Juicebox Reference Architecture

What it does

When your agent uses it

• Tasks that involve Recruiting and HR

Example prompts

• “juicebox architecture”
• “recruiting platform design”
• “/juicebox-reference-architecture”

Requirements

• Compatibility (from SKILL.md): Designed for Claude Code
• Pre-approved tools (allowed-tools): Read, Write, Edit, Grep

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

1. 1Validate each component in a sandbox with synthetic data; reject literal credentials, unapproved integrations, and any contact-level export.
2. 2Enforce least privilege, idempotency, redacted telemetry, suppression checks, and contacts_exported=0 before connecting a new component.
3. 3Run a staged canary and halt on scope, security, quota, policy, or retention drift; restore the prior component revision if it fails.
4. 4Promote only with owner approval and preserve only the redacted architecture receipt after cleanup.

What it can do on your machine

Read from SKILL.md and the folder at commit cfae287. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves these tools, so the agent can use them without asking each time:

  • Read
  • Write
  • Edit
  • Grep

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are typescript).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Links to these hosts (documentation or services it may open):

  • juicebox.ai

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

• Compatibility

  Designed for Claude Code

  From compatibility in the SKILL.md frontmatter.

Context cost

Juicebox Reference Architecture loads about 1.7k tokens when it runs, and up to ~7.4k if it reads all its reference files. Until then it costs about 35 tokens; SKILL.md has 350 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~35

When it runs · the whole SKILL.md, loaded when a task matches

~1.7k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~7.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from jeremylongshore/tons-of-skills-marketplace at commit cfae287, republished under its MIT licence (© jeremylongshore). 350 words, ~1,677 tokens.

name

juicebox-reference-architecture

description

Implement Juicebox reference architecture. Trigger: "juicebox architecture", "recruiting platform design".

allowed-tools

Read, Write, Edit, Grep

compatibility

Designed for Claude Code

version

1.16.0

license

MIT

author

Jeremy Longshore <jeremy@intentsolutions.io>

tags

saas, recruiting, juicebox

Recruiter Dashboard ──→ Search Service ──→ Cache (Redis) ──→ Juicebox API
                             ↓                                /search
                        Queue (Bull) ──→ Analysis Worker      /profiles
                             ↓                                /outreach
                        ATS Export Service ──→ Greenhouse/Lever
                             ↓
                        Webhook Handler ←── Juicebox Events

class CandidateSearchService {
  constructor(private juicebox: JuiceboxClient, private cache: CacheLayer) {}

  async findAndRank(criteria: SearchCriteria): Promise<RankedCandidate[]> {
    const cacheKey = `search:${this.hashCriteria(criteria)}`;
    const cached = await this.cache.get(cacheKey);
    if (cached) return cached;
    const results = await this.juicebox.search(criteria);
    const ranked = results.profiles.map(p => ({ ...p, score: this.scoreCandidate(p, criteria) }))
      .sort((a, b) => b.score - a.score);
    await this.cache.set(cacheKey, ranked, CACHE_CONFIG.searchResults.ttl);
    return ranked;
  }

  async exportToATS(candidates: string[], jobId: string, ats: 'greenhouse' | 'lever'): Promise<ExportResult> {
    const deduped = await this.deduplicateAgainstATS(candidates, jobId, ats);
    return this.juicebox.export({ profiles: deduped, destination: ats, job_id: jobId });
  }
}

const CACHE_CONFIG = {
  searchResults: { ttl: 1800, prefix: 'search' },   // 30 min — candidate pools shift slowly
  profiles:      { ttl: 3600, prefix: 'profile' },   // 1 hr — profile data stable short-term
  analysisRuns:  { ttl: 7200, prefix: 'analysis' },   // 2 hr — analysis results are expensive to recompute
  atsState:      { ttl: 300,  prefix: 'ats' },        // 5 min — ATS pipeline freshness for dedup
  outreach:      { ttl: 60,   prefix: 'outreach' },   // 1 min — sequence status changes frequently
};
// New search invalidates matching cached results; ATS export clears ats cache for that job

class RecruitingPipeline {
  private queue = new Bull('juicebox-events', { redis: process.env.REDIS_URL });

  async onSearchComplete(searchId: string, results: RankedCandidate[]): Promise<void> {
    await this.queue.add('analyze', { searchId, candidateIds: results.map(r => r.id) },
      { attempts: 3, backoff: { type: 'exponential', delay: 2000 } });
  }

  async processOutreachEvent(event: OutreachEvent): Promise<void> {
    if (event.type === 'reply_received') await this.flagForRecruiterReview(event);
    if (event.type === 'bounced') await this.markInvalid(event.candidateId);
    await this.syncStatusToATS(event);
  }
}

interface SearchCriteria   { role: string; skills: string[]; location?: string; experienceYears?: number; companySize?: string; }
interface RankedCandidate  { id: string; name: string; title: string; company: string; score: number; skills: string[]; profileUrl: string; }
interface OutreachSequence { id: string; candidateId: string; jobId: string; steps: OutreachStep[]; status: 'active' | 'replied' | 'bounced' | 'opted-out'; }
interface ExportResult     { exported: number; duplicatesSkipped: number; atsJobId: string; }

© jeremylongshore, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (references) in skills/.curated/juicebox-reference-architecture of jeremylongshore/tons-of-skills-marketplace.

• SKILL.md
• references/implementation-guide.md
• references/service-layer.md

Open the folder on GitHubat commit cfae287