claim-verify_skill

Overview

This skill verifies factual claims against live data sources and returns structured verdicts with confidence scores. It produces a current truth value, freshness indicator, source count, and whether the result was cached. The API supports optional categories for faster routing and offers a trending endpoint for popular queries.

How this skill works

Send a claim to the verification endpoint and the service queries live data sources and caches for up to one hour. It aggregates evidence, computes a confidence score, determines a verdict (confirmed, stale, disputed, false, unknown), and returns metadata such as source_count, freshness, and request_id. Optional category hints (financial, entity, geo, factcheck) route queries to specialized pipelines for quicker, more relevant checks.

When to use it

• Validate numbers or time-sensitive facts (exchange rates, prices, population).
• Check claims before publishing or sharing content publicly.
• Automate fact-checking in workflows and moderation pipelines.
• Compare multiple sources quickly to resolve disputes among datasets.

Best practices

• Provide concise, specific claims to reduce ambiguity (e.g., include units and time context).
• Use category hints for financial, entity, geo, or factcheck to improve routing and speed.」「Batch similar queries when possible to stay within free tier limits and reduce latency.
• Treat confidence scores and source counts as guidance; follow up manually for high-stakes decisions.

Example use cases

• Content moderation: automatically verify assertions in user-generated posts before allowing publication.
• Editorial workflows: fact-check quotes, dates, and figures during publishing.
• Finance bots: verify live exchange rates, crypto prices, and stock quotes before executing trades.
• Research support: confirm entity facts (founding dates, population) for reports and briefs.
• Analytics and dashboards: flag stale or disputed metrics in live reports.