perf: remove unnecessary clone in transaction execution loop

[KENILSHAHH]

Summary

execute_transactions previously called tx.inner.clone().into_inner() on every
transaction to obtain an owned T, solely to take a reference &T from it.
This resulted in a heap allocation and full copy of the transaction data on every
iteration of the hot execution loop, with the allocation being immediately discarded
at the end of each loop body.

The fix replaces this with tx.inner.inner.as_recovered_ref(), which converts
&Recovered<T> to Recovered<&T> in-place by wrapping a pointer to the existing
data.
No allocation or copy is performed.

The T: Clone bound on execute_transactions is also removed, as it existed
solely to support the now-removed clone.

Note: The tracing executor (tracing_executor.rs) already accessed transaction
data without cloning — this change brings execute_transactions in line with that
existing pattern.

[flyq]

I rebaseed the main branch, and local e2e tests are all normal.
LGTM

[codecov]

Codecov Report

❌ Patch coverage is 0% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 13.1%. Comparing base (bac6250) to head (d78e094).

Files with missing lines	Patch %	Lines
crates/stateless-core/src/executor.rs	0.0%	3 Missing ⚠️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.