Check AI-generated claims against your approved Fact Sheet and flag changed quantities, expanded scope, dropped attribution, and unsupported mappings — deterministic, local, and no AI guessing.
Catch AI hallucinations before you hit publish. This skill scans articles, blog posts, marketing copy, and AI-drafted text for unsourced stats, vague "studies show" claims, fabricated quotes, and inflated social proof numbers. It flags each claim by risk level and offers three modes: a fast pre-publish scan, a detailed claim-by-claim report, or honest rewrites. Built-in safeguard: it never invents a source or "verified" status to fill a gap — unverifiable stays labeled unverifiable.
Finds genuinely lesser-known, web-verified facts on any topic you choose, always with a cited source. Responds in your language.
Turns a press release, official announcement, or source article into a publish-ready article while guaranteeing every date, amount, name, and location stays locked to the source — zero fabrication, verified before output.
日々AIの回答・ニュース・SNSの言説を実務で扱う中で、出典のない数値や「最新」を装った古い情報に何度も振り回された経験から、このSkillを設計しました。主張を安定/不安定に分類し、不安定な情報ほど一次ソースを最優先で確認、重要な主張は独立した複数ソースでクロスチェックする手順を体系化しています。AI生成記事が引用元の主張を実際には支持していないという罠も検出します。
This Skill is built on practical expertise in claim-evidence mapping, source authority labeling, quote verification, citation hygiene, recency checks, and hallucination-risk control. It helps AI agents turn research, reports, summaries, briefs, and marketing drafts into auditable claim ledgers where every important claim is matched to a source, date, quote/paraphrase boundary, and confidence label.
在发布前检查 README、论文或报告中的声明是否仍与当前 CSV/JSON 结果一致，输出带证据的 SUPPORTED、STALE 或 UNSUPPORTED 闸门报告。
Forces AI agents to anchor every interview answer in verified company facts, mark all personal examples as [USER_INPUT_REQUIRED] until you provide them, and label every unsourced claim — so nothing fabricated makes it into your next interview.