Re-curate *K. grimontii* reference set + re-run engine (drop misclassified K. pasteurii)

[happykhan]

The Klebsiella oxytoca complex group (2026-05 expert feedback survey) found that the K. grimontii reference dataset includes ~45 K. pasteurii genomes misclassified as K. grimontii in the high-quality set, plus ~5 more in filtered_out. K. pasteurii is closely related taxonomically so the misclassification is subtle, but it pulls the K. grimontii threshold distribution.

Acceptance:

• Get the misclassified-accession list from the KO complex group.
• Remove those genomes from the K. grimontii input set.
• Re-run qualibact-engine; import via scripts/import_engine_output.py.
• Remove the "K. pasteurii misclassification" entry from content/species-notes.yml once landed.
• Spot-check: K. grimontii GC / genome-size distributions should tighten visibly relative to the pre-cleanup engine output.

Survey provenance: see OUTREACH-v1.1.md for the KO complex group's contact info.