Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Deploy Juicebox integrations. Trigger: "deploy juicebox", "juicebox production deploy".
.claude/skills/jeremylongshore-juicebox-deploy-integration/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -17% | 0% |
Deploy a containerized Juicebox AI analysis integration service with Docker. This skill covers building a production image that connects to the Juicebox API for managing datasets, running AI-powered analyses, and retrieving structured insights. Includes environment configuration for dataset access and analysis pipelines, health checks that verify API connectivity and dataset availability, and rolling update strategies for zero-downtime deployments serving real-time analysis results.
dockerfileFROM node:20-slim AS builder WORKDIR /app COPY package*.json ./ RUN npm ci COPY tsconfig.json ./ COPY src/ ./src/ RUN npm run build FROM node:20-slim RUN addgroup --system app && adduser --system --ingroup app app WORKDIR /app COPY --from=builder /app/dist ./dist COPY --from=builder /app/node_modules ./node_modules COPY package*.json ./ USER app EXPOSE 3000 HEALTHCHECK --interval=30s --timeout=5s --retries=3 \ CMD curl -f http://localhost:3000/health || exit 1 CMD ["node", "dist/index.js"]
bashexport JUICEBOX_API_KEY="jb_live_xxxxxxxxxxxx" export JUICEBOX_BASE_URL="https://api.juicebox.ai/v1" export JUICEBOX_WORKSPACE_ID="ws_xxxxxxxxxxxx" export LOG_LEVEL="info" export PORT="3000" export NODE_ENV="production"
typescriptimport express from 'express'; const app = express(); app.get('/health', async (req, res) => { try { const response = await fetch(`${process.env.JUICEBOX_BASE_URL}/datasets`, { headers: { 'Authorization': `Bearer ${process.env.JUICEBOX_API_KEY}` }, }); if (!response.ok) throw new Error(`Juicebox API returned ${response.status}`); res.json({ status: 'healthy', service: 'juicebox-integration', timestamp: new Date().toISOString() }); } catch (error) { res.status(503).json({ status: 'unhealthy', error: (error as Error).message }); } });
bashdocker build -t juicebox-integration:latest .
bashdocker run -d --name juicebox-integration \ -p 3000:3000 \ -e JUICEBOX_API_KEY -e JUICEBOX_BASE_URL -e JUICEBOX_WORKSPACE_ID \ juicebox-integration:latest
bashcurl -s http://localhost:3000/health | jq .
bashdocker build -t juicebox-integration:v2 . && \ docker stop juicebox-integration && \ docker rm juicebox-integration && \ docker run -d --name juicebox-integration -p 3000:3000 \ -e JUICEBOX_API_KEY -e JUICEBOX_BASE_URL -e JUICEBOX_WORKSPACE_ID \ juicebox-integration:v2
| Issue | Cause | Fix | |-------|-------|-----| | 401 Unauthorized | Invalid or expired API key | Regenerate key in Juicebox workspace settings | | 403 Forbidden | Workspace access denied | Verify JUICEBOX_WORKSPACE_ID matches your API key | | 404 Not Found | Dataset or analysis ID not found | Check IDs from Juicebox dashboard | | 429 Rate Limited | Exceeding API rate limits | Implement exponential backoff; batch analysis requests | | Analysis timeout | Large dataset processing | Increase timeout or use async analysis endpoint with polling |
See juicebox-webhooks-events.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→pass | 23,312 | 10,858 | -53% | 1 | 1 | 0% | 2,549 | 2,901 | +14% | 0 | 0 | — |
case-01 | fail→fail | 14,748 | 17,985 | +22% | 1 | 1 | 0% | 2,993 | 3,631 | +21% | 0 | 0 | — |
case-02 | fail→fail | 21,571 | 17,905 | -17% | 1 | 1 | 0% | 3,584 | 3,923 | +9% | 0 | 0 | — |
case-03 | fail→pass | 22,464 | 6,902 | -69% | 1 | 1 | 0% | 3,046 | 2,341 | -23% | 0 | 0 | — |
case-04 | fail→fail | 25,621 | 15,435 | -40% | 1 | 1 | 0% | 2,809 | 4,070 | +45% | 0 | 0 | — |
case-05 | fail→fail | 12,339 | 13,737 | +11% | 1 | 1 | 0% | 1,952 | 2,542 | +30% | 0 | 0 | — |
case-06 | fail→fail | 23,994 | 24,701 | +3% | 1 | 1 | 0% | 3,879 | 5,229 | +35% | 0 | 0 | — |
case-07 | pass→pass | 18,958 | 15,843 | -16% | 1 | 1 | 0% | 2,211 | 2,680 | +21% | 0 | 0 | — |
case-08 | fail→pass | 12,737 | 3,103 | -76% | 1 | 1 | 0% | 2,149 | 1,484 | -31% | 0 | 0 | — |
case-09 | pass→pass | 16,104 | 7,900 | -51% | 1 | 1 | 0% | 1,988 | 2,296 | +15% | 0 | 0 | — |
case-11 | fail→pass | 5,368 | 8,406 | +57% | 1 | 1 | 0% | 909 | 1,587 | +75% | 0 | 0 | — |
case-12 | fail→pass | 13,843 | 8,360 | -40% | 1 | 1 | 0% | 1,780 | 1,469 | -17% | 0 | 0 | — |
case-13 | pass→fail | 11,890 | 6,912 | -42% | 1 | 1 | 0% | 1,101 | 1,236 | +12% | 0 | 0 | — |
case-14 | pass→pass | 10,093 | 9,652 | -4% | 1 | 1 | 0% | 1,790 | 1,659 | -7% | 0 | 0 | — |
case-15 | fail→pass | 19,847 | 8,716 | -56% | 1 | 1 | 0% | 2,155 | 1,673 | -22% | 0 | 0 | — |
case-16 | fail→pass | 16,159 | 3,778 | -77% | 1 | 1 | 0% | 1,515 | 1,695 | +12% | 0 | 0 | — |
case-17 | pass→pass | 10,735 | 2,688 | -75% | 1 | 1 | 0% | 962 | 1,338 | +39% | 0 | 0 | — |
case-18 | pass→pass | 11,101 | 2,848 | -74% | 1 | 1 | 0% | 857 | 1,488 | +74% | 0 | 0 | — |
case-19 | fail→pass | 18,650 | 4,177 | -78% | 1 | 1 | 0% | 2,285 | 1,794 | -21% | 0 | 0 | — |
case-20 | fail→pass | 10,122 | 4,602 | -55% | 1 | 1 | 0% | 1,683 | 1,671 | -1% | 0 | 0 | — |
case-21 | fail→pass | 16,356 | 8,988 | -45% | 1 | 1 | 0% | 1,934 | 1,583 | -18% | 0 | 0 | — |
case-22 | pass→fail | 21,027 | 16,087 | -23% | 1 | 1 | 0% | 2,641 | 2,904 | +10% | 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 +36 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.