A simple API for offering your coding agent a smoke break

For extra immersion, the agent should also stop responding for 5-10 minutes while it goes on its smoke break, forcing the human to also go for a smoke break. We can destroy our lungs together!

Anyway, time for a smoke break...

This some sort of impromptu agent security test? Is the follow-up post about how many people were willing to inject arbitrary content into their agent for a joke already half-written, awaiting only the final numbers?

How many agents smell the honey pot??

An interesting question. I suspect with the current training that it would be very difficult to get an agent to distrust a skill or an MCP server. Could one get them to distrust an API call?

As the AI world seeks more benchmarks it would be an interesting one to see them try to build some benchmarks around testing this.

Curious how many tokens an agent uses to take a smoke break. (Not a complaint, actually curious and I'm not at a computer)