# Literature Search

> Use when a ScanSci user wants to search, compare, or answer questions from a bound local knowledge library with traceable evidence.

- Skill: `rimagination/literature-search` (Agent Skill)
- Install (CLI): `npx skillmds@latest add rimagination/literature-search`
- Raw SKILL.md: https://api.skillmd.com/api/skills/rimagination/literature-search/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-search

---


# Literature Search

Use ScanSci's Knowledge mode for this task.

1. Restate the research question as 2–4 compact retrieval queries.
2. Use the knowledge-library retrieval results supplied by ScanSci; do not invent search hits or citations.
3. Prefer passages that directly answer the question, then compare agreement, disagreement, methods, scope, and limitations.
4. Attach each substantive claim to the evidence identifiers returned by ScanSci.
5. Distinguish source-backed statements, synthesis, and unknowns.
6. If evidence is insufficient, say what is missing and propose a narrower next query.

Do not claim to have searched the web or opened a paper unless ScanSci supplied a corresponding tool result.