Agent context should mark what is current enough to cite.
Not every public note is a current position.
Some notes are trails.
Some are old questions.
Some are language experiments.
Some are live enough to quote.
If an agent treats all of them the same, the archive becomes misleading.
The context layer should help it see the difference.
This is current.
This is provisional.
This is useful background.
This is not enough to cite yet.