On Thu, Oct 15, 2026 at 04:35:27PM +0200, Thorsten Leemhuis wrote: > Can't we prevent all of that somehow -- maybe while also improving the > docs for humans? Is one of those frontier models maybe now or soon > capable of generating and maintaining dedicated skill files for less > capable LLMs from documentation written for humans? Or at least pointing > out inconsistencies when someone updates one side but not the other?
Every release, we ask, "Treat all of Documentation/ as the correct instructions, and examine all of agents/ to find inconsistencies, redundancies, or conflicts with Documentation/, and correct the agents/ docs. Then do a pass to distil instructions in agents/ down to only what is strictly needed for agents to operate." and commit the results? -- Kees Cook
