Juicebox Performance TuningSAFE

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: juicebox-performance-tuning
description: 'Optimize Juicebox performance.

  Trigger: "juicebox performance", "optimize juicebox".

  '
allowed-tools: Read, Write, Edit, Grep
version: 1.16.0
license: MIT
author: Jeremy Longshore <[email protected]>
tags:
- saas
- recruiting
- juicebox
compatibility: Designed for Claude Code
---
# Juicebox Performance Tuning

## Overview

Juicebox's AI analysis API handles dataset uploads, analysis queue wait times, and result pagination. Large dataset uploads (100K+ rows) can block the analysis pipeline, while queue contention during peak hours increases wait times. Result sets from broad queries return thousands of profiles requiring efficient pagination. Caching search results, batching enrichment calls, and managing upload chunking reduces end-to-end analysis time by 40-60% and keeps interactive searches responsive.

## Caching Strategy

```typescript
const cache = new Map<string, { data: any; expiry: number }>();
const TTL = { search: 300_000, profile: 600_000, analysis: 900_000 };

async function cached(key: string, ttlKey: keyof typeof TTL, fn: () => Promise<any>) {
  const entry = cache.get(key);
  if (entry && entry.expiry > Date.now()) return entry.data;
  const data = await fn();
  cache.set(key, { data, expiry: Date.now() + TTL[ttlKey] });
  return data;
}
// Analysis results are expensive — cache 15 min. Searches expire at 5 min.
```

## Batch Operations

```typescript
async function enrichBatch(client: any, profileIds: string[], batchSize = 50) {
  const results = [];
  for (let i = 0; i < profileIds.length; i += batchSize) {
    const batch = profileIds.slice(i, i + batchSize);
    const res = await client.enrichBatch({ profile_ids: batch, fields: ['skills_map', 'contact'] });
    results.push(...res.profiles);
    if (i + batchSize < profileIds.length) await new Promise(r => setTimeout(r, 300));
  }
  return results;
}
```

## Connection Pooling

```typescript
import { Agent } from 'https';
const agent = new Agent({ keepAlive: true, maxSockets: 8, maxFreeSockets: 4, timeout: 60_000 });
// Longer timeout for dataset uploads and analysis queue responses
```

## Rate Limit Management

```typescript
async function withRateLimit(fn: () => Promise<any>): Promise<any> {
  try { return await fn(); }
  catch (err: any) {
    if (err.status === 429) {
      const backoff = parseInt(err.headers?.['retry-after'] || '10') * 1000;
      await new Promise(r => setTimeout(r, backoff));
      return fn();
    }
    throw err;
  }
}
```

## Monitoring

```typescript
const metrics = { searches: 0, enrichments: 0, cacheHits: 0, queueWaitMs: 0, errors: 0 };
function track(op: 'search' | 'enrich', startMs: number, cached: boolean) {
  metrics[op === 'search' ? 'searches' : 'enrichments']++;
  metrics.queueWaitMs += Date.now() - startMs;
  if (cached) metrics.cacheHits++;
}
```

## Performance Checklist

- [ ] Use specific filters (location, skills, title) to narrow search scope
- [ ] Cache search results with 5-min TTL to avoid redundant queries
- [ ] Batch profile enrichment in groups of 50 with 300ms delays
- [ ] Chunk large dataset uploads into 10K-row segments
- [ ] Cache analysis results for 15 min (expensive to recompute)
- [ ] Set 60s timeout for upload and analysis endpoints
- [ ] Monitor queue wait times and schedule uploads during off-peak
- [ ] Paginate results with limit=20 and cursor for interactive UIs

## Error Handling

| Issue | Cause | Fix |
|-------|-------|-----|
| Analysis queue timeout | Peak hour contention | Schedule large analyses off-peak, increase client timeout |
| 429 on bulk enrichment | Too many concurrent enrichment calls | Batch to 50 profiles with 300ms interval |
| Upload failure on large dataset | Payload exceeds limit or connection drop | Chunk into 10K-row segments, retry failed chunks |
| Slow broad search | Unfiltered query returning thousands of results | Add location/skills/title filters, set limit=20 |

## Prerequisites

- An approved performance baseline, synthetic sandbox fixture, bounded test budget, source/destination allowlists, suppression controls, and a rollback owner.

## Instructions

1. Benchmark cache, batching, and pagination changes against synthetic fixtures only; reject live contact export and unapproved destinations.
2. Collect aggregate latency, error, and quota measurements; verify suppression, data minimization, and `contacts_exported=0` before comparison.
3. Run one bounded canary, halt on scope, policy, quota, or retention drift, and restore the prior tuning configuration if it fails.
4. Keep only the redacted benchmark receipt and delete test artifacts after the approved window.

## Output

Produce a performance receipt with environment, baseline and aggregate measurements, tuning settings, suppression/no-export assertions, canary outcome, owner approval, retention/deletion proof, and rollback reference. Exclude queries, records, and credentials.

## Examples

`env=staging; fixture=synthetic; p95_delta=-22%; quota=within-budget; suppression=pass; contacts_exported=0; rollback=available` supports an approval decision.

## Resources

- Juicebox API Docs
- Juicebox Performance Guide

## Next Steps

See `juicebox-reference-architecture`.

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.