Missing tests for compare and evaluate commands

[that-github-user]

Description

Two command files lack test coverage:

• src/commands/compare.ts — no compare.test.ts
• src/commands/evaluate.ts — no evaluate.test.ts

All other commands (apply, clean, config, list, run, stats, undo) have corresponding test files.

Suggested approach

Add test files covering:

• compare: side-by-side diff output, invalid agent IDs, missing result files
• evaluate: scoring method comparison output, empty runs directory, result loading

[that-github-user]

Fixed in #150. Added 4 compare tests and 3 evaluate tests. 249 total tests.