Release surface: legacy status and package-channel cleanup

[yunwei37]

Goal

Clarify GPTtrace package/channel status so old package names do not create confusion.

Current known surface

• GitHub repo exists.
• PyPI gpttrace appears to be an older official package.

TODO

• Decide whether GPTtrace is active, legacy, or folded into AgentSight.
• If active: create updated GitHub release and publish real packages only where the current implementation works.
• If legacy: mark README/package metadata as legacy and point users to the recommended successor.
• Avoid publishing new npm/crates names unless there is a current product/API.

Rule

Do not split brand attention with inactive package names; either revive with a real release or clearly deprecate/migrate.