borghei/Claude-Skills / research/litreview
litreview
VerifiedLiterature review for academic and R&D research: search strategy, source assessment, synthesis, and citation management. Use when conducting a systematic literature review, building a bibliography, or preparing a research-grounded report.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 8.1 KB |
| references/search-strategy-and-prisma.md | markdown | 6.1 KB |
| references/source-quality-assessment.md | markdown | 5.8 KB |
| references/synthesis-and-citation-management.md | markdown | 6.4 KB |
| scripts/search_strategy_builder.py | script | 7.5 KB |
| scripts/source_quality_scorer.py | script | 7.8 KB |
| scripts/thematic_synthesis_builder.py | script | 9.0 KB |
references/synthesis-and-citation-management.md
View on GitHub# Synthesis & Citation Management Reference
Reference for synthesizing literature into themes and managing citations.
## 1. Synthesis approaches
### Narrative synthesis
- Heterogeneous sources; explanatory text
- Default for non-medical reviews
- Risk: subjective; can read as opinion piece
### Thematic synthesis
- Identify themes across sources
- Translate findings into a common framework
- Document evidence strength per theme
- Default for most non-medical systematic reviews
### Meta-analysis
- Quantitative pooling of comparable studies
- Effect sizes, forest plots
- Requires statistical expertise
- Only for sufficiently homogeneous studies
### Realist synthesis
- Context-Mechanism-Outcome configurations
- For complex interventions (educational, organizational)
- Asks: what works, for whom, under what circumstances?
### Scoping review
- Maps a field rather than answers a question
- Less stringent criteria
- Output: themes, gaps, key authors
Pick the approach that fits the question and the literature.
## 2. Thematic synthesis workflow
1. **Read each source carefully.** Note key findings, methodology, limits.
2. **Code findings.** Initial open codes per source.
3. **Identify themes** by clustering codes across sources.
4. **Refine themes** through re-reading.
5. **Translate** findings into the theme framework.
6. **Build evidence table:** theme × source × finding type + quality.
7. **Identify gaps:** themes with thin evidence; themes implied but not studied.
## 3. The evidence table
A core synthesis artifact:
| Theme | Source A | Source B | Source C | ... |
|-------|----------|----------|----------|-----|
| Theme 1 | Supports (high quality) | Supports (med quality) | Partial | |
| Theme 2 | Not addressed | Disconfirms (high quality) | Supports | |
| Theme 3 | Supports (low quality) | Supports (high quality) | | |
Allows quick assessment of:
- Which themes are well-supported (multiple high-quality sources)
- Which themes are contested (mixed evidence)
- Which themes are gaps (single source, low quality)
- Which sources are central (touch many themes)
## 4. Identifying gaps
Gaps emerge from:
- **Themes with no high-quality evidence**
- **Themes implied across sources but not directly studied**
- **Methodological gaps** (e.g., no longitudinal data)
- **Demographic gaps** (e.g., no data on minority populations)
- **Geographic gaps** (e.g., only Western data)
- **Temporal gaps** (e.g., pre-2020 data only)
Document gaps honestly. They drive future research and signal what's
known vs unknown.
## 5. Communicating synthesis
### Per theme, report:
- **Statement of the finding** (what is true across the literature)
- **Evidence strength** (number + quality of supporting sources)
- **Disconfirming evidence** (sources that contradict)
- **Notable studies** (citations to the most rigorous / influential)
- **Gaps** within the theme
### Cross-cutting:
- Methodological assessment of the field
- Where consensus exists; where it doesn't
- Limitations of the review itself
## 6. Citation management
### Tools
- Zotero (free, open-source)
- Mendeley (free with paid tiers)
- EndNote (paid, university-licensed)
- Papers
- Paperpile (web-based)
Pick one early; consistency matters more than features.
### What to track per citation
- Authors (full names)
- Title
- Year
- Venue (journal / conference)
- DOI (essential for stability)
- Abstract
- Keywords
- Your notes
- Themes tagged
- Quality / relevance score
### Format styles
- APA (psychology, education)
- MLA (humanities)
- Chicago (history, social sciences)
- Vancouver (medicine)
- IEEE / ACM (computer science)
- Field-specific journal styles
Stick to one style consistently throughout the document.
## 7. Citation hygiene
### Always:
- Read what you cite (don't cite from secondary sources)
- Cite specifically (page numbers for direct quotes)
- Cite in context (was this from a critical or supportive paper?)
- Use the original publication when possible
### Never:
- Citation chain without reading (you propagate errors)
- Cite to look credible without engagement
- Cite predatory journals
- Inflate citation lists
## 8. The "shoulders of giants" practice
Build on prior work explicitly:
- For each major claim, cite the foundational work
- Acknowledge competing schools of thought
- Distinguish your contribution from prior work
- Cite critics, not just supporters
## 9. Avoiding cite-and-run
A common failure: citing 50 papers but engaging with 5.
- For each cited source: what does it contribute?
- If you can't say, don't cite it
- "Smith (2020) found X" beats "(Smith 2020; Jones 2019; Lee 2018)"
- Use parenthetical citations for background; use named citations for substantive engagement
## 10. Tracking citation cycles
Watch for:
- **Self-citation cycles:** the same small group cites only each other
- **Citation inflation:** seemingly supportive cites that actually don't engage
- **Echo-chamber citation:** all citations from the same school of thought
- **Citation drift:** cited claim differs from original
Run periodic checks: pull cited sources; verify the cite matches.
## 11. Synthesizing across disciplines
Cross-disciplinary synthesis is harder:
- Different vocabularies for the same concept
- Different quality conventions
- Different methodological norms
- Different publication venues
Approach:
- Build a translation glossary
- Note methodological norms per discipline
- Cite within-discipline experts to validate cross-walks
- Be explicit about discipline-specific limitations
## 12. The write-up
### Structure (typical):
- Introduction (the question, why it matters)
- Methods (search, criteria, synthesis approach)
- Results (themes, with evidence and citations)
- Discussion (what's known, what's not, implications)
- Limitations (of the review)
- Conclusions (clear, evidence-based)
- References (complete, accurate)
### Style:
- Lead with the answer; build to support
- Use the literature; don't recite it
- Surface tensions and gaps honestly
- Synthesize rather than summarize source-by-source
## 13. Common pitfalls
- **Source-by-source summary instead of synthesis.** Reader gets a bibliography, not insight.
- **No evidence table.** Reader can't assess support strength.
- **Hidden methodology.** Reader can't reproduce.
- **Conclusions exceeding evidence.** Overreach.
- **No engagement with disconfirming evidence.** Bias.
- **Citation inflation.** Quality > quantity.
- **Citation chains.** Verify against originals.
- **No discussion of review limitations.** Reader can't calibrate trust.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |