We are unable to complete this request because it was blocked under Anthropic’s Usage Policy. Copy Request (aea80a19-1d76-493e-aba6-4a820d10cc89)
Thanks for flagging this. That message is coming from Anthropic’s safety systems for the selected Claude model. Sometimes it’s not always that you were doing something bad but it could just be tripping a safety thing. Something we saw recently was someone building a game where where they used terminology like explode or destroy or something in the context of the game and the model didn’t like that.
If this was a normal coding task and you think it was a false positive, a few things that can help:
- Rephrase the request to make the benign coding intent explicit
- Remove unrelated pasted content that may be triggering the filter
- Split the task into smaller, more specific prompts
- Try a different model
If it keeps happening consistently and you believe it shouldn’t be, please tell us more about the prompt, and we can help you look at this more.