My understanding (to the best of my knowledge) is that Claude blocked Open Claw... How does that affect all the instructions provided in the Open Claw module? This is where I am at "Prompt injection is real. Someone tested this: they sent an email from a random address containing hidden instructions. The AI deleted all their emails. Including the trash folder. This isn't hypothetical. It happened. But here's the good news: ClawdBot's creator specifically recommends Claude Opus 4.5 because Anthropic trained it to resist prompt injection. Their internal testing shows 99% resistance rate."

@Shemeer Ahmad Thanks! That reconciles the conflicting information I have seen/heard.