dineway/dineway-skills

dineway-content-optimize

Optimize one exact Dineway Draft, then produce the canonical native content-only QA report used by Quality Attestation.

First seen Aug 13, 2026

Installation

$ npx skills add dineway/dineway-skills --skill dineway-content-optimize

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from dineway/dineway-skills · top by installs.

npx skills add dineway/dineway-skills

Browse all from dineway/dineway-skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 2
License LICENSE
Default branch main
Open issues 0
Status Active

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 4,544 B
  • docs SUMMARY.md 151 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 4 installs

SKILL.md

Dineway Content Optimize

Run one consolidated SEO/GEO optimization pass against the exact current CMS Draft, then perform one final deterministic content QA and rescore. Use dineway-content-geo-optimize only when the Run explicitly enables deep GEO.

Inputs

  • Current Pipeline Run, Optimization Job, active Assignment/Attempt, and accepted input Result

versions.

  • Exact collection, content ID, Draft Revision ID, revision token, schema fingerprint, content

fingerprint, and full Draft content.

  • Accepted Research, Brief, and Writer Results; current SERP/competitor evidence; Site Context;

selected byline; media; persisted Observations; and active quality/rule policy.

  • Read references/optimization-methodology.md, ../dineway-seo/references/apply-loop.md, and the

Pipeline workflow/artifact contracts.

Stop on stale Assignment, changed Research/Brief selection, schema drift, or Draft drift.

Consolidated audit and optional remediation

  1. Establish one fixed evidence baseline from the approved Brief, current SERP evidence, schema,

links, media, Site Context, and byline samples.

  1. Produce evidence-linked dimensions and suggestions with stable IDs, type, title, description,

optional proposed content, auto-applyability, priority, Observation IDs, decision, and dismissal reason.

  1. Prioritize missing topics, unsupported/thin sections, heading hierarchy, metadata, verified links,

and voice. Reject keyword stuffing, competitor imitation, unsupported expansion, and voice damage.

  1. Audit only by default. When the current Draft passes mandatory checks, do not submit content and

do not write CMS.

  1. When blocking findings require changes, produce one consolidated remediationSource using the

Writer source contract. The backend may create exactly one final Draft Revision and records its lineage from the Writer source Draft. Do not patch cards or accessibility fields one by one.

Score contract

  • seoScore measures SERP-relative topical coverage and on-page search readiness.
  • geoScore measures direct-answer potential, citation worthiness, structured information, and

observed question coverage at normal Optimize depth.

  • score is round((seoScore + geoScore) / 2) when both are measured, the measured side when only

one is available, and null when neither is available.

  • scoreCoverage is the measured component count divided by two.
  • Preserve unavailable values as null, never zero.

Canonical content QA

Build exactly one contentQa report for the final exact Draft through CLI, MCP, and native interfaces. It contains checked time, collection, content ID, Draft Revision ID, schema fingerprint, issues, and every required check:

  • exact Draft identity and unique content identity;
  • schema and Research/Brief coverage;
  • persisted evidence lineage;
  • heading/section structure;
  • SEO metadata and canonical/robots data;
  • internal links and media references;
  • accessibility metadata and structured data;
  • unresolved rules; and
  • final scores.

Each check records passed, an actionable message only when failed, and persisted Observation IDs. The report is the only QA authority consumed by Optimization and derived Quality Attestation. Do not submit separate mandatoryGates or an artifactRef inside the payload.

Do not open a browser, take screenshots, inspect lazy loading, test responsive/viewport layout, evaluate visual rendering, or test interactive behavior. There is no browser fallback in this article-content workflow.

Stage completion

Return the compact assessment: source Draft ID, dimensions, suggestions/decisions, and the exact-Draft content-QA checks. Include remediationSource only for the one blocking remediation path. The server calculates scores, compiles the canonical Optimization Result/report, derives the final fingerprint, and records explicit source/final Draft lineage.

Return to dineway-content-pipeline, which refreshes the exact CMS Draft and calls contentpipelinestage_complete. That transaction validates the report identity and lineage, derives the artifact receipt and atomically creates/accepts Optimization plus Quality Attestation when this is the final quality stage. It otherwise routes to optional deep GEO. Do not call granular Result operations, attest quality separately, approve, schedule, or publish here.