# Eeat

> E-E-A-T specialist — a first-class UPSTREAM consult lens that scores content for Google's Experience, Expertise, Authoritativeness, and Trust signals BEFORE it's finalized, and returns structured fixes. Use when producing or auditing any content, page, or claim that needs to rank / be cited / be trusted (blog, landing page, YMYL, client deliverable). Consumed by the specialist-council, marketing, seo, and geo-optimization; complements source-ingest (credibility of sources) and brand-guardian (terminal accuracy gate). Triggers on "E-E-A-T", "EEAT", "experience expertise authoritativeness trust", "trust signals", "author authority", "does this look credible/trustworthy".

- Skill: `cleanexpo/eeat` (Agent Skill)
- Install (CLI): `npx skillmds@latest add cleanexpo/eeat`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cleanexpo/eeat/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Marketing & Growth
- Author: cleanexpo (https://skillmd.com/u/cleanexpo)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/cleanexpo/eeat

---


# eeat — Experience · Expertise · Authoritativeness · Trust

Google's quality framework (the extra "E", Experience, added Dec 2022), the lens
that decides whether content deserves to rank and be cited — and the strongest
signal for YMYL (your-money-your-life) topics. Until now E-E-A-T lived only
implicitly in `geo-optimization`'s Trust-Signals subsection and `brand-guardian`'s
accuracy gate. This makes it a **first-class, upstream** consult: it reviews a
draft BEFORE it ships and returns structured fixes, rather than being a
back-end afterthought.

**Advise, don't gate.** E-E-A-T returns must_fix + suggestions; the owner
integrates. `brand-guardian` remains the terminal accuracy gate; this is the
upstream quality lens that feeds it.

## When to use
- Producing/auditing content that must rank, be cited by AI engines, or carry
  authority — blog posts, landing pages, service pages, especially **YMYL**
  (health, finance, legal, safety — RestoreAssist/DR/CARSI are YMYL-adjacent).
- As a `specialist-council` consult on any content artifact (routing table:
  content/campaign/FACT tasks).

## When NOT to use
- Pure keyword/SERP data → `seo`. AI-engine visibility infra → `geo-optimization`.
- Fiction / internal notes / non-published copy.

## The four lenses (score each, cite the gap)
1. **Experience** — is there genuine first-hand experience? Real case studies,
   original photos/data, "we did/tested/measured this," specifics only a
   practitioner would know. Generic AI-sounding overviews score low. (This is the
   signal most often missing in generated content.)
2. **Expertise** — a named, credentialed author/entity? Bio, qualifications,
   relevant track record shown on-page. Anonymous or unattributed = weak.
3. **Authoritativeness** — is the site/author cited by others; does the page cite
   authoritative Tier-1/2 sources (pull via `source-ingest`)? Awards, memberships,
   recognized standards bodies (IICRC, ISO) named.
4. **Trust** — the load-bearing one. Accuracy (every load-bearing claim sourced),
   transparency (contact/about, real business, ABN), honest framing (no overclaim
   — ties to the no-fake-as-real rule), reviews/testimonials, secure/updated site.

## Consult-response contract (feeds specialist-council)
```json
{
  "specialist": "eeat",
  "verdict": "pass | needs-work | fail",
  "scores": {"experience": 0, "expertise": 0, "authoritativeness": 0, "trust": 0},
  "must_fix": [{"lens": "experience", "issue": "no first-hand signal — reads generic", "fix": "add a real case study / original data"}],
  "suggestions": [{"lens": "trust", "change": "cite the IICRC standard for the drying claim", "impact": "high"}],
  "evidence": ["source-ingest Tier-1 for the load-bearing stat"]
}
```

## How to use
1. Read the draft. Score each lens 0–100 with the concrete gap.
2. For YMYL, weight Trust + Expertise hardest; a fail there is a fail overall.
3. Return the structured response; the owner folds `must_fix` in before shipping.
4. Cross-consult: pull authoritative citations via `source-ingest`; hand the
   accuracy check to `brand-guardian` as the final gate.

## Guardrails
- **No fabricated credibility** — never invent authors, awards, or citations to
  score higher; flag the gap instead (the fix is to EARN the signal).
- **Sourced trust only** — every "trust" fix that adds a claim must point to a
  real Tier-1/2 source via `source-ingest`.
- Advisory lens, not the gate — `brand-guardian` decides ship/no-ship on accuracy.

## Anti-duplication
Owns the E-E-A-T rubric. Does NOT do SERP/keyword data (`seo`), AI-engine infra
like llms.txt/JSON-LD (`geo-optimization` — which now defers its Trust-Signals
depth here), source scraping (`source-ingest`), or final accuracy gating
(`brand-guardian`). It is the upstream quality lens the others consult.