skills.rest
sparkst
sparkstCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

Fact Checking Protocol

Validates research claims against tiered sources and checks date alignment with a Python script.

Updated Jan 29, 2026
Search Tags:#research#evidence#citation#source verification#claim validation#fact checking#date checking
One-click install
npx skills add https://github.com/sparkst/sparkry-claude-skills --skill fact-checking-protocol-sparkst
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Fact Checking Protocol
Source: https://github.com/sparkst/sparkry-claude-skills/tree/main/plugins/research-workflow/skills/research/fact-check
Command: npx skills add https://github.com/sparkst/sparkry-claude-skills --skill fact-checking-protocol-sparkst

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Research deliverables often contain unverified claims backed by weak, stale, or circular sources, leading to decisions grounded in unreliable evidence. ## Core Features & Use Cases - Claim Budget Enforcement: Limits deliverables to a small set of load-bearing claims so each receives proper validation. - Tiered Source Requirements: Applies stakes-based rules requiring two independent Tier-1 sources for high-stakes claims, with soft-gate warnings instead of hard blocks. - Date Alignment Checking: Uses scripts/date_checker.py to flag stale, mismatched, or future-dated sources relative to claim timeframes. - Use Case: Before finalizing a market research report, validate that the TAM figure is corroborated by two independent analyst sources published within 180 days, and surface warnings for any claim that falls short. ## Quick Start Validate the claims in my research deliverable against the tiered source requirements and flag any stale or weakly supported claims.

Frequently Asked Questions about Fact Checking Protocol

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

FAQPage Schema
How do I validate research claims with multiple sources?▼

Assign each claim a type and stakes level, then attach supporting sources with tier ratings. High-stakes claims require at least two independent Tier-1 sources, while medium-stakes claims accept one Tier-1 plus one Tier-2 or two Tier-2 sources.

What is a load-bearing claim in research validation?▼

A load-bearing claim is a statement that, if wrong, would invalidate the recommendation or materially change the decision. The protocol caps these at a claim budget of ten so each receives proper source validation.

How does the date checker detect stale sources?▼

The date_checker.py script extracts years from claim timeframes and source publication dates, then compares them against a tolerance window. Sources older than the tolerance are flagged as stale, and sources postdating the claim are flagged as future.

Does claim validation block delivery when sources are weak?▼

No, it operates as a soft gate. Claims lacking sufficient Tier-1 corroboration are flagged with a warning or supplemental status, concerns are documented, and the user decides whether to proceed or request more research.

Can Tier-3 sources like Reddit support research claims?▼

Yes, for qualitative claim types such as customer needs and sentiment, Tier-3 sources like Reddit or Hacker News threads are acceptable. They are marked as supplemental and are not sufficient for high-stakes market or financial claims.

Related Skills

Explore More Education & Research Skills→
clawpod-appclawpod-app
46

source-verification

Trace claims to primary sources and assess credibility with recency checks.

Community
Intermediate
scanadyscanady

research-market-analyst

Classify competitive and market signals into confidence-rated strategic intelligence reports.

Community
Intermediate
drewid74drewid74
8

research-analyst

Triangulate claims across multiple sources and assign confidence scores.

Community
Advanced
DevHive1DevHive1
18

web-search

Execute web searches and fetch URL content as structured markdown.

Community
Intermediate