Skip to content

RyvrImmersive/content-eval

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

content-eval — Thought Leadership Quality Scorer

skills.sh

A Claude Code skill that evaluates thought leadership content and returns a structured quality report: a weighted scorecard, evidence-backed dimension analysis, and a prioritised improvement plan.

Works on whitepapers, case studies, blog posts, articles, op-eds, and long-form reports.


What It Does

Scores content across 8 dimensions on a 0–10 scale, produces a weighted overall score out of 100, surfaces specific weaknesses with verbatim quotes from the text, and delivers ranked, actionable recommendations.

Dimension Weight What It Measures
Clarity 15% Is the writing easy to follow? Are sentences crisp?
Credibility 15% Are claims backed by data, citations, or named sources?
Originality 12% Fresh angle, proprietary insight, or contrarian take?
Structure 12% Logical flow, effective headings, good pacing?
Audience Alignment 15% Right tone, vocabulary, and framing for the intended reader?
Executive Value 13% Would a busy senior leader find this worth their time?
Evidence Quality 10% How strong and specific is the supporting evidence?
CTA Effectiveness 8% Clear, motivated next step that feels earned?

Grades: A (85–100) · B (70–84) · C (55–69) · D (40–54) · F (<40)


Installation

npx skills add ryvrimmersive/content-eval -g

Or install from the Skills registry:

npx skills find content-eval

Usage

Once installed, invoke it inside Claude Code:

/content-eval <paste text | file path | URL>

Examples:

# Score a pasted article
/content-eval [paste your article here]

# Score a local file
/content-eval /path/to/whitepaper.pdf

# Score a live URL
/content-eval https://yourcompany.com/blog/post-title

Accepts .txt, .md, .pdf, .docx, and any fetchable URL.


Sample Output

#### Scorecard

| Dimension          | Score /10 | Weight | Weighted |
|--------------------|-----------|--------|----------|
| Clarity            | 8         | 15%    | 1.20     |
| Credibility        | 8         | 15%    | 1.20     |
| Originality        | 9         | 12%    | 1.08     |
| Structure          | 7         | 12%    | 0.84     |
| Audience Alignment | 9         | 15%    | 1.35     |
| Executive Value    | 8         | 13%    | 1.04     |
| Evidence Quality   | 7         | 10%    | 0.70     |
| CTA Effectiveness  | 3         | 8%     | 0.24     |
| **OVERALL**        | —         | 100%   | **77/100** |

**Grade:** B
**One-line verdict:** Sharp contrarian argument and strong audience calibration —
held back by an absent CTA and a slightly under-resolved structure.

Followed by:

  • Dimension-by-dimension analysis with verbatim quotes as evidence
  • 4–7 priority-ordered recommendations with specific rewrites
  • Quick wins (word/sentence-level fixes)

Content Type Adjustments

The skill applies different emphasis depending on content type:

  • Whitepaper — Credibility and Evidence Quality weighted higher in commentary
  • Case Study — Audience Alignment and Evidence Quality lead; specificity of metrics matters most
  • Blog Post — Clarity and Originality lead; CTA should feel natural
  • Op-Ed / POV — Originality and Credibility dominate; author voice must be unmistakable
  • Article (journalism) — Structure and Evidence Quality are paramount

Calibration Philosophy

  • A 10 is rare — anchored to Fortune 500 CMO whitepapers and major analyst reports
  • A 7 means genuinely good
  • A 5 means mediocre but functional
  • A 3 means significant problems
  • A 1 means nearly absent

Scores are intentionally honest. The tool is designed to surface real gaps, not to validate.


Files

content-eval/
├── SKILL.md                      # Skill definition and scoring rubrics
├── references/
│   └── scoring-examples.md       # Annotated calibration examples per dimension
└── evals/
    └── evals.json                 # 3 test cases (weak blog, strong case study, CEO op-ed)

Evals

Three test cases are included to verify skill performance:

ID Content Expected Score Range
0 Generic AI blog post (weak) < 65
1 Fintech case study with metrics (strong) 75–90
2 CEO LinkedIn op-ed with Edelman data (strong) 75–88

Run evals with:

npx skills eval content-eval

Author

Built by RYVR Immersive. Feedback and PRs welcome.

About

Analyse B2B thought leadership content on quantitative and qualitative aspects

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors