content-audit

name	content-audit
description	Content quality audit with anti-slop detection and brand consistency scoring. Use when user wants to check content quality, audit a draft, score content, verify brand consistency, detect AI-generated patterns, or run a quality check before publishing.

Content Quality Audit

Overview

This skill audits content across three dimensions, each scored out of 100:

1. Content Quality -- Anti-slop score, readability, hook strength, structural variety
2. Brand Voice Consistency -- How well the content matches the voice profile in brand/VOICE.md
3. Brand Visual Consistency -- For visual content (presentations, videos, infographics), how well it matches brand/VISUAL.md

The composite score combines all applicable dimensions. For text-only content, the composite is based on dimensions 1 and 2. For visual content, all three dimensions contribute.

---

Modes

Quick Score (/content score)

Runs a fast assessment and returns:

• Composite score (out of 100)
• Top 3 issues to fix
• One-sentence summary

Use this for a quick pulse check on a draft.

Full Audit (/content audit)

Runs a comprehensive review and returns:

• Detailed scores for each dimension with subscores
• Line-level feedback with specific callouts
• Forbidden phrase detection
• Readability metrics
• Concrete fix suggestions for every issue found
• Before/after examples for the most critical issues

Use this before publishing.

---

Before You Start

1. Load brand references: Read brand/VOICE.md and brand/VISUAL.md if they exist. These define the scoring criteria for brand consistency. If they do not exist, score brand consistency against reasonable defaults and note that brand files should be created for more accurate scoring.

2. Identify the content type: Blog post, thread, LinkedIn post, newsletter, video script, slide deck, etc. Different content types have different quality expectations.

3. Get the content: The user should provide the draft text or a path to the file containing the content.

---

Audit Flow

Stage 1: Content Quality Audit

Run the content through these checks. Reference slop-detector.md for the full pattern list and scoring.md for the scoring rubric.

Anti-Slop Detection

Scan the content for AI-generated patterns grouped by severity:

Critical (major deductions):

• Opening cliches ("In today's fast-paced world," "Let's dive in," etc.)
• Closing cliches ("In conclusion," "To sum up," "At the end of the day")
• Buzzwords used as substance substitutes ("leverage," "unlock," "game-changer," "paradigm shift")
• Filler phrases that add no meaning ("It's worth noting that," "It goes without saying," "Needless to say")
• Em-dash abuse (more than 2 em-dashes per 500 words)

Warning (moderate deductions):

• Overused transitions ("Furthermore," "Moreover," "Additionally" appearing more than once)
• Generic metaphors ("Think of it as..." "It's like..." without a genuinely illuminating comparison)
• Vocabulary tells ("delve," "tapestry," "landscape," "myriad," "plethora")
• Structural tells (all paragraphs within 10% of the same length)

Info (minor deductions):

• Passive voice (acceptable at <15% of sentences, flagged above that)
• Predictable cadence (every sentence following the same rhythm)
• Missed opportunities for concrete examples

For each detection, report:

• The flagged text (exact quote with location)
• The severity level
• Why it is a problem
• A specific suggested fix

Readability Analysis

• Sentence length variety: Standard deviation of sentence lengths. Higher variety = better. Flag if >60% of sentences are within 5 words of the average.
• Paragraph length variety: Mix of short (1-2 sentences) and longer (3-5 sentences) paragraphs. Flag if all paragraphs are the same length.
• Flesch-Kincaid Grade Level: Report the grade level. For most content, aim for grades 8-12. Technical content can go higher, but flag anything above grade 16.
• Average words per sentence: Report the average. Flag if >25 (too dense) or <10 (too choppy).

Hook Strength

Evaluate the opening (first 2-3 sentences) on a 1-10 scale:

Score	Criteria
9-10	Specific, surprising, creates genuine curiosity or tension. Reader must continue.
7-8	Good hook with a clear promise of value. Could be sharper but works.
5-6	Functional but generic. Gets to the topic but does not create urgency.
3-4	Weak. Generic opening, throat-clearing, or vague promise.
1-2	Cliche opener, "In this article we will explore...", or no hook at all.

Structure Analysis

• Does the content have clear section headings (for long-form)?
• Is information logically ordered?
• Does the closing land with impact (CTA, insight, question, callback)?
• Are there concrete examples, data, or evidence for claims?

Stage 2: Brand Voice Consistency

Reference brand-consistency.md for the full methodology.

Compare the content against brand/VOICE.md on these dimensions:

Voice Match

• Formality level: Does the content match the brand's formality (casual, conversational, professional, academic)?
• Sentence patterns: Does the sentence rhythm match? (Short and punchy? Long and flowing? Mixed?)
• Vocabulary alignment: Does the content use the brand's preferred vocabulary? Avoid its banned terms?
• Personality markers: Does the content reflect the brand's personality traits (witty, serious, provocative, warm, etc.)?
• Humor style: If the brand uses humor, is it the right kind? If the brand is serious, is inappropriate humor present?

Forbidden Phrase Detection

Check the content against the banned phrase list in brand/VOICE.md. Report every match with:

• The exact phrase found
• Its location in the content
• The suggested replacement from the banned list

Voice Score Components

Component	Weight	Description
Formality match	30%	Alignment with target formality level
Vocabulary alignment	25%	Use of preferred terms, avoidance of banned terms
Sentence rhythm	20%	Match with brand's sentence pattern profile
Personality consistency	15%	Brand personality reflected throughout
Forbidden phrases	10%	Penalty for each banned phrase found

Stage 3: Brand Visual Consistency (Visual Content Only)

For presentations, video compositions, infographics, or any content with visual elements. Reference brand-consistency.md.

Check against brand/VISUAL.md:

• Color palette: Are all colors used from the approved palette? Flag any off-brand colors with their hex values.
• Typography: Are the correct fonts used for headings, body, and mono text? Are font sizes consistent with the brand scale?
• Logo usage: Is the logo present where expected? Is it the correct version (light/dark, full/icon)? Does it have sufficient clear space?
• Spacing and layout: Does the layout follow the brand's grid and spacing system?
• Image style: Do images match the brand's visual style (photography style, illustration style, icon style)?

---

Scoring

Reference scoring.md for the detailed rubric.

Score Breakdown

CONTENT QUALITY AUDIT
=====================

Composite Score: [XX]/100

Dimension 1: Content Quality          [XX]/100
  - Anti-slop score:                   [XX]/100
  - Readability:                       [XX]/100
  - Hook strength:                     [XX]/100
  - Structure:                         [XX]/100

Dimension 2: Brand Voice Consistency   [XX]/100
  - Formality match:                   [XX]/100
  - Vocabulary alignment:              [XX]/100
  - Sentence rhythm:                   [XX]/100
  - Personality consistency:           [XX]/100
  - Forbidden phrases:                 [XX]/100

Dimension 3: Visual Consistency        [XX]/100  (if applicable)
  - Color palette:                     [XX]/100
  - Typography:                        [XX]/100
  - Logo usage:                        [XX]/100
  - Layout/spacing:                    [XX]/100

TOP ISSUES (ranked by impact)
-----------------------------
1. [Issue description with specific example and fix]
2. [Issue description with specific example and fix]
3. [Issue description with specific example and fix]
...

Composite Score Calculation

For text-only content:

Composite = (Content Quality * 0.60) + (Voice Consistency * 0.40)

For visual content:

Composite = (Content Quality * 0.40) + (Voice Consistency * 0.30) + (Visual Consistency * 0.30)

Score Interpretation

Score	Rating	Action
90-100	Excellent	Ready to publish. Minor polish optional.
80-89	Good	Fix the top 1-2 issues, then publish.
70-79	Acceptable	Several issues need attention. Revise and re-audit.
60-69	Needs Work	Significant issues. Major revision required.
Below 60	Poor	Rewrite recommended. Fundamental issues present.

---

Feedback Format

For each issue found, provide:

[SEVERITY: critical/warning/info]
Location: [Paragraph number, sentence, or line reference]
Issue: [What the problem is]
Flagged text: "[exact text from the content]"
Why: [Why this is a problem -- what effect it has on the reader]
Fix: [Specific rewrite or action to take]

For the most critical issues, provide before/after examples:

BEFORE: "In today's rapidly evolving digital landscape, it's worth noting
that leveraging AI can be a real game-changer for content creation."

AFTER: "AI cut our content production time from 8 hours to 2. Here's
exactly how we set it up."

---

Output Checklist

Before delivering the audit, verify:

• All three applicable dimensions are scored
• Anti-slop detection has been run against the full pattern list
• Every flagged issue includes the exact text, location, reason, and fix
• Forbidden phrase detection has been run against brand/VOICE.md banned list
• Hook strength is evaluated on the 1-10 scale with rationale
• Readability metrics are reported (sentence length variety, Flesch-Kincaid, etc.)
• Before/after examples are provided for the top 3 most critical issues
• Composite score is calculated using the correct formula
• Score interpretation is provided with clear next steps
• Feedback is specific and actionable (not "improve the writing")