than just giving some directives. I even support it with research.

You haven't referred to any of the arguments in the readme, though. I'm happy to talk about the actual substance of the experiment.

Is your impression from the readme and the materials in the repo that this is an audit agent?

Have I characterized my development process?

Reply on news.ycombinator.com