> The only reliable countermeasures are outside the LLMs but they restrain agent... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		Esophagus4 6 months ago \| parent \| context \| favorite \| on: Agentic Browser Security: Indirect Prompt Injectio... > The only reliable countermeasures are outside the LLMs but they restrain agent autonomy. Do those countermeasures mean human-in-the-loop approving actions manually like users can do with Claude Code, for example?

veganmosfet 6 months ago [–]

Yes, adding manual checkpoints between the LLM and the tools can help. But then users get UI fatigue and click 'allow always'.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact