Y
Hacker News
new
|
ask
|
show
|
jobs
by
clarus
97 days ago
The solution for this might be to add a ME.md in addition to AGENT.md so that it can learn and write down our character, to know if a question is implicitly a command for example.