Skills clarity-gate
📦

clarity-gate

Safe

Validate RAG content quality before ingestion

Prevent low-quality or hallucinated content from entering your RAG knowledge base. This skill provides a systematic verification framework with nine quality checkpoints and human-in-the-loop validation.

Supports: Claude Codex Code(CC)
📊 70 Adequate
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "clarity-gate". A technical documentation page about API authentication

Expected outcome:

  • Quality Assessment: APPROVED with minor recommendations
  • - Factual accuracy: Verified against official API docs
  • - Source credibility: Official documentation
  • - Information completeness: Missing rate limiting section
  • - Logical consistency: Pass
  • - Temporal relevance: Current (2024)
  • - Technical depth: Appropriate for developer audience
  • - Audience appropriateness: Clear and well-structured
  • - Citation quality: Good - includes code examples
  • - Redundancy check: No duplicate content found
  • Recommendation: Add rate limiting documentation before ingestion.

Using "clarity-gate". A blog post about industry trends from 2019

Expected outcome:

  • Quality Assessment: REJECTED
  • - Temporal relevance: FAIL - Content is 5 years old
  • - Factual accuracy: Cannot verify - sources outdated
  • - Recommendation: Reject due to age. Consider updating with recent data if the information is still valuable.

Security Audit

Safe
v1 • 2/25/2026

All static analysis findings are false positives. The skill contains only documentation content (SKILL.md) with no executable code. Hardcoded URLs are legitimate GitHub repository links in documentation. No actual cryptographic algorithms or security-sensitive patterns present.

1
Files scanned
23
Lines analyzed
0
findings
1
Total audits
No security issues found
Audited by: claude

Quality Score

38
Architecture
100
Maintainability
87
Content
31
Community
100
Security
83
Spec Compliance

What You Can Build

Enterprise knowledge base management

Teams building RAG applications can implement quality gates to prevent low-quality documents from entering their corporate knowledge base.

Research content curation

Researchers and academic teams can verify the quality and reliability of sources before adding them to retrieval systems.

Customer support knowledge base

Support teams can ensure documentation quality and accuracy before publishing to AI-powered customer support systems.

Try These Prompts

Basic content verification 
Review this document for quality issues before adding it to our RAG knowledge base. Check for clarity, accuracy, and completeness.
Nine-point quality assessment 
Evaluate this content using the clarity-gate framework: 1) Factual accuracy 2) Source credibility 3) Information completeness 4) Logical consistency 5) Temporal relevance 6) Technical depth 7) Audience appropriateness 8) Citation quality 9) Redundancy check. Provide scores for each dimension.
Two-round HITL validation 
Round 1: Perform automated quality checks on this document and flag any concerns. Round 2: Present flagged items for human review and incorporate feedback into final approval decision.
Batch content audit 
Review this batch of documents for our RAG system. Apply the clarity-gate verification framework to identify which documents meet quality standards and which need revision or rejection.

Best Practices

  • Define specific quality criteria for your domain before implementing verification
  • Use the two-round HITL workflow for critical content where accuracy is essential
  • Document rejection reasons to improve content quality over time
  • Calibrate verification strictness based on your use case and risk tolerance

Avoid

  • Do not use this skill as a substitute for proper source evaluation and fact-checking
  • Avoid treating all content types with the same quality standards - adjust criteria per domain
  • Do not skip human review for high-stakes content like medical or legal information
  • Never ingest content without verification when building systems that influence decisions

Frequently Asked Questions

What is the 9-point verification framework?
The framework evaluates content across nine dimensions: factual accuracy, source credibility, completeness, logical consistency, temporal relevance, technical depth, audience appropriateness, citation quality, and redundancy. This ensures comprehensive quality assessment before content enters your RAG system.
How does the two-round HITL workflow function?
The first round performs automated quality checks and flags potential issues. The second round presents flagged items for human review, allowing experts to make final approval decisions. This combines AI efficiency with human judgment for critical content validation.
Can I customize the verification criteria for my domain?
Yes, the skill provides a framework that you can adapt. Adjust the nine verification points based on your specific requirements, industry standards, and risk tolerance. For example, medical content may prioritize source credibility more heavily than technical documentation.
Does this skill automatically filter content?
No, this is a documentation skill that provides guidance and patterns. You must implement the actual verification logic in your RAG pipeline. The skill helps you design effective quality gates and verification workflows.
When should I use strict vs. relaxed verification?
Use strict verification with full HITL review for high-stakes content like medical, legal, or safety-critical information. Use relaxed verification for low-risk content like internal notes or draft materials where minor inaccuracies are acceptable.
How do I integrate this into my existing RAG pipeline?
Implement verification as a pre-processing step before document chunking and embedding. Use Claude to evaluate each document against your quality criteria, and route rejected or flagged content to human review queues. Only ingest content that passes verification.

Developer Details

Author

sickn33

License

MIT

Repository

https://github.com/sickn33/antigravity-awesome-skills/tree/main/web-app/public/skills/clarity-gate

Ref

main

File structure

📄 SKILL.md