Does it rewrite my entire document every time?

No, it is designed to be surgical. It identifies failing sections via a manifest and rewrites only those specific files to avoid churning high-quality existing work.

How does it handle citations?

It enforces strict citation hygiene by validating keys against your ref.bib file and ensuring citations stay within the scope defined in your evidence bindings.

What is the main benefit of using writer-selfloop?

It prevents 'hollow writing' by ensuring every section meets strict quality gates and evidence requirements before passing, rather than just generating generic prose.

What is 'generator voice' and how does this skill fix it?

Generator voice refers to repetitive LLM patterns like 'In this section we will...' or 'Next, we move to...'. This skill identifies these 'style smells' and replaces them with argument-driven transitions.

When should I skip the writer-selfloop?

You should skip this skill if you are still in the early outline or evidence collection phase (pre-C2), as it requires a stable evidence substrate to function correctly.

Writer Self-loop

Name: Writer Self-loop
Author: WILLOSCAR

byWILLOSCAR

•

Aprendizaje y Documentación

Automates the iterative refinement of research paper sections by identifying and fixing specific quality failures without rewriting the entire document.

Acerca de

The Writer Self-loop skill implements a sophisticated self-correction mechanism for academic and technical research writing. By treating writing as a series of semantic execution units, it identifies specific failures such as thin evidence, 'generator voice' fillers, or citation-scope violations. Instead of performing costly global rewrites, the skill targets only failing markdown sections, ensuring that the final output converges toward a high-quality, evidence-first draft that maintains strict citation hygiene and professional academic tone.

Características Principales

Strict citation validation against local .bib files and evidence bindings
89 GitHub stars
Automated semantic triage to classify writing versus evidence failures
Targeted rewriting of specific H3 body files to minimize content churn
Detection and removal of LLM-specific 'generator voice' and filler phrases
Iterative quality gate enforcement for consistent paper-wide voice

Casos de Uso

Maintaining citation integrity and academic tone across multi-author research drafts
Refining long-form technical surveys where sections lack sufficient evaluation anchors
Correcting 'hollow' writing by routing thin sections back to the evidence pipeline

Acerca de

Características Principales

Strict citation validation against local .bib files and evidence bindings
89 GitHub stars
Automated semantic triage to classify writing versus evidence failures
Targeted rewriting of specific H3 body files to minimize content churn
Detection and removal of LLM-specific 'generator voice' and filler phrases
Iterative quality gate enforcement for consistent paper-wide voice

Casos de Uso

Maintaining citation integrity and academic tone across multi-author research drafts
Refining long-form technical surveys where sections lack sufficient evaluation anchors
Correcting 'hollow' writing by routing thin sections back to the evidence pipeline