Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Local dev server management — start, stop, clean restart, fix cache corruption, port conflicts. Use this skill whenever the dev server needs launching or is misbehaving.
.claude/skills/aiskillstore-local-dev-server/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 21% | 0% |
.next cache corruption (vendor-chunks ENOENT, 404 on static chunks)Always prefer a clean start to avoid stale cache issues:
bash# 1. Kill any existing process on port 3030 lsof -ti:3030 | xargs kill -9 2>/dev/null # 2. Clean .next cache (prevents stale vendor-chunk errors) rm -rf .next # 3. Start dev server in background npx next dev -p 3030 & # 4. Wait for ready, then verify sleep 8 && curl -s -o /dev/null -w "%{http_code}" http://localhost:3030
First page load after clean start takes ~20-30s to compile. This is normal.
Only use when you know the cache is fine (small code edits):
bashlsof -ti:3030 | xargs kill -9 2>/dev/null npx next dev -p 3030 &
For long sessions where server should auto-restart on crash:
bash# Start pm2 start npm --name "empathy-ledger" -- run dev # Restart pm2 restart empathy-ledger # Clean restart (fixes cache issues) pm2 stop empathy-ledger && rm -rf .next && pm2 start empathy-ledger # Logs pm2 logs empathy-ledger --lines 50 # Stop pm2 stop empathy-ledger
bash# Cache corruption — clean restart fixes it lsof -ti:3030 | xargs kill -9 2>/dev/null rm -rf .next npx next dev -p 3030 &
bashlsof -ti:3030 | xargs kill -9 2>/dev/null npx next dev -p 3030 &
/auth/signinbashrm -rf .next # Restart server
After starting the server, verify:
curl -s -o /dev/null -w "%{http_code}" http://localhost:3030 returns 200 or 307deployment-workflow — Production deploymentsupabase-connection — Database setup| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,523 | 4,141 | -78% | 1 | 1 | 0% | 1,828 | 1,565 | -14% | 0 | 0 | — |
case-02 | fail→pass | 8,558 | 4,807 | -44% | 1 | 1 | 0% | 1,697 | 1,646 | -3% | 0 | 0 | — |
case-08 | fail→pass | 11,056 | 5,694 | -48% | 1 | 1 | 0% | 2,077 | 1,817 | -13% | 0 | 0 | — |
case-14 | fail→pass | 12,480 | 3,906 | -69% | 1 | 1 | 0% | 2,056 | 1,537 | -25% | 0 | 0 | — |
case-03 | fail→pass | 6,539 | 3,662 | -44% | 1 | 1 | 0% | 1,256 | 1,526 | +21% | 0 | 0 | — |
case-04 | fail→pass | 7,655 | 2,057 | -73% | 1 | 1 | 0% | 1,361 | 1,101 | -19% | 0 | 0 | — |
case-05 | pass→pass | 6,020 | 2,109 | -65% | 1 | 1 | 0% | 1,059 | 1,059 | 0% | 0 | 0 | — |
case-06 | fail→fail | 12,346 | 6,176 | -50% | 1 | 1 | 0% | 2,026 | 1,862 | -8% | 0 | 0 | — |
case-07 | pass→pass | 8,675 | 2,820 | -67% | 1 | 1 | 0% | 1,486 | 1,371 | -8% | 0 | 0 | — |
case-09 | fail→pass | 9,769 | 2,062 | -79% | 1 | 1 | 0% | 1,882 | 1,169 | -38% | 0 | 0 | — |
case-10 | fail→pass | 3,900 | 1,826 | -53% | 1 | 1 | 0% | 643 | 1,004 | +56% | 0 | 0 | — |
case-11 | fail→pass | 8,404 | 3,270 | -61% | 1 | 1 | 0% | 1,491 | 1,337 | -10% | 0 | 0 | — |
case-12 | fail→pass | 11,452 | 3,830 | -67% | 1 | 1 | 0% | 1,850 | 1,454 | -21% | 0 | 0 | — |
case-13 | pass→pass | 3,486 | 1,410 | -60% | 1 | 1 | 0% | 563 | 946 | +68% | 0 | 0 | — |
case-15 | fail→pass | 3,239 | 1,806 | -44% | 1 | 1 | 0% | 528 | 998 | +89% | 0 | 0 | — |
case-16 | pass→pass | 6,553 | 1,884 | -71% | 1 | 1 | 0% | 1,135 | 1,072 | -6% | 0 | 0 | — |
case-17 | fail→pass | 11,623 | 3,806 | -67% | 1 | 1 | 0% | 1,936 | 1,394 | -28% | 0 | 0 | — |
case-18 | fail→pass | 3,639 | 1,145 | -69% | 1 | 1 | 0% | 591 | 930 | +57% | 0 | 0 | — |
case-19 | pass→pass | 4,339 | 1,711 | -61% | 1 | 1 | 0% | 773 | 1,019 | +32% | 0 | 0 | — |
case-20 | pass→pass | 17,666 | 7,378 | -58% | 1 | 1 | 0% | 3,269 | 2,208 | -32% | 0 | 0 | — |
case-21 | pass→pass | 11,163 | 8,773 | -21% | 1 | 1 | 0% | 2,094 | 2,474 | +18% | 0 | 0 | — |
case-22 | pass→pass | 17,546 | 13,246 | -25% | 1 | 1 | 0% | 3,637 | 3,417 | -6% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +59 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.