Follow this workflow end-to-end unless the user explicitly asks to skip steps

1) Intake and scope

• Confirm input artifacts: TeX source path(s), PDF path, supplementary files, target repository, and expected outputs.
• State assumptions explicitly when information is missing.
• Keep approach adaptable to the specific paper; do not force a fixed dependency stack or rigid project template.
• Check whether the working folder already contains paper source files (.tex, .bib, style files, figures).
• Check whether the working folder contains user-preprocessed documents (.md, .json, images such as .png, .jpg, .svg).
• Source priority rule (read in order, stop when sufficient):
  1. TeX sources (preferred): If usable TeX source files (.tex, .bib, style files) are present, use them as the primary source.
  2. User-preprocessed documents (secondary): If TeX is absent or incomplete, read user-provided documents (.md, .json) and images (.png, .jpg, .svg) that may contain pre-extracted paper content.
  3. PDF fallback (last resort): Only when both TeX and user-preprocessed documents are unavailable or insufficient, fall back to PDF extraction.

2) Source extraction (TeX → preprocessed docs → PDF)

• TeX path (highest priority):

  • Parse and read the main TeX project structure first (main.tex or equivalent entry file and includes).
  • Preserve original scientific wording when converting relevant content to Markdown notes.
  • Resolve equations, theorem blocks, citations, and appendices from source files whenever possible.
  • Record unresolved include/bibliography issues explicitly; do not invent missing content.

• User-preprocessed documents path (secondary):

  • Read Markdown files (.md) that may contain paper content extracted by the user.
  • Read JSON files (.json) that may contain structured paper data (metadata, sections, references).
  • View images (.png, .jpg, .svg) that may contain paper figures, tables, or scanned pages.
  • Preserve original content; do not summarize or paraphrase.
  • Note the source of each piece of information (which file, which section).

• PDF fallback path (lowest priority, when all else fails):

  • Extract paper content page by page into Markdown, preserving the original wording.
  • Do not summarize, paraphrase, or rewrite scientific statements.
  • Preserve structure faithfully:
    • Title, authors, affiliations, abstract, sections, subsections.
    • Equations (LaTeX-friendly when possible), theorem/lemma/proposition blocks.
    • Tables, figure captions, references, appendices, footnotes.
  • If a PDF is scanned or partially unreadable:
    • Run OCR and mark uncertain spans clearly.
    • Never silently invent missing text.
  • Include image references/placeholders when figures cannot be represented as plain text.
  • Produce one primary output file such as paper_fulltext.md.

3) Extraction quality checks