# Academic Researcher Writing

> Academic Researcher & Writer

- Skill: `ziadnagar/academic-researcher-writing` (Agent Skill, multi-file: 6 files)
- Install (CLI): `npx skillmds@latest add ziadnagar/academic-researcher-writing`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ziadnagar/academic-researcher-writing/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ZiadNagar (https://skillmd.com/u/ziadnagar)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/ziadnagar/academic-researcher-writing

---


# Academic Researcher & Writer

A unified skill for the full academic pipeline: **find → synthesize → write → refine**.

Covers two phases that often overlap:

1. **Research Phase** — query formulation, source discovery, RAG-style synthesis, gap analysis
2. **Writing Phase** — structure selection, drafting, citation formatting, peer review response

---

## When to Use Which Module

| User says… | Go to |
|---|---|
| "Find papers on X", "what does the literature say about Y" | → [Research Phase](#research-phase) |
| "Help me write my intro / lit review / methods" | → [Writing Phase](#writing-phase) |
| "Respond to reviewer comments" | → [Peer Review Response](#peer-review-response) |
| "Write a grant proposal / abstract / thesis chapter" | → [Writing Phase](#writing-phase) |
| "I need to research AND write about X" | → Do both phases in sequence |

---

## Research Phase

### Step 1 — Clarify the Research Question

Before searching, pin down:
- **Topic + scope**: What domain? What time frame?
- **Question type**: Empirical ("does X cause Y?"), theoretical ("how does X explain Y?"), methodological ("what methods are used to study X?"), or applied ("how is X used in practice?")
- **Depth**: Quick overview vs. comprehensive systematic review
- **Output format**: Summary, annotated bibliography, synthesis table, or draft section

### Step 2 — Query Formulation

Translate user intent into effective search strings:

```
Boolean logic:  (machine learning OR deep learning) AND (climate change OR global warming)
Field-specific: ("neural network" AND "interpretability") NOT survey
Date-bounded:   after:2020 before:2025
```

**Key databases by discipline:**

| Field | Primary | Secondary |
|---|---|---|
| CS / AI / ML | ArXiv, Semantic Scholar | ACM DL, IEEE Xplore |
| Life sciences | PubMed / MEDLINE | bioRxiv, PLoS |
| Social sciences | SSRN, JSTOR | Google Scholar |
| All fields | Google Scholar | Semantic Scholar |
| Open access | OpenAlex | CORE, Unpaywall |

### Step 3 — Source Evaluation (SIFT + CRAAP)

For each source, check:
- **Currency**: Publication date, last update
- **Relevance**: Directly addresses the question?
- **Authority**: Peer-reviewed? Author credentials? Journal impact factor?
- **Accuracy**: Methodology sound? Data available? Reproducible?
- **Purpose**: Research vs. opinion vs. advocacy?

Citation counts and forward citations (who cites this paper?) are strong relevance signals.

### Step 4 — Synthesis

Don't summarize papers one-by-one. Instead:

1. **Theme map**: Group papers by argument, finding, or method
2. **Tension detection**: Where do sources agree? Where do they conflict?
3. **Gap identification**: What question is unanswered? What population/context is missing?
4. **Timeline**: How has consensus evolved?

Output a **synthesis matrix** if the user has 5+ sources:

```
| Source | Method | Sample/Context | Key Finding | Limitations |
|--------|--------|----------------|-------------|-------------|
| ...    | ...    | ...            | ...         | ...         |
```

---

## Writing Phase

### Document Type Selection

| Task | Structure |
|---|---|
| Empirical paper | IMRaD (Intro, Methods, Results, Discussion) |
| Review / survey | Intro → Scope → Themes → Gaps → Conclusion |
| Theoretical paper | Problem → Framework → Analysis → Implications |
| Grant proposal | Specific Aims → Background → Approach → Significance |
| Conference abstract | Background → Objective → Methods → Results → Conclusion (150–300 words) |
| Thesis chapter | Chapter intro → lit → methods → findings → summary |

### IMRaD Deep Dive

**Introduction** — build the "inverted triangle":
1. Broad context (why does this field matter?)
2. Specific gap or problem (what is not yet known?)
3. Study rationale (why this approach?)
4. Research question / hypothesis (explicit statement)
5. Paper overview (optional 1–2 sentences)

**Methods** — enough to replicate:
- Participants / data sources (sampling, inclusion/exclusion criteria)
- Procedure / instruments
- Analysis approach (statistical tests, models, frameworks)
- Ethical considerations if applicable

**Results** — present, don't interpret:
- Lead with the answer to each research question
- Tables and figures before prose discussion of them
- Negative/null results deserve equal reporting

**Discussion** — interpret and contextualize:
1. Restate key findings (briefly)
2. Relate to prior literature (support or contradict?)
3. Explain unexpected findings
4. Acknowledge limitations honestly
5. Implications (theoretical and/or practical)
6. Future work
7. Conclusion paragraph

### Literature Review (Standalone)

Structure:
```
1. Introduction (scope + purpose)
2. Search strategy (transparent and reproducible)
3. Thematic sections (not paper-by-paper summaries)
4. Critical analysis (tensions, gaps, methodological issues)
5. Conclusion + research agenda
```

**Rules:**
- Every paragraph should advance an argument, not just report
- Use topic sentences that make a claim; then support with citations
- Avoid "Smith (2020) found that…" as a paragraph opener — lead with the idea

### Grant Proposal

Read the call FIRST. Then:

**Specific Aims** (1 page — most important):
- Hook sentence (the problem)
- Long-term goal
- Objective of this grant
- Central hypothesis
- 3–4 specific aims (verb + measurable outcome)
- Innovation + significance statement

**Research Strategy**:
- Significance: Why now? What changes if successful?
- Innovation: What's new about your approach?
- Approach: Aim-by-aim with methods, expected outcomes, potential problems, alternatives

**Common mistakes to avoid:**
- Vague aims ("we will study X") → make them testable
- No preliminary data → include even small pilot results
- Ignoring reviewer perspective → answer "so what?" constantly

---

## Peer Review Response

### General Strategy

Reviewers want to see:
1. You read every comment carefully
2. You either fixed it or have a principled reason not to
3. Changes are easy to find in the manuscript

### Response Structure

```
Dear Editor and Reviewers,

Thank you for the thorough and constructive feedback on [title]. 
We have addressed all comments and summarize our responses below.

---

REVIEWER 1

Comment 1: [Quote reviewer comment verbatim]
Response: [Your response — direct and specific]
Manuscript change: [Where/what changed, or "No change — see rationale below"]

Comment 2: ...
```

### Tone Rules
- Never defensive, never dismissive
- "We agree that..." or "We respectfully disagree because..." — both are valid
- If you disagree, provide evidence or logical argument
- Acknowledge when a comment improved the paper

---

## Citation & Style Quick Reference

| Style | Common in | Key feature |
|---|---|---|
| APA 7 | Social/behavioral sciences, education | Author-date in text |
| MLA 9 | Humanities | Author-page in text |
| Chicago / Turabian | History, some humanities | Footnotes or author-date |
| Vancouver | Medicine, biomedical | Numbered references |
| IEEE | Engineering, CS | Numbered, bracketed |
| Harvard | Business, sciences (UK) | Author-date variant |

For AI/ML papers: Use the citation style of the target venue (NeurIPS, ICML = custom BibTeX; ACL = ACL anthology format).

---

## Output Quality Checklist

Before finalizing any piece of academic writing, verify:

- [ ] Research question is explicit and answerable
- [ ] Claims are supported by cited evidence
- [ ] Methods are described reproducibly
- [ ] Limitations are honestly stated
- [ ] Hedging language used where appropriate ("suggests", "may", "appears to")
- [ ] Abstract stands alone (no undefined acronyms, no citations)
- [ ] Correct citation style applied consistently
- [ ] Passive vs. active voice matches field conventions
- [ ] Word count within target range
- [ ] Figures/tables have captions that stand alone

---

## Reference Files

- `references/research-databases.md` — Full database guide with access tips
- `references/writing-patterns.md` — Sentence-level patterns for each IMRaD section
- `templates/abstract-template.md` — Structured abstract for 5 common formats
- `templates/peer-review-response.md` — Full response letter template
- `templates/grant-aims-page.md` — NIH-style specific aims template

---

## Quick Prompts (Copy-Ready)

**To kick off research:**
> "I'm researching [topic]. My research question is [RQ]. Target audience: [journal/field]. I need [overview / comprehensive lit review / specific papers on subtopic]. Please start with Step 1 of the Research Phase."

**To kick off writing:**
> "I'm writing a [paper type] about [topic] for [venue/course]. I have [sources / no sources yet]. Please guide me through the Writing Phase starting with structure selection."

**To get peer review help:**
> "I have reviewer comments for my paper on [topic]. I'll paste the comments — please help me write a professional, point-by-point response letter."