Create Introduction

/create-introduction

Produce a publication-quality introduction as a LaTeX fragment. The section has one job: leave a reader, expert or not, comfortable enough to understand everything that follows.

Before anything else

1. Load the style rules. Invoke paper-writing-style and read references/human-voice.md and references/latex-conventions.md. If it is not installed, use the condensed rules at the bottom of this file.
2. Lock the language. Detect the language of the user's first message and keep it for the whole run. Never switch mid-run. Output LaTeX is always English.
3. Read the context. Find .tex files, the .bib file, and the results and discussion section if it exists. Read the .bib keys: they are the literature you have to work with. Detect the citation style. Do not ask for anything you can find yourself.

Write this section third, not first

The course puts the introduction third in the writing order, after results and discussion and after conclusions, for one reason: writing it once you know what the results say produces an introduction aimed directly at them, so the connection between the gap you state and the gap you answer is never lost. If the results section does not exist yet, say so and offer to continue anyway, but note that the introduction will probably need revising later.

The full order is in paper-writing-style, and /write-paper walks it.

The four movements

The state of the art is the movement most often missing. Without it the reader cannot judge whether the paper advances anything, because they do not know where the frontier was.

The three dimensions

Every introduction has to work on all three at once. Check each before delivering.

1. Information. The four movements above, each actually present.

2. Flow: general to specific. The section funnels. It opens on the broad field and narrows continuously to the purpose, which is the most specific thing in it. This works inside paragraphs too, not just across them. A worked example of narrowing within one paragraph:

While the vast majority of layer-by-layer systems exhibit linear growth, certain polyelectrolyte combinations grow exponentially. Recently, a modified version of the technique was introduced by our group. A substantially smaller and more recent subset of the layer-by-layer literature describes \ldots

Each sentence closes the aperture. Never widen back out until the conclusions section.

3. Citation and referencing. Three classes of source belong in an introduction, and knowing which is which is how you choose from hundreds of candidates:

If the .bib file has no recent entries, say so. An introduction whose newest citation is four years old reads as out of touch, and reviewers notice.

Workflow

1. Gather what exists

Accept a short description, paths to the manuscript or results section, a .bib file, a list of key papers, or nothing. Read everything given before asking a question.

2. Interview

Batches of at most four questions, using AskUserQuestion where answers are selectable. Continue until all four movements can be written without inventing anything. Cover:

• Target journal, which sets how much contextualization is needed. A specialist venue's audience already knows why the field matters; a broad one such as Nature does not and the introduction has to establish it.
• The broad field, and the terms, jargon and methods a reader needs defined.
• The gap, pressed hard. This is the movement that determines whether the paper looks necessary. What is unresolved, unexplained, contested, or prohibitively expensive? A gap of "more research is needed" is not a gap.
• The state of the art. Which recent papers define the current frontier? Ask for keys or citations. If the researcher cannot name any from the last year or two, raise it.
• Seminal work that founded the field or the technique.
• The purpose, and confirmation that it answers the stated gap.
• Length, in paragraphs or words, if the journal constrains it.

3. Draft

Emit one LaTeX fragment, in English, with movement labels:

% CONTEXTUALIZATION
Layer-by-layer assembly produces ultrathin polymer films with nanometre control over
thickness and composition \citep{decher1997}, and has become a standard route to
functional coatings for biomedical devices \citep{lee2019}.
% GAP
However, the mechanical stability of these films under physiological shear remains poorly
understood, which limits their use in implantable devices \citep{costa2021}.
% STATE OF THE ART
Recently, significant advances have been reported in crosslinking strategies that improve
film cohesion \citep{park2024,mendes2025}, although these have been characterized only
under static conditions.
% PURPOSE
In this paper, we report the shear stability of crosslinked layer-by-layer films under
physiological flow, measured over 30 days.

Show it in chat in a fenced latex block. Then state the paragraph count and, in a few lines, which choices you made: how much context you spent and why, which papers you placed where, and how the funnel narrows.

4. Check and refine

Ask what to change. Expect pushback on the gap and on the choice of state-of-the-art citations; both are judgement calls only the researcher can settle. Show the full fragment each round.

5. Verify against the results

Before finishing, check the introduction against the results and discussion section if it exists. Every relevance claim, gap and state-of-the-art reference in the introduction should have its counterpart later in the paper. The two sections must be tightly connected, and an introduction that promises something the results never address is the most common structural fault in a manuscript.

6. Offer to write it out

Only after approval, offer to write the fragment to a file or patch the manuscript.

Hard rules

• Never invent a citation key or a reference. If a claim needs a source you do not have, emit % TODO: cite <what is needed>. Never attach the nearest plausible key from the .bib.
• Never invent the state of the art. If you do not know what was published recently in the field, say so and ask. Fabricating a frontier is worse than leaving a gap.
• Never emit an em dash (---). -- only in numeric ranges.
• Never copy text from a cited source. Read the idea, state it in your own words, cite the author. Citation says where the idea came from, never where the text came from. The only exception is a direct quotation in quotation marks, which is rare in this genre.
• The funnel never widens. Check the draft narrows monotonically.
• Past tense for what was done, present for what is currently known.
• Cut it is important to note that, plays a key role, novel, and any sentence that would fit unchanged in a different paper.

Condensed style rules, if paper-writing-style is not installed

Clarity and concision govern. One idea per sentence, short enough that the reader reaches the full stop with the opening in mind. Be specific: a sentence that could appear in any paper in the field says nothing. Put the action in the verb. Cut redundancy and anything the editor already assumes. Prefer plain words: use over utilization, to over in order to, most over the majority of, because over due to the fact that. Avoid delve, harness, leverage, underscore, pave the way, crucial role. Keep however, moreover, in contrast, in this paper, here we report: correct scientific register, not filler. Never stack passive voice. Vary sentence openers between subject-verb, adverb, prepositional phrase, dependent clause and infinitive phrase. Every it and this needs one possible referent. Thin space between quantity and unit (10\,nm), escape \%.