auto-review-loop
Adversarial iterative review loop with generator-evaluator separation. Up to 4 rounds of independent review, improvement, and…
Converts the final Markdown manuscript from `paper-draft` / `paper-review-loop` into a submission package for the target venue — modular LaTeX (one file per section), compiled PDF, and Word `.docx`. Venue is read from `output/PAPER_PLAN.md` (or argument) and routed through a
$ npx -y skills add grind-lab-core/night_owl_research_agent --skill paper-covert --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/paper-covertContext preview
The summary Claude sees to decide when to auto-load this skill.
Converts the final Markdown manuscript from `paper-draft` / `paper-review-loop` into a submission package for the target venue — modular LaTeX (one file per section), compiled PDF, and Word `.docx`. Venue is read from `output/PAPER_PLAN.md` (or argument) and routed through a
name: paper-covert description: Converts the final Markdown manuscript from `paper-draft` / `paper-review-loop` into a submission package for the target venue — modular LaTeX (one file per section), compiled PDF, and Word `.docx`. Venue is read from `output/PAPER_PLAN.md` (or argument) and routed through a small YAML profile. Does not rewrite prose, score, or invent citations. argument-hint: [venue-or-manuscript-path] allowed-tools: Bash(*), Read, Write, Edit, Grep, Glob
Convert the final manuscript in `output/manuscript/` into a submission package for: **$ARGUMENTS**
> Naming: the pipeline in `paper-writing-pipeline` references this as `/paper-convert`. Folder name is `paper-covert` per invocation; they resolve to the same skill.
---
Inline override: `/paper-covert — venue: ISPRS, mode: latex`. Modes: `all` (default), `latex`, `pdf`, `docx`.
---
Primary:
Supporting (read when present):
---
output/submission/ ├── SUBMISSION_MANIFEST.md # source, venue, files, readiness, unresolved items ├── latex/ │ ├── main.tex # lean entry: \input's only │ ├── preamble.tex # packages + placeholder env │ ├── metadata.tex # title, authors, keywords, highlights │ ├── references.bib # copied or synthesized; else stub │ ├── build.sh # latexmk → pdflatex fallback │ ├── sections/ │ │ ├── 01_abstract.tex │ │ ├── 02_introduction.tex │ │ ├── 03_related_work.tex │ │ ├── 04_data.tex │ │ ├── 05_methodology.tex │ │ ├── 06_experiments.tex │ │ ├── 07_results.tex │ │ ├── 08_discussion.tex │ │ ├── 09_limitations.tex │ │ ├── 10_conclusion.tex │ │ ├── 11_declarations.tex │ │ └── appendix_*.tex │ └── figures/ # copied from output/figures/ per FIGURE_MANIFEST ├── pdf/MANUSCRIPT_SUBMISSION_READY.pdf └── docx/MANUSCRIPT_SUBMISSION_READY.docx
Filenames are stable; overwrite in place. Two-digit zero-padded section prefixes match `paper-draft`.
---
Pick `SRC` by priority above. Read `PLAN` for title, keywords, venue, declarations, figure/table plan. Load `PROFILE`. Record the source path, venue, and profile choice at the top of `SUBMISSION_MANIFEST.md`.
Section file mapping adapts to the manuscript's actual section list; `PLAN` §19 is authoritative when it differs from the default 01–11 layout. Appendices use `appendix_<letter>.tex`.
Write `build.sh` from `./templates/build_script_template.sh`. Run it (`latexmk -pdf`; fallback to `pdflatex + bibtex + pdflatex ×2`). Copy `main.pdf` to `pdf/MANUSCRIPT_SUBMISSION_READY.pdf`. On failure, keep the LaTeX tree intact and record the first error + likely cause in the manifest under **Build**.
Prefer `pandoc latex/main.tex -o docx/MANUSCRIPT_SUBMISSION_READY.docx --citeproc --bibliography=latex/references.bib`. If Pandoc rejects venue LaTeX, fall back to `pandoc` on the source Markdown (concatenated `SECTIONS_SRC/*.md`). If Pandoc is unavailable, record `docx: blocked` in the manifest.
Write `SUBMISSION_MANIFEST.md` from `./templates/SUBMISSION_MANIFEST_TEMPLATE.md`:
A fully automatic, domain-aware AI research agent for Geoscientists, Remote Sensing researchers, and GIScientists — powered entirely by Claude Code skills.
Adversarial iterative review loop with generator-evaluator separation. Up to 4 rounds of independent review, improvement, and…
Discover, evaluate, and download publicly available datasets from the internet. Infers data needs from a research question or…
Deploy and run experiments for ML/DL training (local, remote, or Modal GPU) AND spatial data science / GIScience experiments…
Run an end-to-end workflow that chains the skills `refine-research` and `experiment-design`. Use when the user wants a one-shot…
Turn a refined GIScience / remote sensing / spatial data science proposal into a detailed, claim-driven experiment roadmap. Use…
Complete 4-stage end-to-end research pipeline. Orchestrates idea-discovery-pipeline → deploy-experiment → auto-review-loop →…