Skip to content

About

Author-led, evidence-linked manuscript editing.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

3 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

weaver-text badge

weaver-text

AI editing help for scientists who write their own manuscripts and grant proposals. The author supplies the ideas, argument, and data and makes the final decisions on wording. The skill edits, reviews, checks citations, and prepares the submission, at the scope each task needs.

Part of a research suite: weaver-text (editing support for author-written text), honeybee-lit (general literature reviews), and wasp-lit (post-GWAS candidate gene-phenotype literature analysis).

Why a weaver?

A weaver bird builds its nest strand by strand, and a poorly built nest is inspected and rebuilt. weaver-text helps an author tighten each strand of their own text: every claim ties to its evidence, and every review finding comes with a concrete revision the author can accept or reject.

Reproducibility and standardization

  • Saved sources. Retrieved guidelines and references are saved locally with provenance, including during reviews (Module 08).
  • Current official rules. Journal and agency instructions are retrieved and followed. Requirements are never inferred from common practice.
  • Reproducible figures. Figures and tables are generated from supplied results, with source records, never from invented values.
  • Standard review format. Every finding, including minor ones, gives a location, rationale, severity, and suggested revision.
  • Reporting-standard checks. Statistics, causal claims, and study design are checked against the relevant reporting guidelines (Module 02).
  • Decision logs. Intake actions, edit batches, and author decisions are recorded in log/intake.md, log/edits.md, and log/decisions.md.
  • Verified output. Rendered documents are inspected before any submission-readiness claim, and blocked checks are reported.

What it does

Task What the author gets
Edit Clearer, tighter versions of the author's own sentences or sections, within the requested scope.
Review Findings with suggested revisions; the author's text stays unchanged until they approve a change.
Production Help assembling the author's material into a submission: figures from supplied results, references, compliance checks, export, and packaging.
  • Line editing for clarity, concision, tense, and tone, keeping the author's voice and meaning.
  • Structural feedback on sections and argument flow, offered as suggestions.
  • Citation verification, claim-support checks, and publication-status checks.
  • Simulated scientific and funding reviews to stress-test a draft before submission, plus help with rebuttals.
  • Markdown for early drafts; LaTeX, DOCX, and ODT for production. Existing formats and bibliography systems are kept unless the author approves a change.

The author stays in charge

  • The skill does not invent facts, methods, results, or references. Gaps become [AUTHOR QUERY: ...] markers for the author to resolve.
  • Edits stay within the requested scope. Larger changes are proposed, not made.
  • Review suggestions are applied only with the author's approval.
  • Text is never cut solely to meet a length limit; cuts are proposed for the author to choose.
  • Software installation and Git changes require permission.
  • Changes the skill makes to files are tracked (DOCX/ODT tracked changes, latexdiff for LaTeX, diffs for plain text) so the author can accept or reject each one.

Package contents

weaver-text/
├── SKILL.md                    # Agent instructions and module routing
├── README.md
├── LICENSE
├── assets/weaver-text.jpg
├── evals/                      # Behavior checks for scope, approvals, and provenance
└── skill-modules/
    ├── modules/                # Modules 01 to 08 and the writing-sources register
    └── templates/              # Guidelines README, requirements matrix, results manifest, search record

Project workspace

Create only the folders a task needs, and keep any existing layout (Module 01).

Purpose Folders
Receive files intake/
Establish context guidelines/, references/, results/
Plan and write plan/, drafts/, figures/
Evaluate reviews/
Deliver final/
Record log/

Files in intake/ are sorted without changing contents. results/ holds supplied findings, not an analysis pipeline. When a reanalysis is needed, the skill recommends that the author perform it.

Companion skills

Skill Role
weaver-text AI editing, review, and submission help for author-written manuscripts and grant proposals.
honeybee-lit General literature reviews, from citation checks to narrative, scoping, and systematic reviews.
wasp-lit Post-GWAS literature analysis of candidate gene-phenotype relationships.

Installation

Clone the repository, unchanged, into a skills location your agent scans:

# Pi
git clone https://github.com/chen42/weaver-text.git ~/.pi/agent/skills/weaver-text
# Claude Code (user or project scope)
git clone https://github.com/chen42/weaver-text.git ~/.claude/skills/weaver-text

In Pi, run /reload after cloning. Export to DOCX, ODT, or PDF depends on locally available tools such as LaTeX, Pandoc, or LibreOffice.

License

MIT

About

Author-led, evidence-linked manuscript editing.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages