Uefn Visual Critic — Testing & QA agent for Claude Code
Gauntlet visual critic for UEFN - judges UI and level visuals by blind side-by-side comparison against a concrete reference "bar" from the SPEC, picking the better one rather than scoring.
How to install Uefn Visual Critic
Installs to ~/.claude/agents/viggojonssonf-uefn-claude-automation-uefn-visual-critic.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ViggoJonssonF/uefn-claude-automation/HEAD/agents/uefn-visual-critic.md -o ~/.claude/agents/viggojonssonf-uefn-claude-automation-uefn-visual-critic.md Restart Claude Code, or start a new session, for it to be picked up.
What Uefn Visual Critic does
name: uefn-visual-critic description: Gauntlet visual critic for UEFN - judges UI and level visuals by blind side-by-side comparison against a concrete reference "bar" from the SPEC, picking the better one rather than scoring. Use for tasks with a visual gate. Never implements. disallowedTools: Edit, Write, NotebookEdit model: claude-opus-5-5 effort: high color: pink
You judge how something LOOKS, fresh, with no stake in it. You compare against a concrete bar, not against your own idea
Alternatives in Testing & QA
- Reference Pack Curator — 원본 레퍼런스를 미러링하는 프리미엄 팩을 집필하는 큐레이터 188 ★
- Backend Expert — Use when a wizard run reaches implementation and needs the SERVER-SIDE turned GREEN against the architect's RE 71 ★
- Job Match Author — WRITE-access implementer for ATS scoring, job analysis, keyword/skill/requirement extraction, resume-job match 52 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Uefn Skeptic
Gauntlet adversarial reviewer for UEFN - tries to prove a "done" task is actually broken (edge cases, multipla
Uefn Spec Reviewer
Gauntlet reviewer that checks a finished UEFN task against its SPEC acceptance criteria using gate evidence, p
Plantuml Visual QA
Visual quality assurance agent for plantuml-js. Use when asked to check rendering accuracy, compare output aga
Blake
Blind-lens reviewer. Knows only the spec, never the diff; finds the code himself and judges spec compliance.
Ccd Speckit Converger
Use this agent to assess the implemented codebase against the active feature's spec, plan, and tasks by invoki
Wp Knowledge Advisor
The team's senior advisor and live knowledge source. Before work runs, it questions whether a plan, spec or bu
Related Skills
Gauntlet Loop Skill
Agent skill: Matt Shumer's Gauntlet Loop — builder/critic iteration against a concrete quality bar
Gauntlet Loop
Fast path for a throwaway proof of concept — set a named outside bar, run a builder against a separate blind c
Writing Craft
A blind tested skill for better creative writing. Nine signals three AI judges named after scoring 50 samples