draft_authoring

Writes article drafts from intake records and source sets with claim-to-source notes.

4|Updated Jan 31, 2026
One-click install
npx skills add https://github.com/synthaicode/XRefKit --skill draft-authoring-synthaicode
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: draft_authoring
Source: https://github.com/synthaicode/XRefKit/tree/main/skills/packs/editorial-ops/draft_authoring
Command: npx skills add https://github.com/synthaicode/XRefKit --skill draft-authoring-synthaicode

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Drafting articles from scattered sources often hides unsupported claims inside polished prose, making factual gaps invisible to reviewers. This Skill produces a draft where every concrete assertion is tied to a source and every unsupported claim stays explicitly marked. ## Core Features & Use Cases - Structured Drafting: Builds an article spine from the intake goal and audience, then writes the draft in that order. - Claim-to-Source Traceability: Attaches claim-to-source notes for concrete assertions, numbers, names, and URLs. - Visible Uncertainty: Keeps unsupported claims marked instead of converting missing evidence into confident prose. - Use Case: An editorial team has an intake record and a vetted source set for a new article. The Skill produces a date-prefixed draft under work/editorial_ops/ with unresolved questions listed, ready for fact review and reader experience review. ## Quick Start Use the draft_authoring skill to write a first article draft from the current intake record and source set, keeping unsupported claims marked and returning the draft path with unresolved questions.

Frequently Asked Questions about draft_authoring

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write an article draft from a source set without inventing facts?▼

Provide an intake record and a source set, and the Skill builds an article spine from the goal and audience, then writes the draft in that order. Unsupported claims stay explicitly marked instead of being smoothed into confident prose.

What inputs does the draft authoring workflow need before starting?▼

It requires an intake record defining the article boundary, a source set sufficient to support a first draft, and the channel intent. An optional existing draft is accepted when the target is a revision rather than a new article.

Where does the draft output get saved?▼

Drafts are written under work/editorial_ops/ with a date-prefixed filename by default. You can supply an explicit output path to override this location.

What happens after the draft is produced?▼

The Skill returns the draft path, claim-to-source notes, and unresolved authoring questions. It then hands off to the fact_review and reader_experience_review steps as the defined next actions.

When should I not use this drafting approach?▼

Do not use it when no intake record exists or the source set cannot support even a first draft, since the startup checks will block execution. It is also unsuited for tasks that require hiding factual gaps behind polished wording.