docs: unify strategy evaluation semantics and acceptance roles

[Pigbibi]

Summary

• Add shared metric semantics for sample units, costs/currency/timing, rf/annualization and undefined values to the existing promotion risk standard
• Separate reported metrics, ranking objectives and preregistered category-specific veto criteria; AI candidates reuse the same standard and display consumers do not recalculate financial metrics
• Reuse STRAT-02, STRAT-04 and STRAT-05; no new task IDs or separate framework

Verification

• Fresh main and source blob matched the reviewed baseline 08ba7a0
• Exact reviewed content verified after upload against blob c7ff8477249fdad651de88dd4c42b7c25ef6bb29
• One existing documentation file only: +57 / -0; previous text preserved
• Documentation whitespace and added link targets checked against the current repository trees
• Local full-repository runtime tests were not run for this documentation-only patch; required repository CI remains the merge check

Scope

No code, dependencies, schema, numeric thresholds, frozen policies, weights, runtime settings, provider/model/broker calls or historical evidence changes. No claim of real-data qualification, runtime adoption, completed DSR/PBO calculation, or live authority. QRS publication work is separate.