From plaited-plaited
Use You.com Search, conditional livecrawl/contents, and narrow Research API follow-ups through the official client first, with direct HTTP only as fallback. - MANDATORY TRIGGERS: You.com, youdotcom, YDC, YDC API, You.com API, livecrawl, contents, research lite, research deep, cited web research, direct API integration - Use when: web-grounded search or extraction is needed, or when a repo script should call You.com programmatically
npx claudepluginhub joshuarweaver/cascade-code-general-misc-3 --plugin plaited-plaitedThis skill is limited to using the following tools:
Use You.com through the official client when possible. In this repo, prefer
Applies Acme Corporation brand guidelines including colors, fonts, layouts, and messaging to generated PowerPoint, Excel, and PDF documents.
Builds DCF models with sensitivity analysis, Monte Carlo simulations, and scenario planning for investment valuation and risk assessment.
Calculates profitability (ROE, margins), liquidity (current ratio), leverage, efficiency, and valuation (P/E, EV/EBITDA) ratios from financial statements in CSV, JSON, text, or Excel for investment analysis.
Use You.com through the official client when possible. In this repo, prefer
@youdotcom-oss/api inside Bun scripts and let varlock inject
YDC_API_KEY.
Use raw fetch only when:
Default behavior in this repo should be:
This matters because:
research lite prompts often return weak or no-result outputfetch against You.com Search can show transport/decompression instability
under longer batch runs@youdotcom-oss/api client is the better default scripted path
before adding custom retry logic@youdotcom-oss/api over raw fetch for Bun scripts.varlock to inject YDC_API_KEY.In this repo:
const apiKey = process.env.YDC_API_KEY
if (!apiKey) throw new Error('YDC_API_KEY is required')
Typical command surface:
bunx varlock run -- bun scripts/some-script.ts
Use the official client first:
import { fetchSearchResults } from '@youdotcom-oss/api'
const response = await fetchSearchResults({
searchQuery: {
query: 'student behavior incident tracking software',
count: 5,
},
YDC_API_KEY: process.env.YDC_API_KEY,
getUserAgent: () => 'Plaited/7.x (You.com)',
})
console.log(response.results.web)
fetch plumbingfetch path we were using during Slice 14 work| Endpoint | Method | URL | Use |
|---|---|---|---|
| Search | GET | https://api.you.com/v1/agents/search | default discovery step |
| Research | POST | https://api.you.com/v1/research | narrow cited synthesis |
| Contents | POST | https://ydc-index.io/v1/contents | explicit URL extraction |
JSON Schemas remain in assets/:
Use this decision tree:
Prefer:
Examples:
student behavior incident tracking softwarereference metadata cleanup validation export workflowfont library browser preview catalogBetter flow:
Use research lite only when:
Avoid research lite first for:
Use deep only when:
Use raw fetch only when the client is not suitable:
const query = new URLSearchParams({
query: 'student behavior incident tracking workflow',
count: '3',
livecrawl: 'web',
livecrawl_formats: 'markdown',
})
const response = await fetch(`https://api.you.com/v1/agents/search?${query}`, {
headers: { 'X-API-Key': process.env.YDC_API_KEY! },
})
If this raw path shows transport or decompression instability in a batch script:
Prefer Search + livecrawl when:
Prefer Contents when:
Do not use both by default.
When using contents on very large single pages (for example full arXiv HTML
papers), the returned payload can be truncated and end before valid JSON close.
Treat large single-page extraction as potentially partial unless validated.
Use this operational pattern:
contents (for example .../html/...)jq, script parse, or schema parse)contents on a compact variant
(for arXiv, prefer https://arxiv.org/abs/<id>) to recover clean metadataThis keeps extraction robust without switching away from the official
@youdotcom-oss/api path.