references/REFERENCE.md
GitHubseo-audit-full — Reference Guide
Detailed field definitions, audit modules, scope boundaries, and agent instructions
for the seo-audit-full advanced SEO audit skill.
Positioning
seo-audit-full is the advanced tier — it runs ALL checks from seo-audit (basic)
first, then adds extra modules. Both target a single page.
Compared to seo-audit (basic):
| Dimension | seo-audit (Basic) | seo-audit-full (Advanced) |
|---|---|---|
| Scope | Core signals | Basic + extra modules |
| Technical SEO | robots.txt, sitemap, 404, canonical, HTTPS, hreflang | ★ Same (inherited) |
| On-page | H1, title, meta desc, canonical, slug, images, word count | ★ Same (inherited) |
| Schema | JSON-LD extraction + field validation | ★ Same (inherited) |
| E-E-A-T pages | Trust page existence + reachability | ★ Same (inherited) |
| Social Tags | Not included (banned) | ★ OG Tags + Twitter Card |
| E-E-A-T Content | Not included | ★ Experience/Expertise/Authority/Trust scoring (LLM) |
| Duplicate Content | Not included | ★ Near-duplicate signals (LLM) |
| Anchor Text | Not included | ★ Internal link anchor text quality (LLM) |
| Scripts | 4 Python scripts | 4 inherited + 1 own (check-social.py) |
| Report depth | Summary + Priority Actions + Insights | Full findings with effort/impact priority matrix |
Audit Scope — Full Report Modules
Module 1: Technical SEO (inherited from Basic scripts)
| Check | Script | Notes |
|---|---|---|
| robots.txt | check-site.py |
RFC 9309 group parsing, Allow directive support |
| sitemap.xml | check-site.py |
Tracks robots.txt Sitemap directives |
| 404 Handling | check-site.py |
Soft 404 detection |
| URL Canonicalization | check-site.py |
HTTP→HTTPS, www, trailing slash, canonical match |
| i18n / hreflang | LLM check on HTML | BCP 47 codes, reciprocal symmetry, x-default |
Module 2: On-Page SEO (inherited from Basic scripts)
| Check | Script | Notes |
|---|---|---|
| Title Tag | check-page.py |
Length, keyword presence/position |
| Meta Description | check-page.py |
Length, keyword, quality |
| H1 Tag | check-page.py |
Uniqueness, keyword match |
| URL Slug | check-page.py |
Keyword presence, readability |
| Canonical Tag | check-page.py |
Self-referencing validation |
| Image Alt Text | LLM check on HTML | Parse <img> tags from static HTML |
| Word Count | LLM check on HTML | < 100 fail, 100–499 warn, ≥ 500 pass |
| Keyword Placement | LLM check on HTML | Present in first 100 body words |
| Heading Structure | LLM check on HTML | H2 count, keyword in H2, H3/H2 ratio |
| Internal Links | LLM check on HTML | Same-origin <a> count (excl. nav/footer) |
Module 3: Structured Data (inherited from Basic scripts)
| Check | Script | Notes |
|---|---|---|
| Schema (JSON-LD) | check-schema.py |
@type detection, required/recommended field validation |
Module 4: Social Tags (★ Full-only — scripted)
| Check | Script | Notes |
|---|---|---|
| OG Tags | check-social.py |
og:title, og:description, og:image, og:type, og:url |
| Twitter Card | check-social.py |
twitter:card type, title/desc/image with OG fallback |
OG Tags status logic:
- Pass: og:title + og:description + og:image + og:type all present and valid
- Warn: og:url missing or og:url/canonical mismatch, or length exceeds limits
- Fail: og:title or og:image completely missing
Twitter Card status logic:
- Pass: twitter:card present with valid type, title/desc/image present or OG fallback
- Warn: missing optional fields with no OG fallback
- Fail: twitter:card tag completely missing
Module 5: E-E-A-T Trust Pages (inherited from Basic workflow)
| Page | Required |
|---|---|
| About Us | Yes |
| Contact | Yes |
| Privacy Policy | Yes |
| Terms of Service | Yes |
| Media / Partners | No — include only if present |
Module 6: LLM-Only Advanced Checks (★ Full-only)
These checks require semantic judgment and cannot be scripted:
| Check | What to assess |
|---|---|
| E-E-A-T Content Quality | Experience signals, expertise depth, authority indicators, trust markers |
| Duplicate Content Signals | Near-duplicate paragraphs, boilerplate ratio, unique content percentage |
| Anchor Text Quality | Are internal link anchors descriptive and keyword-relevant? |
Agent Instructions
General quality rules
- Concrete over abstract. "og:title is 112 chars, exceeding the 95-char limit" > "og:title is too long."
- Proportional depth. More detail for high-impact issues.
- No false certainty. Mark assumptions with
[ASSUMPTION]or[UNVERIFIED]. - Priority matrix. Include effort/impact tags:
Low Effort / High Impact, etc.
Handling missing data
If API keys are not available:
"GSC data is not available — search performance analysis is not included in this audit. To enable: set
GSC_API_KEYenvironment variable."
If CWV data is not supplied:
"Core Web Vitals data is not available — this report does not include field performance measurements."
Finding Format Reminder
**Finding: [Title]**
- **Evidence:** [Observable fact, data point, or marked assumption]
- **Impact:** [SEO / UX consequence]
- **Fix:** [Actionable recommendation with example]
For Priority Actions:
1. [High Impact / Low Effort] Fix og:image — social shares currently show no preview.
Limitations Disclosure
Always include a limitations section:
This audit is based on publicly accessible page signals at the time of analysis. Depending on data availability, the following may not be included: source code review, JavaScript rendering analysis, Core Web Vitals field measurements, Google Search Console data, crawl log analysis, or competitive benchmarking. All findings marked [UNVERIFIED] or [ASSUMPTION] indicate areas where additional data collection is recommended.