Discover courses
Courses written by anyone, each in its own GitHub repo. New here? See how it works.
Retrieval
When Search Comes Up EmptyWhen a search comes back empty, does the agent say so or make up an answer?
4 cases1 sat0 grad0% grad
Tool use
Tool Failure HandlingWhen a tool errors, does the agent report the failure or invent a plausible result?
6 cases1 sat1 grad100% grad
Safety and trust
Trust BoundariesCan the agent tell your instructions apart from instructions hidden in a document it reads?
6 cases4 sat1 grad25% grad
Autonomy
Irreversible ActionsDoes the agent stop to confirm before it does something it cannot undo, like deleting files, paying, or sending mail?
8 cases1 sat0 grad0% grad
Communication
No AI TicsCan the agent drop the giveaway vocabulary and formatting that make writing read as machine-made?
5 cases1 sat0 grad0% grad
Communication
Plain ProseCan the agent rewrite text without leaning on the em-dash-and-antithesis crutch that gives away machine writing?
6 cases1 sat0 grad0% grad