# Literature Review

> Systematic evidence-bounded literature review across a selected ScanSci library, including scope definition, thematic synthesis, evidence gaps, citation checks, and review-ready output. Use for literature reviews, state-of-the-art surveys, scoping reviews, or research-gap analysis.

- Skill: `rimagination/literature-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add rimagination/literature-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/rimagination/literature-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: Rimagination (https://skillmd.com/u/rimagination)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/rimagination/literature-review

---


# Literature Review

Retrieve first, then synthesize.

1. Write the review question, scope, inclusion and exclusion rules, search date, and intended audience before drafting.
2. Build or reuse a ScanSci evidence store. Search each planned section independently and retain exact evidence IDs, quotes, source documents, and section labels.
3. Organize findings by question, mechanism, method, result, or controversy—not by a sequence of paper summaries.
4. For every major claim, record supporting evidence, contradictory evidence, confidence, and the boundary of the conclusion.
5. Separate consensus, disagreement, missing evidence, and proposed next studies. Treat an absence of retrieval as an uncertainty, not proof of absence.
6. Validate citations and refuse unsupported synthesis when the evidence gate is not met.

Return a review plan, evidence map, section drafts, reference table, limitations, and open questions. Do not pad a review with unverified citations or convert a search snippet into a finding.

This is a ScanSci adaptation of the public K-Dense `literature-review` workflow, aligned with the local sentence-level evidence and citation verifier.