HOMESKILLSBLOGGITHUB
// SKILL

FACT CHECKING

Verify every statistic and claim in your blog post. The fact checker extracts all statistical claims, fetches each cited source URL, and scores match confidence. Exact matches score 1.0, approximate matches lower. No more fabricated data.

$
/blog factcheck ./content/blog/my-post.md

REQUIRES CLAUDE BLOG INSTALLED IN CLAUDE CODE

// HOW IT WORKS

VERIFY. SCORE. TRUST.

Running /blog factcheck scans your blog post for every statistical claim, percentage, date, and factual assertion. For each claim, it identifies the cited source, fetches the page, and compares the data. You get a confidence score for every claim in your post.

Claims without sources are flagged for manual addition. Broken source URLs trigger alternative source suggestions. The result is a blog post where every number is verifiable.

01
CLAIM EXTRACTION
Scans your post for all statistical claims: numbers, percentages, dates, rankings, and factual assertions. Each claim is isolated and paired with its cited source URL.
02
SOURCE FETCHING
Fetches each cited URL, parses the content, and searches for the specific data point. Handles HTML, PDF, and plain text sources. Reports inaccessible URLs.
03
CONFIDENCE SCORING
Each claim receives a 0.0 to 1.0 confidence score. 1.0 means exact match found in source. Below 0.5 means the claim needs manual verification or a better source.
04
REPORT GENERATION
Produces a verification report listing every claim, its source, confidence score, and recommendation. Unsourced claims and broken URLs get alternative source suggestions.
// USAGE

HOW TO FACT CHECK

Step 1: Install Claude Blog

git clone --depth 1 https://github.com/AgriciDaniel/claude-blog.git && bash claude-blog/install.sh

Step 2: Open Claude Code

claude

Step 3: Run the Fact Checker

/blog factcheck ./content/blog/my-post.md

The fact checker scans your post, fetches all cited sources, and produces a verification report with confidence scores for every claim.

// FAQ

QUESTIONS ABOUT FACT CHECKING

The fact checker scans your blog post for statistical claims (numbers, percentages, dates, rankings). For each claim, it identifies the cited source URL, fetches the page content, and searches for the specific data point. It then compares the claim in your post against the source and produces a match confidence score.
The confidence score ranges from 0.0 to 1.0. A score of 1.0 means the exact number was found in the source. Scores between 0.7 and 0.99 indicate approximate matches where the source contains similar but not identical data. Scores below 0.5 suggest the claim could not be verified and needs manual review.
The fact checker can fetch and parse most public web pages. It handles HTML pages, PDFs, and plain text. Pages behind paywalls, login walls, or heavy JavaScript rendering may not be fully accessible. The checker reports when a source URL cannot be fetched so you can verify manually.
When a source URL returns a 404 or the claim cannot be verified, the fact checker suggests alternative source URLs where the same data point might be found. These suggestions are based on the claim context and common authoritative sources in the topic area.
// RELATED SKILLS

EXPLORE MORE

BLOG ANALYZE

Deep analysis of individual blog posts with readability scoring, keyword density, and content quality assessment.

LEARN MORE →

BLOG WRITE

AI-powered blog post writing with SEO optimization, persona-aware tone, and automatic internal linking.

LEARN MORE →

BLOG REWRITE

Rewrite and refresh existing blog posts with updated information, improved structure, and persona alignment.

LEARN MORE →

VIEW ALL SKILLS →

VERIFY YOUR CLAIMS
IN 30 SECONDS.

$
git clone --depth 1 https://github.com/AgriciDaniel/claude-blog.git && bash claude-blog/install.sh
VIEW ON GITHUB ALL SKILLS >