juicebox-cost-tuning — Claude Code skill: Juicebox costs
The juicebox-cost-tuning skill helps control Juicebox spending by cutting redundant searches and profile enrichment, then checking changes against a quota baseline.
MIT 1 installs ★ 2,935 updated 2026-10-29
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a claude-codeWhat juicebox-cost-tuning does
Juicebox charges are tied to search queries, enriched profiles, contact lookups, and outreach messages. The skill recommends caching search results, filtering candidates before enrichment, and requesting contact details only for finalists. Use it when you need to reduce unnecessary processing and monitor quota usage.
Before a canary, prepare an approved budget, synthetic test data, a quota baseline, and an owner who can roll back the change. Compare aggregate usage with that baseline and verify source, destination, and suppression controls. Stop the run if its scope or policy drifts; on a budget or quota anomaly, cancel queued work and restore the approved configuration.
Keep a redacted receipt with the environment, aggregate usage, cache or batch setting, approval, and rollback details. Do not include query text, contact data, or credentials. A run that violates source or suppression controls must be rejected.
What it helps with
- cache repeated searches to reduce query usage
- filter candidates before enriching profiles
- compare canary usage with an approved quota baseline
- stop processing when scope or suppression controls drift
Example prompts for the agent
- Review Juicebox costs using synthetic staging data and compare aggregate usage with the quota baseline.
- Suggest ways to reduce profile enrichment while keeping the approved processing scope.
- Create a redacted cost-control receipt without query text, contacts, or credentials.
How to install juicebox-cost-tuning
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a claude-code2Manually: put the files in place
~/.claude/skills/juicebox-cost-tuning/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/juicebox-cost-tuning/
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a codex2Manually: put the files in place
.agents/skills/juicebox-cost-tuning/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a cursor2Manually: put the files in place
.cursor/skills/juicebox-cost-tuning/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install jeremylongshore/tons-of-skills-marketplace skills/.curated/juicebox-cost-tuning2Manually: put the files in place
.github/skills/juicebox-cost-tuning/SKILL.md1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a gemini-cli2Platform CLI
gemini skills install https://github.com/jeremylongshore/tons-of-skills-marketplace --path skills/.curated/juicebox-cost-tuningManually: put the files in place
.devin/skills/juicebox-cost-tuning/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about juicebox-cost-tuning
How do I install juicebox-cost-tuning in Claude Code?
Run npx skills add jeremylongshore/tons-of-skills-marketplace --skill juicebox-cost-tuning -a claude-code. The skill file is at ~/.claude/skills/juicebox-cost-tuning/SKILL.md.
What does juicebox-cost-tuning do?
It helps reduce unnecessary Juicebox searches and profile processing, monitor quota use, and assess cost changes with a canary. The run should stop if it departs from approved scope or controls.
What should I prepare before a cost canary?
Prepare an approved budget, synthetic test data, a current quota baseline, and a rollback owner. The process also checks source, destination, and suppression controls.
Source text
SKILL.md · MIT
--- name: juicebox-cost-tuning description: 'Optimize Juicebox costs. Trigger: "juicebox cost", "juicebox billing", "juicebox budget". ' 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 Cost Tuning ## Cost Factors | Feature | Cost Driver | |---------|-------------| | Search | Per query | | Enrichment | Per profile | | Contact data | Per lookup | | Outreach | Per message | ## Reduction Strategies 1. Cache search results (avoid duplicate queries) 2. Use filters (fewer wasted enrichments) 3. Only enrich top-scored candidates 4. Only get contacts for final candidates ## Quota Monitoring ```typescript const quota = await client.account.getQuota(); console.log(`Searches: ${quota.searches.used}/${quota.searches.limit}`); if (quota.searches.used > quota.searches.limit * 0.8) console.warn('80% quota used'); ``` ## Overview Use cost controls as a safety boundary: validate them first in a sandbox, reduce unnecessary processing, and never treat cost optimization as permission to expand data or outreach scope. ## Prerequisites - An approved budget, synthetic sandbox fixture, source/destination allowlists, suppression controls, current quota baseline, and a rollback owner. ## Instructions 1. Measure cache and batching changes with synthetic inputs and aggregate counters only; reject unapproved sources, destinations, or contact exports. 2. Compare each canary to the approved quota baseline, verify suppression and `contacts_exported=0`, and stop on material scope or policy drift. 3. Promote cost settings only after owner approval; restore the prior configuration and cancel queued work on failure. 4. Retain a redacted summary and delete test artifacts at the end of the approved window. ## Output Record a cost-control receipt with environment, baseline and aggregate usage, cache/batch setting, suppression/no-export results, owner approval, rollback action, and retention/deletion proof. Exclude query text, contacts, and credentials. ## Error Handling | Condition | Response | |---|---| | Budget or quota anomaly | Pause the canary, cancel queued work, and restore the approved configuration. | | Source, destination, or suppression drift | Reject the run and investigate with redacted aggregate telemetry only. | ## Examples `env=staging; fixture=synthetic; cache=enabled; usage_delta=-18%; suppression=pass; contacts_exported=0; rollback=available` supports an approval decision. ## Resources - [Pricing](https://juicebox.ai/pricing) ## Next Steps See `juicebox-reference-architecture`.
© Jeremy Longshore · MIT · jeremylongshore/tons-of-skills-marketplace