web-search

Solid

USE WHEN the user asks for current facts, recent news, prices/rates, or anything that may have changed since the training cutoff. Plans the query, runs WebSearch / WebFetch (or an MCP fetch tool), and synthesizes a cited answer with date + source URL.

AI & Automation 50 stars 6 forks Updated today MIT

Install

View on GitHub

Quality Score: 78/100

Stars 20%
57
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
86
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# web-search — fresh, cited answers ## When to use Trigger when the question depends on **current state**: - prices, rates, exchange rates, stock quotes, ETF NAV - recent news (last 6 months) - newly released product specs, library versions - regulations / policies that may have changed - comparison ("X vs Y") where one side could have shipped updates Do NOT use for stable facts (math, history, well-settled science) — answer directly. Web searches add latency and cost. ## How to run 1. **Plan the query**. Convert vague Chinese questions into 2–4 targeted English keywords. Strip filler words. Example: "美元兑人民币今天什么价" → `USD CNY exchange rate today`. 2. **Search**. Prefer in this order: - `WebSearch` tool (if available) - `mcp__fetch__fetch` (MCP fetch server) for known URLs - `WebFetch` for known docs / official sources 3. **Verify recency**. Open at least one source, confirm the publication date is within the relevant window. If only stale results, say so explicitly — do NOT pretend the answer is current. 4. **Synthesize**. Output: - One-sentence answer with the **number / fact** - Bullet of 2–3 source citations: `[Title](URL) — YYYY-MM-DD` - If sources disagree, surface the disagreement ## Failure modes to avoid - Quoting yesterday's number as "today's" - Citing a self-media blog when an official source exists (always prefer central-bank / exchange / vendor official pages) - Doing zero searches but writing a confident-sounding paragra...

Details

Author
hesorchen
Repository
hesorchen/muselab
Created
2 months ago
Last Updated
today
Language
Python
License
MIT

Integrates with

Similar Skills

Semantically similar based on skill content — not just same category