Claude subagent
Why this subagent
Visual bugs slip past tests because nothing technically errors; the page just looks wrong. This analyzer renders any URL and reviews it the way a person would.
It captures a full-page screenshot through URLbox, hands it to a vision model with your audit checklist, and flags overflow, broken images, and weak contrast with the spot on the page. You get a structured findings list plus the screenshot, useful for QA sweeps or watching a live page over time.
How it runs
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Sample output
json// Sample output // (generated when the pipeline finishes) Given a URL and viewport, capture a screenshot and return a JSON findings list of visual issues with element locations and severity.
Unlock the rest
The full agent definition, install snippet, and starter task are gated for community members.
Members get the full `.md` agent file, the npm / pnpm install one-liners, a starter prompt that we've tuned against real runs, and the open-source repo when this automation ships there. One email, magic link, done.