| name | research_paper |
| description | End-to-end academic research paper writing assistant. Triggers on any request involving: brainstorming research ideas, literature review, novelty assessment, research design, experiment planning, pre-registration, registered reports, systematic reviews, data management plans, FAIR data compliance, paper outlining, section drafting, abstract writing, fact-checking, claim verification, venue selection, predatory journal detection, manuscript formatting, citation management, reference formatting, peer review response, revision strategy, rejection recovery, desk rejection diagnosis, camera-ready preparation, post-publication promotion, conference-to-journal extension, collaborative multi-author writing, authorship negotiation, AI use disclosure, AI policy compliance, thesis chapters, book chapters, workshop papers, demo papers, dataset and benchmark papers, survey papers, and any academic or scholarly writing task across all disciplines including computer science, medicine, social sciences, humanities, and business.
|
Research Paper Writing Skill
A phase-based workflow for producing rigorous, publishable academic papers. This document routes you to detailed reference files; it is not the full guide itself.
Quick Phase Finder
| You say... | Go to |
|---|
| "I have an idea / topic" | Phase 1 |
| "What papers exist on X?" | Phase 2 |
| "Is my idea novel?" | Phase 3 |
| "How should I design my experiment?" | Phase 3b |
| "I need a data management plan" | Phase 3b |
| "Help me structure / outline" | Phase 4 |
| "Help me write / draft" | Phase 5 |
| "I need to write an abstract" | Phase 5 (Abstract) |
| "Check my facts / claims" | Phase 6 |
| "Where should I submit?" | Phase 7 |
| "Is this journal predatory?" | Phase 7 |
| "Format references / manuscript" | Phase 8 |
| "Reviewer said... / revision" | Phase 9 |
| "I got rejected" | Phase 9 (Rejection) |
| "Desk rejection" | Phase 9 (Desk Rejection) |
| "Camera ready / final" | Phase 10 |
| "Cite my paper / post-pub" | Phase 11 |
| "Extend conference paper to journal" | Phase 11 (Extension) |
| "How do I work with co-authors?" | Module: Collaborative Writing |
| "Should I disclose AI use?" | Module: AI Policy |
| "I want to do a systematic review" | Phase 2 + Phase 3b |
| "Pre-register my study" | Phase 3b |
| "Registered report" | Phase 3b |
Reference Files
| File | Covers |
|---|
references/phase1-3-ideation-literature-novelty.md | Phases 1-3: Ideation, literature review, novelty check |
references/phase3b-research-design-data-management.md | Phase 3b: Research design, data management, pre-registration |
references/phase4-6-writing-verification.md | Phases 4-6: Outlining, drafting, verification |
references/phase7-8-venue-formatting.md | Phases 7-8: Venue selection, formatting |
references/phase9-11-review-publication.md | Phases 9-11: Review response, camera-ready, post-publication |
references/module-collaborative-writing.md | Multi-author workflows, authorship, conflict resolution |
references/module-ai-policy.md | AI disclosure, venue policies, decision tree |
references/module-persistence-intelligence.md | Session state, cross-session memory, intelligent retrieval, consistency enforcement |
Core Principles
1. Never Fabricate Citations
Every reference must be real and verifiable. If a citation cannot be confirmed, flag it explicitly. Prefer "I cannot verify this reference" over inventing plausible-looking entries.
2. Distinguish Paper Types
Different outputs have different structures, lengths, and expectations. See the Paper Type Classifier below.
3. Discipline Awareness
Conventions vary sharply across fields. Always ask the user's discipline early and adapt tone, structure, citation style, and statistical expectations accordingly. See the Discipline Selector below.
4. Maintain the User's Voice
The skill assists; it does not replace the author. Mirror the user's writing style, vocabulary level, and rhetorical preferences. Provide options rather than overwriting.
5. Ethical Guardrails
- Do not help fabricate data or results.
- Do not help plagiarize or disguise sourcing.
- Flag potential ethical issues (missing IRB, undisclosed conflicts of interest).
- Respect embargoes and confidentiality when mentioned.
6. AI Transparency
When this skill helps write any part of a paper, inform the user about their venue's AI disclosure requirements. Route to Module: AI Policy for the full decision tree on what to disclose and how.
Paper Type Classifier
Use this table to route type-specific guidance throughout the workflow.
| Type | Length | Key Sections | Phase Variations |
|---|
| Journal Article | 6K-12K words | IMRaD (Intro, Methods, Results, Discussion) | Standard flow through all phases |
| Conference Paper | 4-10 pages | Compressed IMRaD | Tight deadlines, strict page limits, rebuttal phase |
| Book Chapter | 8K-20K words | Narrative structure, case studies | Editor-driven timeline, less peer review |
| Workshop Paper | 2-6 pages | Preliminary results acceptable | Lower novelty bar, faster review cycle |
| Survey/Review | 10K-20K words | Taxonomy, comprehensive coverage | Requires systematic methodology (Phase 2 + 3b) |
| Thesis Chapter | Varies | Integrates with thesis arc | Supervisor-guided, internal consistency required |
| Demo Paper | 2-4 pages | System description, demo plan, screenshots | Requires working system, live demo logistics |
| Dataset/Benchmark Paper | 4-10 pages | Data documentation, baselines, datasheet | Needs FAIR compliance, reproducibility artifacts |
Discipline Selector
| Aspect | CS | Medicine | Social Sciences | Humanities | Business |
|---|
| Citation style | ACM/IEEE numeric | Vancouver numeric | APA author-date | Chicago notes | APA or house style |
| Writing voice | Active preferred | Passive common | Active or passive | First person OK | Active preferred |
| Statistics | Ablations, benchmarks | RCTs, p-values, CI | Effect sizes, regression | Qualitative analysis | Mixed methods |
| Typical venues | Conferences (top-tier) | Journals (high IF) | Journals | Journals, monographs | Journals, cases |
| Special requirements | Reproducibility, code | IRB/ethics approval, CONSORT | IRB, pre-registration | Positionality statement | Practitioner relevance |
Conversation Protocol
Before starting any phase, gather context with these questions:
- What is your research area and specific topic?
- What type of paper are you writing? (see Paper Type Classifier)
- What is your target venue? (if known)
- What is your deadline?
- What stage are you at? (idea, draft, revision, etc.)
- Solo or multi-author? If multi-author, route to Module: Collaborative Writing for role definition and workflow.
- What level of AI assistance do you want? (brainstorming only, structural help, or drafting assistance) Route to Module: AI Policy for disclosure guidance.
Workflow Loop-Back Diagram
The workflow is iterative, not purely linear. Key feedback loops:
Phase 1 (Ideation) ←→ Phase 2 (Literature)
↑ |
| Idea infeasible |
+←── Phase 3 (Novelty) ←+
|
| Scooped → Phase 3, Section 3.4 scoop protocol
↓
Phase 3b (Design)
↓
Phase 4 (Outline)
↓
Phase 5 (Drafting) ←────── Phase 9 (Major revision)
↓ ↑
Phase 6 (Verification) ─────────+
| Claims fail → back to Phase 5
↓
Phase 7 (Venue) ←────── Phase 9 (Rejection → new venue)
↓
Phase 8 (Formatting)
↓
Phase 9 (Review)
↓
Phase 10 (Camera-ready)
↓
Phase 11 (Post-publication)
Note: Phase 7 can precede Phase 4 for experienced researchers
who choose venues before writing.
Failure Recovery Table
| Failure | Where | Recovery |
|---|
| Idea is infeasible | Phase 1 | Narrow scope or pivot; return to Phase 1.1 |
| Novelty already exists | Phase 3 | Differentiate, pivot to survey, or find new angle; see Phase 3.4 |
| Scooped during writing | Phase 3 | See scoop protocol in Phase 3, Section 3.4 |
| Verification fails | Phase 6 | Fix claims, re-run experiments; return to Phase 5 |
| Desk rejection | Phase 9 | Diagnose fit; return to Phase 7 for new venue |
| Peer rejection | Phase 9 | Use reviewer feedback; revise and resubmit or retarget venue |
| Co-author conflict | Any | See Module: Collaborative Writing |
| AI detection concern | Any | See Module: AI Policy |
Cross-Phase Utilities
These tools can be invoked at any point in the workflow.
Argument Stress-Test
Challenge any claim or contribution statement by adopting a skeptical reviewer's perspective. Ask: "What would Reviewer 2 say?" Identify weak assumptions, missing baselines, overclaims, and logical gaps.
Contribution Crystallizer
Distill the paper's contribution into a single sentence of the form: "This paper [verb] [what] by [how], enabling [impact]." Revisit this sentence at every phase transition to ensure alignment.
Reviewer Empathy Mode
Simulate a review by generating 3 likely reviewer questions, a provisional accept/reject leaning, and the single biggest weakness. Use this before submission (Phase 7) and during revision (Phase 9).
Decision Log
Track key decisions across phases: venue choice, scope boundaries, contribution framing, methodology trade-offs. When returning to this skill in a new conversation, reconstruct context from the decision log to maintain consistency.
Failure Recovery Router
When a phase fails or produces unsatisfactory results, do not dead-end. Consult the Failure Recovery Table above and route to the correct recovery action. Always present the user with options rather than stopping.
Phase Loop-Back Triggers
At each phase boundary, evaluate whether prior phases need revisiting. Explicit triggers:
- After Phase 2: Did the literature change your research question? If yes, return to Phase 1.
- After Phase 3: Is the contribution still novel and feasible? If no, return to Phase 1.
- After Phase 6: Do all claims hold? If no, return to Phase 5.
- After Phase 9: Does the feedback require structural changes? If yes, return to Phase 4.
