A free Claude Code skill that scores your content against the same 5-layer framework used to build content infrastructure for founder-led companies.
Your content gets a weighted score out of 100, using the same weights as whystrohm.com/scan. You see exactly what's broken and why. Then one piece gets rewritten live, layer by layer, so you see the difference, not just read about it.
| Phase | What Happens |
|---|---|
| 1. Scrape | Pulls your website to analyze voice, positioning, and content patterns. Also searches public pages for how your buyer talks about the problem |
| 2. Profile | Asks 3 targeted questions about your buyer, differentiator, and content |
| 3. Score | Rates your content against 5 layers (each 1-10), then weights them into a total out of 100. Voice Consistency uses your saved brand/voice-profile.json if you have one |
| 4. Diagnose | Shows exact quotes from your content with specific problems identified |
| 5. Rewrite | Rewrites your lowest-scoring piece live with before/after comparison |
| 6. Reveal | Explains the full 5-layer framework so you understand the system |
| 7. Save | Writes the findings to brand/audit-findings.json in your current folder, in a shared format. storyboard-architect in Shotkit can take it as a brief input |
| Layer | Weight | What It Checks |
|---|---|---|
| Vocabulary | 15% | Hype words, hollow verbs, vendor-centric language, proof density |
| Structure | 20% | Buyer context, problem specificity, stakes, evidence, clear action |
| Proof Density | 25% | Claims-to-evidence ratio. Every assertion needs adjacent proof |
| Voice Consistency | 20% | Does your content match your actual brand, or a generic template? |
| Buyer Alignment | 20% | Written FOR your buyer, or talking about yourself? |
Total = round(sum of layer score x 10 x weight). Each layer point is worth 1.5 to 2.5 points of the 100.
One command:
git clone https://github.com/whystrohm/whystrohm-audit.git ~/.claude/skills/whystrohm-auditOr manually: Download this repo and copy the whystrohm-audit/ folder to ~/.claude/skills/.
- Claude Code (CLI, desktop app, or IDE extension)
- No API keys. No external tools. No configuration.
/whystrohm-audit
You'll be asked for:
- Your website URL
- Your ideal buyer (role, size, industry)
- Your differentiator (what you do differently)
- One piece of published content (LinkedIn post, email, blog paragraph, anything)
The audit runs in one Claude Code session. It uses WebFetch and WebSearch on public pages only. It reads brand/voice-profile.json if one is in your current folder, and writes brand/audit-findings.json there at the end. See SECURITY.md.
Example output. The numbers below are for illustration only. Your score depends on your content.
═══════════════════════════════════════
YOUR CONTENT INFRASTRUCTURE SCORE
═══════════════════════════════════════
52/100
Common for founder-led brands. Good instincts, no system.
Vocabulary: 7/10 ███████░░░ x15%
Structure: 4/10 ████░░░░░░ x20%
Proof Density: 3/10 ███░░░░░░░ x25%
Voice Consistency:7/10 ███████░░░ x20%
Buyer Alignment: 6/10 ██████░░░░ x20%
═══════════════════════════════════════
Followed by:
- Per-layer breakdown with exact quotes from your content
- Your top 3 content gaps ranked by business impact
- A live rewrite of your weakest piece with before/after scores
- The complete 5-layer framework explained
brand/audit-findings.jsonsaved in your current folder
whystrohm-audit/
├── SKILL.md # Orchestrator: controls the two-phase flow
├── CLAUDE.md # Contributor guide for AI agents
├── rules/
│ ├── scoring-rubric.md # The 5-layer framework with scoring criteria
│ ├── voice-analysis.md # How brand voice is inferred from your website
│ └── rewrite-rules.md # Layer-by-layer transformation rules
├── templates/
│ ├── audit-report.md # Output formatting (score display, breakdowns)
│ └── cta.md # Closing pitch (configurable booking link)
├── contracts/
│ ├── audit-findings.v1.schema.json # Format of brand/audit-findings.json
│ ├── voice-profile.v1.schema.json # Format of brand/voice-profile.json
│ └── README.md # Which skill writes and reads each file
├── examples/
│ └── audit-findings.example.json # A filled example of the findings file
├── .github/ # Issue templates, PR template, CODEOWNERS, contracts CI
├── CHANGELOG.md # Version history
├── CONTRIBUTING.md # How to contribute
├── SECURITY.md # Security policy
├── LICENSE # MIT
└── README.md
This skill scores one piece of content.
A full content infrastructure install scores your entire library, maps voice drift across your platforms, writes the guardrails down as checks, and hands you a system that produces brand-consistent content without routing everything through the founder.
What the full system includes:
- Brand voice written as enforceable rules (not a PDF)
- Content guardrails that reject hype and require proof before publish
- Video production pipeline that renders branded content from code
- Scheduled publishing across platforms
- Templates, calendars, and a content system your team owns
See how it is built: whystrohm.com/system
Score your site first. No email needed to see the result:
WhyStrohm builds content infrastructure for founder-led companies. You own what gets built.
| Skill | What It Does | Install |
|---|---|---|
| Ritual · NEW | Scans your machine, ranks the top 5 recurring tasks worth scheduling, and drafts a Claude Code scheduled trigger that runs this audit across your brand repos on a set cadence. | Download .skill ↗ |
| Digital Twin | Reverse-engineer how you think and talk. Stress-tested AI System Prompt of yourself. | git clone https://github.com/whystrohm/digital-twin-of-yourself.git ~/.claude/skills/digital-twin |
| Voice Scorer | Measure voice drift between your website and social content. | git clone https://github.com/whystrohm/whystrohm-voice-scorer.git ~/.claude/skills/whystrohm-voice-scorer |
This skill is one piece of the brand infrastructure I build for founder-led brands: voice extraction, programmatic video, and scheduled publishing. You own everything.
MIT. See LICENSE.