# Recipe Academic Literature Review

> Map the gap between published research and filed patents — a scholarly sweep across Semantic Scholar, arXiv, and OpenAlex aligned against a matching patent search, output centered on the published-versus-protected gap rather than a ranked novelty list. Trigger when the user asks for a literature review, a state-of-the-art survey, or a comparison of academic research against patents.

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

---


# Recipe: Academic Literature Review

Combine academic and patent searches to map what a field has published against
what it has protected.

Prefer `academic` (Semantic Scholar + arXiv) for CS/ML papers and preprints;
prefer `npl` (OpenAlex) for broad cross-disciplinary journal coverage and
open-access filtering.

## Steps

### Step 1: Academic Search

```bash
flowleap --json academic search "<research topic>" --limit 20
flowleap --json academic search "<research topic>" --source arxiv --from-year 2020
```

### Step 2: Widen to OpenAlex NPL

```bash
flowleap --json npl "<research topic>" --from-year 2020 --limit 10
```

### Step 3: Patent Search for the Same Topic

Write the CQL yourself from the topic's discriminating terms (the specific
subject matter, not the field name — method in `flowleap-patent`), probe the
count, then search:

```bash
flowleap --json tools run search_patents query='<self-written CQL>' range=1-1 details=false
flowleap --json patent search --query "<self-written CQL>" --limit 20
```

### Step 4: Synthesize the Gap

Align the academic themes against the patent CPC and assignee clusters. Flag
topics heavily published but lightly patented (open R&D space) and topics
heavily patented but lightly published (crowded IP). Done when each major theme
is classified on the published-versus-protected axis.

## Visual deliverable

To hand off Step 4's published-versus-protected map as a shareable artifact,
follow this recipe's analysis through to the end, then render it with
`recipe-custom-dashboard` — the **filing-trends** template (multi-series
comparison over time) is the closest fit, adapted to plot academic output
against patent filings per theme rather than per applicant. Analysis logic
stays here; the dashboard skill only owns presentation.

## Output

- Academic papers (title, authors, year, source, citation counts)
- Related patents (publication number, title, applicant, date, CPC)
- A published-versus-protected map: themes tagged as open R&D space or crowded IP