Testing: Benchmark tests and E2E integration test suite

[erdemtuna]

Source: codeFindings.md Section 4

Two testing gaps:

1. No benchmark tests: Performance-critical paths lack benchmarks — semver comparison, SHA-256 hashing, YAML parsing, tree rendering. Important for detecting regressions as the codebase grows.

2. No E2E integration tests: No tests exercise the full CLI workflow (init → add → install → update → remove). Unit tests cover individual functions well, but the end-to-end contract is untested.