↓Skip to main content

Run a literature review from your desk

Playbook
·3 mins
From scoped question through citations and synthesis, every layer stacks into a defendable literature review.

My first AI-assisted lit pass read beautifully until I clicked the DOIs. Two 404’d. One paper matched the title but contradicted the claim beside it. The model had written confident prose on vapor.

I fixed it with structure: scoped searches against the real indexes, DOIs and NCT IDs in the table, synthesis tied to what those queries actually returned—no more treating one chat thread as a literature database.

When I reach for this #

I need a structured survey of published research on a topic — for a project proposal, a competitive landscape review, or to answer a question like “what’s the current evidence for X?” and I need the answer to hold up under scrutiny.

What I need before starting #

  • Claude Code with the /literature-review skill installed
  • MCP connectors for PubMed and bioRxiv enabled (these come from the claude.ai connector set)
  • A scoped research question — not “tell me about cancer” but “what are the current Phase II+ trials combining X and Y in indication Z?”

What I do #

1. Scope the question #

The quality of the review depends entirely on the quality of the question. I write the research question as a single sentence with:

  • The population or domain
  • The intervention or method
  • The outcome or comparison I care about
  • A time window if relevant

Example: “What published evidence exists for NRF2 pathway modulation as a therapeutic strategy in neurodegeneration, published 2020–2026?”

2. Launch the review #

/literature-review

The skill asks clarifying questions about scope, then dispatches parallel searches across:

  • PubMed — peer-reviewed published literature
  • bioRxiv/medRxiv — preprints (not peer-reviewed, but often months ahead of publication)
  • ClinicalTrials.gov — active and completed trials
  • Semantic Scholar — citation network and related papers

Each source is searched by a separate subagent running in parallel. The searches complete in 2-3 minutes total.

3. Review the synthesis #

The skill produces a structured output:

  • Summary — key findings across all sources, with source counts
  • Evidence table — each finding with its source, DOI/NCT ID, date, and relevance rating
  • Gaps — what the search didn’t find, or areas where evidence is thin
  • Methodology notes — search terms used, databases queried, date ranges

Every citation links to a real DOI or trial ID. I can click through and verify.

4. Refine if needed #

Too broad — tighten terms, indication, or time window, then rerun.

Too narrow — relax one constraint at a time so the search can breathe.

Cadence — parallel searches keep each iteration in minutes, not afternoons.

What goes wrong #

  • Question too broad — “what’s known about inflammation” returns hundreds of results with no useful synthesis. Narrow to a specific mechanism, indication, or time window
  • Preprint bias — bioRxiv results aren’t peer-reviewed. The synthesis notes which sources are preprints, but I still flag them separately when citing in formal documents
  • Missing databases — the skill searches the major life sciences databases. For chemistry-specific literature (reaction databases, patent searches), I supplement with targeted searches
  • Connector issues — if a database search fails silently, the synthesis may be incomplete. Check the methodology notes to confirm all databases were actually queried

Notes #

Deep read — upload short-listed PDFs to NotebookLM for source-grounded Q&A on methods. This playbook finds candidates; NotebookLM is where I do close reading on the full text.

Share-out — export synthesis to Quarto (or similar) so the evidence table ships as a sortable artifact:

  • Posit Connect
  • internal static hosting
  • a private GitHub Pages branch

Match whatever surface your org treats as the living report home.