Errors: what did not work
Show everything an AI tried in this person's Trillion that did not work, with why, which AI and model did it, and the chat it works in
A screen: the AI activity list narrowed to what failed (a call that failed, or a tool run that ended failed). An AI reads the same with read_activity failed true.
What it asks for
Nothing: your AI starts it with your words.
Use it from your AI
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to show everything an AI tried in this person's Trillion that did not work, with why, which AI and model did it, and the chat it works in”
What your AI calls: run_tool {"tool_id":"tool:errors"}