# Manuscript Ingest

> Manuscript Ingest is an Agent Skill for documents published in the WILLOSCAR/research-units-pipeline-skills repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Use when `paper-review` needs a canonical manuscript text artifact before claim extraction. **Trigger**: ingest paper, manuscript text, provide paper, paper.md, 输入论文, 导入稿件, 审稿输入. **Use when**: You are running the `paper-review` pipeline and need `output/PAPER.md` before `claims-extractor`. **Skip if**: `output/PAPER.md` already exists and looks like the full manuscript text. **Network**: none. **Guardrail**: Do not summarize or rewrite the paper; store the raw text (or a faithful extraction) so claims stay traceable.

- Skill name: `manuscript-ingest`
- Category: [Documents](https://www.openparable.com/skill/documents/1)
- Source: https://github.com/WILLOSCAR/research-units-pipeline-skills/blob/HEAD/.codex/skills/manuscript-ingest/SKILL.md
- Repository: WILLOSCAR/research-units-pipeline-skills (487 GitHub stars when collected)

## SKILL.md

# Manuscript Ingest

Transforms a manuscript source file into the canonical text artifact used by `paper-review`.

## Inputs

One manuscript source from the workspace, typically:
- `inputs/manuscript.md`
- `inputs/manuscript.txt`
- `inputs/manuscript.pdf`

## Output

- `output/PAPER.md`

## Script boundary

`scripts/run.py` should:
- find the simplest available manuscript source
- extract faithful text
- write the full text to `output/PAPER.md`

It should not summarize, critique, or reformat the manuscript into a review.

## Contract

The output must preserve:
- paper body text
- section headings when available
- page markers when extractable

## Acceptance

- `output/PAPER.md` exists
- it contains the manuscript body rather than only title/abstract

## Non-goals

- claim extraction
- evidence auditing
- review writing

---

Source: https://www.openparable.com/skill/documents/manuscript-ingest (OpenParable). Cite the canonical URL above.