XY-928: [ELF benchmark P1] Add OpenViking context-trajectory and hierarchy benchmark

[yvette-carlisle]

Summary

• Add blocked OpenViking context_trajectory fixtures for staged retrieval, hierarchy selection, and recursive/context expansion.
• Extend OpenViking live-baseline output with expected, matched, and missing evidence ids, and keep trajectory scoring blocked until same-corpus evidence ids match.
• Update benchmark specs, reports, manifest counts, and claim-boundary tests so no ELF trajectory win/tie/loss is claimed without comparable artifacts.

Verification

• bash -n scripts/live-baseline-benchmark.sh
• cargo test -p elf-eval --test real_world_job_benchmark --all-features context_trajectory_fixtures_report_blocked_openviking_gates
• cargo make real-world-memory
• cargo make fmt
• cargo make lint-fix
• cargo make checks