Live data from Hacker News

Show HN: Browser Harness – Gives LLM freedom to complete any browser task

github.com

71–77 of 77 posts

Re: Show HN: Browser Harness – Gives LLM freedom to complete any browser task

#72
"Minimal infrastructure, max LLM freedom" works great for personal automation. The same shape under enterprise security review collapses on the question of what cannot happen, which is exactly what the prompts-and-vibes school doesn't have a structural answer for. Direct CDP hands the model the keys; the harness around it is what should decide which doors the keys open.

Most agent stacks at AI startups have that layer as llm driven glue rather than an owned surface, and it shows up as a re-architecture cost on every model release. model should be replaceable, the integrations and guardrails specific to the customer's environment should not.

Re: Show HN: Browser Harness – Gives LLM freedom to complete any browser task

#73
like how thin this harness is. Letting the LLM work directly against CDP with just helpers.py + SKILL.md feels much closer to how I want “skills” to look: small, inspectable, and self-extending when the agent actually needs something (like that upload_file moment). Curious to see how far this pattern can go beyond the browser.

Re: Show HN: Browser Harness – Gives LLM freedom to complete any browser task

#74

Remote debugging will trigger bot detection. So this won't work for use cases like booking a flight/hotel on the major platforms.

i dont think its detectable if done well. We’ve built involved automations with no problems. i used libretto not browserbase tho

I have too, although I wont describe my methods here. The more people do it, the more bot detection will improve.
Post reply on HN