Skip to content

indranilbanerjee/digital-marketing-pro / skills/content-engine

content-engine

Verified

Draft content in brand voice through a gated pipeline: blogs, ads, emails. Brief only → content-brief. "write a blog post about X"

FileKindSize
SKILL.mdskill md35.1 KB
accessibility.mdmarkdown3.5 KB
ad-copy.mdmarkdown4.5 KB
ai-content-quality.mdmarkdown4.9 KB
brand-voice.mdmarkdown4.5 KB
case-studies.mdmarkdown16.9 KB
content-calendar.mdmarkdown4.5 KB
content-decay.mdmarkdown4.5 KB
draft-deliverables.mdmarkdown5.1 KB
email-automation.mdmarkdown4.3 KB
email-infrastructure.mdmarkdown5.5 KB
email-sequences.mdmarkdown5.5 KB
evals/evals.jsonconfig2.6 KB
landing-pages.mdmarkdown4.5 KB
multilingual.mdmarkdown4.7 KB
personalization.mdmarkdown20.0 KB
seo-content.mdmarkdown5.5 KB
social-content.mdmarkdown5.0 KB
video-scripting.mdmarkdown16.5 KB

ai-content-quality.md

View on GitHub
# AI Content Quality — Management & Review Framework

## AI Content Quality Checklist

Before publishing any AI-assisted content, verify:

### Factual Accuracy
- [ ] All statistics are verified against original sources
- [ ] All dates, names, and company references are accurate
- [ ] No hallucinated facts, studies, or quotes
- [ ] All URLs and links are real and working
- [ ] Product features/pricing are current and correct

### Originality & Value
- [ ] Content adds genuine insight beyond what's already published
- [ ] Unique perspective, framework, or data is included
- [ ] Not a generic rehash of top search results
- [ ] Contains at least one element readers can't get elsewhere
- [ ] Voice and perspective feel authentic, not generic

### Brand Voice
- [ ] Matches brand voice dimensions (formality, energy, humor, authority)
- [ ] Uses preferred vocabulary, avoids restricted words
- [ ] Personality traits are evident in the writing
- [ ] Tone is appropriate for the target channel

### Readability & Structure
- [ ] Appropriate grade level for target audience
- [ ] Paragraphs are concise (2-4 sentences)
- [ ] Logical heading hierarchy
- [ ] Lists and tables used where appropriate
- [ ] No repetitive sentence structures
- [ ] Transitions flow naturally between sections

---

## Common AI Content Pitfalls

| Pitfall | How to Spot It | How to Fix It |
|---------|---------------|--------------|
| **Generic language** | Phrases like "in today's fast-paced world" | Replace with specific, concrete statements |
| **Hallucinated facts** | Statistics without verifiable sources | Fact-check every number and claim |
| **Lack of POV** | Content reads as neutral summary | Add expert opinion, take a position |
| **Repetitive structure** | Every section follows same pattern | Vary formats (list, narrative, table, example) |
| **Over-hedging** | "May," "might," "could potentially" everywhere | Make definitive statements where supported |
| **Missing specificity** | "Many companies" instead of "67% of B2B firms" | Replace vague claims with data |
| **Filler paragraphs** | Sections that don't add new information | Delete or replace with substantive content |
| **Inconsistent depth** | Some sections detailed, others superficial | Balance depth across all sections |

---

## Human Review Workflow

### AI Handles (Draft Phase)
- Initial research and outline
- First draft of body content
- Generating content variations (headlines, CTAs)
- Formatting and structure
- SEO optimization (keyword placement, meta descriptions)

### Human Reviews (Edit Phase)
- Factual accuracy verification
- Brand voice alignment
- Unique insight and perspective injection
- Tone and sensitivity review (cultural, legal, brand safety)
- Final approval for publication

### Review Roles
| Role | Responsibility | Time Required |
|------|---------------|-------------|
| **Subject matter expert** | Verify accuracy, add expertise | 15-30 min per piece |
| **Brand voice reviewer** | Check voice consistency, tone | 10-15 min per piece |
| **Editor** | Readability, structure, flow | 20-30 min per piece |
| **Compliance reviewer** | Legal claims, disclosures, regulations | 10-20 min (if applicable) |

---

## AI Content Disclosure

### When to Disclose
- **Required**: When regulations mandate disclosure (evolving — check jurisdiction)
- **Required (EU)**: EU AI Act Article 50 has applied since 2 August 2026; generative systems that existed before then have until 2 December 2026 to mark their output under Article 50(2) — AI-generated content targeting EU audiences needs machine-readable + visible disclosure. The `eu_disclosure_if_ai` gate enforces this at publish time; use `/digital-marketing-pro:c2pa-metadata` to embed the machine-readable provenance manifest
- **Recommended**: When content is substantially AI-generated
- **Not needed**: When AI is used as a writing tool but human significantly edits

### Disclosure Language Options
- "This article was written with AI assistance and reviewed by [Expert Name]"
- "Research and initial drafting assisted by AI. Verified and edited by our editorial team"
- Keep disclosure simple, honest, and non-defensive

---

## Quality Scoring for AI Outputs

| Dimension | Weight | 0 (Fail) | 5 (Adequate) | 10 (Excellent) |
|-----------|--------|----------|-------------|----------------|
| Factual accuracy | 25% | Contains hallucinations | All facts correct | Adds verifiable data beyond prompt |
| Originality | 20% | Generic rehash | Some unique angles | Genuinely novel insights |
| Brand voice match | 20% | Off-brand | Mostly aligned | Unmistakably on-brand |
| Readability | 15% | Wrong level for audience | Appropriate level | Perfectly calibrated |
| Actionability | 10% | Theoretical only | Some actionable advice | Specific, implementable guidance |
| Structure | 10% | Poor organization | Well-organized | Exceptional flow and formatting |

**Score 80+**: Publish with minor edits. **60-79**: Significant human editing needed. **Below 60**: Regenerate or rewrite.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall