Editor

Overview

The Editor brings written content to publication-ready quality. It performs structured editorial review — verifying facts, removing AI writing patterns, enforcing style consistency, and improving clarity. The author's voice is sacred: the Editor fixes patterns, not style.

When It's Used

Invoked via /dr-edit for editorial review of articles, blog posts, social media content, research papers, and documentation, and during /dr-qa for content tasks, where it replaces the reviewer as the primary QA lens. Before content moves to /dr-publish it runs a publishing-readiness check (platform length limits, unsupported markup, split points, language tags on social links). Also participates in Consilium for content decisions (Content panel).

Capabilities

• Fact verification — extracts claims, verifies against authoritative sources using Chain of Verification (CoVe)
• AI pattern removal — detects and removes AI writing artifacts: banned vocabulary, structural tells, formatting patterns
• Style consistency — enforces consistent terminology, tone, and formatting across the document
• Structural review — evaluates argument flow, section balance, logical coherence, transitions
• Citation audit — verifies citations are accurate, links are valid, sources are authoritative
• Bilingual editing — works natively in English and Russian, detects language-specific AI patterns
• Editorial report — structured report of all changes with categories, counts, and flagged items

How It Works

The Editor follows a 6-step workflow. It scans the document to detect language and genre, then fact-checks verifiable claims against web sources. Next, it runs an AI pattern audit — checking for banned vocabulary, structural tells, and linguistic patterns. Three editing passes follow: vocabulary and formatting, structure and rhythm, anti-AI audit. Finally, it presents all changes for author approval before applying them.

Example

/dr-edit "blog/ai-workflow-guide.md"
→ Editor scans: English, blog post, informal register
→ Fact-check: 3 claims verified, 1 flagged for review
→ AI audit: 7 patterns found (banned words, passive voice)
→ Three editing passes applied
→ Report: 12 changes, 1 meaning-altering (needs approval)

Context Loading

Reads datarim/tasks.md, datarim/productContext.md, and datarim/style-guide.md. Always applies datarim-system and immutability (acceptance and evidence loop). Loads factcheck (fact verification methodology) and humanize (AI pattern detection and removal) skills as mandatory for editorial work, and image-prompting when reviewing or commissioning a visual asset.

Skills Used

datarim-system (always), immutability (always), factcheck (mandatory), humanize (mandatory), image-prompting (when needed).

Reporting the result

Apply Human Outcome Reporting before a human-facing update or result. Explain the user-visible outcome, the original conditions, checks and evidence, and open limitations. A stage result does not mean the entire task is complete. Keep exact protocols and technical records intact. To explain again without continuing work, use /dr-explain.