Skip to content

bench: visual verification of Studio comparison matrix UI #990

@christso

Description

@christso

Objective

Visually verify the Studio comparison matrix UI (PR #984) renders correctly in a browser. The API has been verified but the React component has not been tested in an actual browser.

Context

PR #984 adds a "Compare" tab to the Studio with an experiment × target matrix. The backend API returns correct data (verified via curl), but the React component (CompareTab.tsx) has not been tested in a browser because the dev environment lacks a display server for headless Chrome.

What to verify

  1. Start studio: agentv studio --multi
  2. Register agentv-bench-skills project (has experiment data)
  3. Navigate to the project → "Compare" tab
  4. Verify:
    • Matrix renders with experiments as columns, targets as rows
    • Cells are color-coded: green (>80%), yellow (50-80%), red (<50%)
    • Best/worst performer indicators (▲/▼) appear
    • Clicking a cell expands to show per-test-case breakdown
    • Empty state renders correctly for projects with no experiments
    • Error state renders when API fails
    • Loading skeleton shows while data loads

Acceptance criteria

  • Screenshot evidence of the comparison matrix with real data
  • Expand/collapse works
  • Color coding is correct
  • Responsive on different viewport sizes

Metadata

Metadata

Assignees

No one assigned

    Labels

    in-progressClaimed by an agent — do not duplicate workwuiRelates to the browser dashboard / web UI runtime

    Type

    No type

    Projects

    Status

    Done

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions