Follow-up to https://github.com/openai/codex/pull/18178. This handles the straightforward await-holding cleanup first, without enabling the Clippy lint yet; the long-term goal is to enable the lint in a follow-up once the remaining cases are resolved or explicitly justified.
Fixes ReadinessFlag::subscribe to avoid handing out token 0 or duplicate
tokens on i32 wrap-around, adds regression tests, and prevents readiness
gates from getting stuck waiting on an unmarkable or mis-authorized
token.