intro-related-writing
Skill2 starsChanged 3 months agoMITView on GitHub
Write or revise a paper's Introduction and Related Work sections using recipes distilled from dissecting top ML researchers' papers. Use when drafting an intro, a contributions list, or a related-work section, or when restructuring how a paper positions against rivals. Applies the seven-question funnel, the credit-then-contrast pattern, and mechanical checklists.
In this repository9 other files it gives its agents.
- abstract-writingSkill
- caption-writingSkill
- conclusion-writingSkill
- figure-creationSkill
- method-writingSkill
- rebuttal-writingSkill
- results-writingSkill
- sentence-auditSkill
- supplement-writingSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Follow intro-related-guide.md in this folder, the complete plain-English method.
- Distinctive: line 12
Contributions first: Before any intro prose, write the contribution bullets: each one a refutable claim with a strong bold opening line, …
- Distinctive: line 24
Every rival gets credit-then-contrast: name what they introduced (warmly), state what you keep, then the one difference and its consequence.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Introduction + Related Work (master method)
Follow intro-related-guide.md in this folder, the complete plain-English method.
This file is the operating summary.
Process
- Contributions first. Before any intro prose, write the contribution bullets: each one a refutable claim with a strong bold opening line, not a parts list. The intro is then built backwards from them, every paragraph earns a bullet.
- Draft the funnel: seven reader questions, in order: why care (hook) / what world (shared ground) / what is missing and what it costs (the gap, named as concrete costs) / what you built (the artifact, named once) / how it works (ONE mechanism, concept only, details go in Method) / what you found (one headline number, stated early, never saved for a surprise) / how it differs (one clause; the full comparison lives in Related Work).
- Weave related work by theme, far to near. Group by idea, not by paper; order groups from distant background to the closest rival, closest LAST. For engaged rivals make the system the sentence subject; for background use citation clusters. Every rival gets credit-then-contrast: name what they introduced (warmly), state what you keep, then the one difference and its consequence. Never write an annotated bibliography (a list of papers with no relation to your work).
- Run the checklists in the guide (intro + related work), then read both cold, in one pass, as a stranger.
Hard rules
- The intro and the abstract must tell the same story; if the abstract's thesis changed, the intro's prose funnel changes with it, not only the bullets.
- Attach every number to the exact claim it measures; numbers travel in pairs.
- One named axis for your key idea, coined once and reused verbatim.
- Your closest rival is your likeliest reviewer: credit their paper warmly and place its group last, immediately before your differentiation.
- Mechanism details (constants, per-step mechanics) belong in Method; the intro keeps only the concept.
Reference file
intro-related-guide.md: full guide: term translations, the funnel, three intro recipes with fill-in templates, the related-work recipe with worked credit-then-contrast examples, checklists, tense table, assembly line.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.