genre-skill-builder_skill

This skill helps researchers generate genre-analysis-based writing skills from corpus conclusions, producing structured phases, cluster guides, and technique

14

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill nealcaren/social-data-analysis --skill genre-skill-builder

  • SKILL.md10.6 KB

Overview

This skill builds new writing skills by analyzing a corpus of article sections and turning discovered genre patterns into a structured, reusable skill. It produces phased workflows, cluster profiles, and sentence-level technique guides grounded in quantitative and qualitative analysis. The result is a practical authoring tool tailored to a target section and venue.

How this skill works

The skill ingests a corpus of same-type sections (introductions, discussions, methods, etc.), computes descriptive statistics, and applies systematic coding to identify recurring moves and features. It runs cluster discovery to define 3–6 distinctive writing strategies, then generates a complete skill directory: phased workflows, cluster guides, and technique templates based on measured benchmarks and exemplars. Human review points are built into every phase so users confirm scope, codes, and cluster names.

When to use it

  • You want a writing guide for a specific article section (abstract, discussion, methods, etc.).
  • You need guidance derived from an empirical corpus rather than intuition or anecdotes.
  • You want a skill that follows a phased, repeatable workflow for research writing.
  • You plan to support multiple writing styles via cluster-based guidance.
  • You are preparing discipline- or venue-specific writing benchmarks and exemplars.

Best practices

  • Assemble a focused corpus of 30+ same-type sections from target venues when possible.
  • Select an existing model skill as a structural template before analysis begins.
  • Iterate on the codebook: run a small pilot coding pass, then refine categories for clarity.
  • Respect required human pauses: confirm scope, codes, and cluster names before generation.
  • Validate benchmarks and exemplars against held-out samples and author judgment.

Example use cases

  • Create a 'discussion-writer' skill from 50 Social Problems discussions to surface common closing moves and benchmarks.
  • Produce a methods-guide for qualitative articles by clustering approaches to participant recruitment and reflexivity.
  • Generate an abstract-writing skill for a conference by analyzing accepted abstracts and deriving tight opening/closing moves.
  • Adapt a model skill like lit-writeup to a new discipline by mapping corpus-derived parameters into templates.

FAQ

Aim for 30+ sections for stable clusters; minimum cluster rules allow smaller sets but expect less reliable differentiation.

Can I use mixed-section corpora?

No. The corpus should contain only one section type to keep feature coding and cluster discovery coherent.

How many clusters will be created?

Typically 3–6 clusters. Fewer than 3 suggests homogeneity or a coarse codebook; more than 6 risks over-differentiation.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational