hashgraph-online/awesome-codex-plugins / plugins/avalonreset/seo-dungeon/skills/seo-hreflang
seo-hreflang
VerifiedHreflang and international SEO audit, validation, and generation. Detects common mistakes, validates language/region codes, and generates correct hreflang implementations. Use when user says "hreflang", "i18n SEO", "international SEO", "multi-language", "multi-region", or "language tags".
| File | Kind | Size |
|---|---|---|
| LICENSE.txt | markdown | 143 B |
| SKILL.md | skill md | 12.4 KB |
| references/content-parity.md | markdown | 4.0 KB |
| references/cultural-profiles.md | markdown | 4.8 KB |
| references/locale-formats.md | markdown | 3.4 KB |
| references/machine-translation-qa.md | markdown | 3.3 KB |
references/machine-translation-qa.md
View on GitHub# Machine-translation QA flag (September 11, 2025 QRG)
Google's January 23, 2025 Quality Rater Guidelines update added
explicit language about **machine-translated content with no human
review** as a form of scaled content abuse (§4.6.5). This is also the
**enforceable webmaster-facing spam policy** (updated 2026-05-15), which
names "automated transformations like synonymizing, **translating**, or
other obfuscation techniques" as scaled content abuse:
https://developers.google.com/search/docs/essentials/spam-policies
> "Using automated tools (generative AI or otherwise) as a low-effort
> way to produce many pages that add little-to-no value."
Machine-translated content is fine **when reviewed by a human
speaker and corrected**. Untranslated MT - or "lightly post-edited"
output that still contains hallucinated terms, wrong gender/number
agreement, or untranslated proper nouns - is treated as scaled
content abuse.
## Signals the seo-hreflang audit should surface
| Signal | Severity | Notes |
|---|---|---|
| Multiple `hreflang` alternates point to URLs whose content is identical except for header chrome | Critical | Indicates the body wasn't translated; just template wrapped. |
| `lang="xx"` attribute on `<html>` doesn't match the body language | High | Translation pipeline output without a final QA step. |
| Auto-translated `<meta name="description">` longer than 160 chars (untrimmed) | Medium | The translator overran the snippet limit - no human reviewer caught it. |
| `lang` attribute is `auto` or missing entirely | Medium | Pages that don't declare their language confuse hreflang + AI crawlers. |
| Untranslated proper nouns or product names sprinkled in body | Low (heuristic) | Common MT failure mode; hard to detect automatically. |
| Schema.org `inLanguage` field absent or wrong | Medium | Multi-language audits should cross-check `inLanguage` vs. body. |
## What the audit should NOT flag
- Sites that have a few MT pages clearly labelled as MT (a
human-translation-fallback pattern). Google's QRG explicitly
permits MT *when honestly labelled and clearly scoped*.
- Machine-translated UI strings - those are i18n, not "content".
- Content with `lang="auto"` if the audit can't fetch a fallback
signal (be conservative; don't claim what we can't verify).
## Cross-skill delegation
- For per-page hreflang validation, stay inside `seo-hreflang`.
- For broader scaled-content scoring (entropy of translated pages,
AI-pattern detection in body), defer to `seo-content` via
`claude-seo run content_quality.py`.
- For "is this translated by Google's own auto-translate widget"
detection, look for the `.goog-te-banner-frame` iframe. There is **no
per-mechanism exemption**: in June 2025 Google removed the
robots.txt-blocking guidance for auto-translated pages, and auto/AI-
translated content is now judged **case-by-case on user value**. Use
page-level `noindex` for low-value translated pages, not sitewide blocks;
the widget also produces poor passage-citability.
## Primary sources
- Current QRG (September 11, 2025; 182 pages; nothing newer through 2026-07-09): https://services.google.com/fh/files/misc/hsw-sqrg.pdf
- John Mueller on MT (multiple SOTR episodes, 2024-2025): MT is OK
when reviewed by a human; bulk MT without review is abuse.
Last verified: 2026-07-09.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |