[geo-optimizer] GEO Audit Report — 2026-08-03

[github-actions[bot]]

GEO Audit Report — github/gh-aw

Audit Date: 2026-08-03
Run: §30835042672

---

📊 Scores

Target	Score	Band
Docs site (github.github.com/gh-aw/)	44/100	Foundation
README (github.com/github/gh-aw)	55/100	Foundation
Docs sitemap average (20/215 pages sampled)	37.75/100	Foundation (14 pages), Critical (6 pages)

---

✅ Top Strengths

• Meta tags: Docs site scores 14/14 — complete title, description, canonical, and Open Graph tags.
• Schema JSON-LD: Docs homepage has WebSite, Organization, SoftwareApplication, and FAQPage schema (13/16).
• Signals: Docs site has lang="en", RSS feed, and a freshness date (6/6).
• Brand & entity: Docs homepage links to Wikipedia, Wikidata, LinkedIn, and Crunchbase (all 4 Knowledge Graph pillars present).
• Platform citability: Google AI scores 80/100 and Perplexity 65/100 on the docs homepage, driven by strong meta tags, freshness, and SSR-accessible content (no JS-rendering dependency; framework: astro).
• README: Has both robots.txt (15/18) and llms.txt (14/18) already configured — most repos lack these entirely.

---

🚨 Critical Gaps

1. No robots.txt on the docs site (0/18) — AI crawlers (GPTBot, ClaudeBot, PerplexityBot) have no explicit allow rules.
2. No /llms.txt on the docs site (0/18) — missing the emerging standard for LLM-friendly site summaries.
3. No AI Discovery files anywhere (0/6 on both docs and README) — missing /.well-known/ai.txt, /ai/summary.json, /ai/faq.json, /ai/service.json.
4. README has zero Schema JSON-LD (0/16) — no WebSite/Organization/FAQPage structured data on the GitHub repo homepage.
5. Hidden text detected on the docs site (a "CtrlK" search-shortcut label) — flagged as a cloaking-risk pattern that AI crawlers may penalize.
6. 6 of 20 sampled docs/blog pages fall into "Critical" band (score 33–35), mostly older blog posts with thin schema (3/16) and weak brand/entity scores (2–3/10).

---

🔧 Recommended Fixes (prioritized by impact)

1. Add robots.txt with explicit Allow rules for GPTBot, ClaudeBot, and PerplexityBot on github.github.com/gh-aw/ — this alone recovers up to 18 points on every audited page (largest single lever; currently 0 across all 20 sampled pages).
2. Generate /llms.txt for the docs site using geo llms --base-url https://github.github.com/gh-aw — worth up to 18 points and directly requested by the audit tool.
3. Add Schema JSON-LD (WebSite + Organization + FAQPage) to the README's rendered page — currently 0/16, the single lowest-scoring category on the README.
4. Create AI discovery files (/.well-known/ai.txt, /ai/summary.json, /ai/faq.json, /ai/service.json) — 0/6 on both properties; low effort, standardized format.
5. Remove or properly hide the "CtrlK" hidden text (search shortcut label) to eliminate the cloaking-pattern penalty (−3 negative_penalty on docs site).
6. Front-load key information in the first 30% of docs homepage content — has_front_loading: false, and front_loaded_ratio is only 0.33; reordering intro paragraphs improves AI snippet selection.
7. Add numerical data/statistics to docs content — numbers_count: 0 on the homepage; audit estimates +40% AI visibility from this change.
8. Fix broken links on the README — 9 broken links detected, a negative signal affecting trust.

📋 Full Breakdown by Category — Docs Site Homepage (score 44/100)

Category	Score	Max
Robots.txt	0	18
llms.txt	0	18
Schema JSON-LD	13	16
Meta Tags	14	14
Content	7	12
Signals	6	6
AI Discovery	0	6
Brand & Entity	7	10
Negative penalty	−3	—

Trust Stack: Technical 3/5, Identity 3/5, Social 4/5, Academic 2/5, Consistency 4/5 → Composite 16, Grade C, medium trust level.

Platform citation scores: ChatGPT 50, Perplexity 65, Google AI 80 (all citability 78/100 underlying).

Content decay: evergreen_score 100, decay risk low.

📋 Full Breakdown by Category — README (score 55/100)

Category	Score	Max
Robots.txt	15	18
llms.txt	14	18
Schema JSON-LD	0	16
Meta Tags	11	14
Content	12	12
Signals	3	6
AI Discovery	0	6
Brand & Entity	3	10
Negative penalty	−3	—

Negative signals: hidden text, popup signals detected, 9 broken links, boilerplate ratio 0.19.

📄 Sitemap Page Scores (20 of 215 discovered URLs sampled)

Average score: 37.75/100 (14 Foundation, 6 Critical). Average breakdown: robots 0, llms 0, schema 5.5/16, meta 14/14, content 10.75/12, signals 5.35/6, ai_discovery 0/6, brand_entity 3.1/10.

Top pages:

URL	Score	Band
/gh-aw (homepage)	44	foundation
/blog/2026-01-13-meet-the-workflows-continuous-improvement	44	foundation
/blog/2026-01-13-meet-the-workflows-continuous-refactoring	43	foundation
/blog/2026-01-12-welcome-to-pelis-agent-factory	41	foundation
/blog/2026-01-13-meet-the-workflows-advanced-analytics	41	foundation

Worst pages:

URL	Score	Band
/blog/11	33	critical
/blog/9	34	critical
/blog/8	35	critical
/blog/7	35	critical
/blog/4	35	critical

⚠️ Sitemap contains 215 URLs but only 20 were audited (--max-urls cap). Increase the cap for full-site coverage in future runs.

---

Automated audit powered by geo-optimizer-skill · Run logs

Generated by 🌍 GEO Optimizer Daily Audit · auto · 23.4 AIC · ⌖ 5.37 AIC · ⊞ 7.2K · ◷

• expires on Aug 6, 2026, 9:10 AM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-08-06T17:10:25.727Z.

Closed by Workflow