From a5d3709fdb73b2f1973b8a7ba602c2ad099814e7 Mon Sep 17 00:00:00 2001 From: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com> Date: Fri, 20 Nov 2026 21:13:43 -0800 Subject: [PATCH] docs(skills): document running mobile:audit CLI in test-t3-mobile Co-authored-by: Cursor --- .agents/skills/test-t3-mobile/SKILL.md | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/.agents/skills/test-t3-mobile/SKILL.md b/.agents/skills/test-t3-mobile/SKILL.md index a11eee4e3f4e..ddfb45fa39c1 100644 --- a/.agents/skills/test-t3-mobile/SKILL.md +++ b/.agents/skills/test-t3-mobile/SKILL.md @@ -82,3 +82,27 @@ Confirm the intended projects appear, exercise the affected flow, and capture evidence. Retain the app and environment while iterating. At teardown, remove the disposable connection, close the AgentDevice session, call `device_close`, and stop only your backend and Metro processes. + +## Accessibility audit + +For mobile UI changes, run the local `mobile:audit` CLI on the touched screens +with the session's AgentDevice launcher: + +```bash +vp run mobile:audit capture --platform --screen \ + --out /tmp/mobile-audit/after --agent-device "$agent_device_command" \ + -- "${agent_device_target_args[@]}" + +# Compare against a base capture (exits non-zero on regressions): +vp run mobile:audit compare --before /tmp/mobile-audit/before \ + --after /tmp/mobile-audit/after --out /tmp/mobile-audit/diff \ + --fail-on-regression + +# Or annotate a single capture when there is no baseline: +vp run mobile:audit report --dir /tmp/mobile-audit/after \ + --out /tmp/mobile-audit/report +``` + +Attach the generated `report.md` summary and annotated PNG(s) to your response +to the user (embedded by absolute file path) and to the PR description (uploaded +via `ghdrop` or forge assets; never commit generated audit artifacts).