القائمة
Drive the Boardex UI against the mock runner in a headless browser to verify a change end-to-end.
يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.
اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
يتجاوز الأمر المباشر Prompt المخصّص للمراجعة. افحص المصدر قبل تشغيله.
npx skills add https://github.com/ankayca/boardex --skill verifyيبقى الأمر في سطر واحد. مرّر أفقيًا لمراجعته كاملًا قبل النسخ.
تفضّل نسخة محلية؟ نزّل الملفات المتاحة حاليًا لدى SkillsMP.
عرض SKILL.md
استنادا إلى تصنيف SOC المهني
| name | verify |
| description | Drive the Boardex UI against the mock runner in a headless browser to verify a change end-to-end. |
Ports 4319 (mock runner default), 4321, 5173, 5199 are often occupied by leftover
dev processes — pick fresh ports and check with ss -tln first. Never
pkill -f "tsx src/index.ts" to clean up: the pattern matches every mock runner
on the machine, including ones you don't own. Kill by PID.
# mock runner (SPEED scales fixture delays; acceptance runs use SPEED=5)
cd tools/mock-runner && SPEED=5 PORT=4333 npx tsx src/index.ts
# UI, pointed at it
cd apps/ui && VITE_RUNNER_URL=http://localhost:4333 npx vite --port 5333 --strictPort
# health checks
curl -s http://localhost:4333/health # {"ok":true,...,"runnerKind":"mock"}
No browser tooling in the repo. Install Playwright in the session scratchpad
(NOT the repo): npm init -y && npm install playwright && npx playwright install chromium (skip --with-deps; sudo is unavailable and the WSL box already has
the libs). Viewport ≥1280px wide for the three-zone workspace grid; 1100px to
check the right rail stacking under center.
Useful handles (accessible roles, stable):
header span[data-kind="status"],
complementary "Board context", step rows are <li> with an expand button named
by step title, log tabs role=tab (Agent/Build/Flash/Serial/RTT), panes
role=log named " — log", lines under [data-index] (virtualized
— only ~30 rows render; short logs render fully).GET /runs/{id}/events?afterSeq=0, find unresolved approval.requested, then
POST /runs/{id}/approvals/{aid} {"status":"approved"}.The full fixture at SPEED=5 completes in ~80s with prompt approvals. Collect
page.on('console') errors — the §7.2/§7.3 done-criteria include "no console
errors". Mid-run reload is the standard replay check: completed steps' log
content must be byte-identical before/after.