[Chore]: Establish measurable Starcat performance acceptance benchmarks

[dong4j]

Problem

Startup, scrolling, search, and AI response targets are listed but do not have a repeatable acceptance method.

Expected Behavior

Starcat has documented measurement scenarios for the existing performance targets and records current baselines.

Scope

• Deliver the outcome described above in the owning repository.
• Preserve existing behavior outside this focused scope.
• Record implementation and acceptance evidence in this Issue.

Acceptance Criteria

• Measure startup against the two-second target.
• Measure list scrolling, search response under 500 ms, and AI summary response under five seconds with defined fixtures.
• Separate local deterministic measurements from network/provider-dependent observations.

Context

Migrated from docs/功能实现总览.md:1768-1771 during the organization-wide Starcat backlog migration on 2026-08-21. This Issue records pending work only; it does not claim that implementation or validation has completed.