web-search-verifier

Cross-verify macroeconomic numeric data from three independent web sources.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/yjang-git/HoneyPot --skill web-search-verifier-yjang-git
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: web-search-verifier
Source: https://github.com/yjang-git/HoneyPot/tree/main/plugins/investments-portfolio/skills/web-search-verifier
Command: npx skills add https://github.com/yjang-git/HoneyPot --skill web-search-verifier-yjang-git

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill standardizes web search workflows to prevent hallucinated numeric data when collecting macroeconomic indices, exchange rates, and interest rates by enforcing multi-source cross-verification and strict citation rules.

Core Features & Use Cases

  • Three-source cross-check: Requires at least three independent sources with one tier-1 official source for every numeric value before acceptance.
  • Original-text provenance: Mandates including the exact quoted original_text from search results alongside parsed numeric values and source URLs.
  • Fail-safe validation: Enforces date alignment, ±1% variance tolerance, blocklist avoidance, retry policies, and clear failure codes for automated handling.
  • Use Case: An index-fetcher agent queries S&P 500, runs three parallel searches with mcp_websearch_web_search_exa, extracts original_text from each result, computes variance, and returns a verified JSON payload or a FAIL with diagnostic code.

Quick Start

Ask the agent to call mcp_websearch_web_search_exa for the target metric, extract original_text and URLs from three independent sources including one official source, verify date and ±1% variance, and return the structured verified result.

Frequently Asked Questions about web-search-verifier

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent hallucinated macroeconomic data when using web search agents?

To prevent hallucinated macroeconomic data, enforce multi-source cross-verification by requiring agents to collect numeric values from at least three independent sources, including one official tier-1 source, and validate them before acceptance.

How do I verify exchange rates and interest rates from web search results?

Verify exchange rates and interest rates by extracting exact original_text quotes and source URLs from three independent search results, enforcing date alignment, and checking that values fall within a ±1% variance tolerance.

What is the best way to validate numeric data from mcp_websearch for investment workflows?

The best way to validate numeric data for investment workflows is to cross-verify values across three independent sources, mandate exact original-text provenance alongside parsed numbers, and return a structured JSON payload or an explicit FAIL code.

Can I use automated web search to collect macroeconomic indices without source attribution?

No, collecting macroeconomic indices without source attribution risks hallucination. You must include exact original_text quotes alongside parsed numeric values and source URLs from at least three independent sources for reliable reporting.

Why does my agent return inconsistent numeric values for macroeconomic indices?

Inconsistent numeric values for macroeconomic indices occur when agents lack strict cross-verification. Requiring three independent sources, enforcing ±1% variance tolerance, and checking date alignment prevents this data inconsistency.

What happens when web search verification fails to find three independent sources for interest rates?

When web search verification fails to find three independent sources for interest rates, the workflow returns an explicit FAIL code with diagnostics, enforcing retry policies and blocklist avoidance for automated handling instead of accepting unverified data.