earthquake-plate-calculation-5: instruction, verifier, and reference solution use inconsistent target plates

[Ichiyo4]

Hi, I noticed a possible inconsistency in
earthquake-plate-calculation-5 at commit
a0da045a8bf64b8a8ff20730c4d6ef10dc4e2c5b.

1. instruction.md first asks for the plate with the highest number of
   earthquakes, but later defines distance_km as the distance to the
   Pacific plate boundary.

2. tests/test_outputs.py expects:

   • target_plate = "SU"
   • earthquake_count = 134
   • id = "us7000myfa"
   • distance_km = 419.32

   These values appear to follow the Sunda-plate interpretation, although
   the verifier comments refer to the Pacific plate.

3. solution/solve.sh explicitly selects the Pacific plate, filters
   earthquakes within it, selects PA boundaries, and computes distance
   to those Pacific boundaries.

Could you clarify whether the intended task is:

• to dynamically select the plate with the most earthquakes and calculate
  distance to that plate's boundary; or
• to analyze only the Pacific plate?

It may be helpful to align the instruction, expected verifier values,
verifier comments, and reference solution.

Thank you for releasing the benchmark.