latte-review-guide
安装方式
先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。
用 AI 助手安装
复制提示词,让助手先核对固定版本和依赖。
查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "latte-review-guide":https://skillpi.com/zh/skills/deepskill-c3ae94513550877cde2c8d762a2d96b181f68b6ded86a259f6b3bb0fd26f6878。安装前核对上游来源和版本:https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/research/paper-review/latte-review-guide/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。
此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。
npx skills add https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/tree/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/research/paper-review/latte-review-guide --skill latte-review-guide来源与配置
查看 GitHub 源码brycewang-stanford/Auto-Empirical-Research-Skills固定提交 bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba
配置方式待核实
请查看原始来源中的适用工具、依赖和权限要求。
安装实测待补充
本条目面向文献与学术工具场景,提供“latte-review-guide”技能的目录入口。公开标签:learning、文献综述、AI代理、PRISMA。请通过固定上游版本核对实际功能、依赖与权限;此页为 Skillπ 独立编写的用途介绍和原站链接,不代表运行安全保证。原文与附件依各自许可开放。
文件浏览
---
name: latte-review-guide
description: "Automate systematic literature reviews with LatteReview AI agents"
metadata:
openclaw:
emoji: "☕"
category: "research"
subcategory: "paper-review"
keywords: ["LatteReview", "systematic review", "literature screening", "AI review", "title screening", "PRISMA"]
source: "https://github.com/PouriaRouzrokh/LatteReview"
---
# LatteReview Guide
## Overview
LatteReview is a low-code Python package that uses AI agents to automate systematic literature reviews. It handles title/abstract screening, full-text assessment, data extraction, and PRISMA-compliant reporting — tasks that typically consume hundreds of researcher-hours. Supports multiple LLM backends (Anthropic, OpenAI, local models).
## Installation
```bash
pip install lattereview
```
## Core Workflow
### Step 1: Initialize Review
```python
from lattereview import ReviewProject
# Create a new review project
project = ReviewProject(
name="ML in Medical Imaging Review",
research_question="What deep learning architectures are used for "
"medical image segmentation?",
inclusion_criteria=[
"Uses deep learning for medical image segmentation",
"Published in peer-reviewed venue",
"Reports quantitative evaluation metrics",
],
exclusion_criteria=[
"Review/survey articles",
"Non-English publications",
"Conference abstracts only",
],
)
```
### Step 2: Import Papers
```python
# Import from various sources
project.import_papers("scopus_export.csv", source="scopus")
project.import_papers("pubmed_export.csv", source="pubmed")
# Or from a DataFrame
import pandas as pd
df = pd.read_csv("papers.csv")
project.import_from_dataframe(df,
title_col="title",
abstract_col="abstract",
year_col="year",
)
print(f"Imported {project.total_papers} papers")
```
### Step 3: AI Screening
```python
from lattereview.agents import ScreeningAgent
# Configure screening agent
screener = ScreeningAgent(
llm_provider="anthropic",
model="claude-sonnet-4-20250514",
criteria=project.inclusion_criteria,
exclusion=project.exclusion_criteria,
)
# Title/abstract screening
results = screener.screen(
project.papers,
mode="title_abstract",
confidence_threshold=0.7,
)
# Results include: decision, confidence, reasoning
for paper in results[:3]:
print(f"{paper.title}")
print(f" Decision: {paper.decision} "
f"(confidence: {paper.confidence:.2f})")
print(f" Reason: {paper.reasoning}")
```
### Step 4: Data Extraction
```python
from lattereview.agents import ExtractionAgent
extractor = ExtractionAgent(
llm_provider="anthropic",
fields={
"architecture": "Deep learning architecture used",
"dataset": "Medical imaging dataset",
"modality": "Imaging modality (CT, MRI, X-ray, etc.)",
"dice_score": "Best Dice similarity coefficient reported",
"sample_size": "Number of images/patients",
},
)
extracted = extractor.extract(project.included_papers)
# Export structured data
extracted.to_csv("extracted_data.csv")
```
### Step 5: Generate Report
```python
# PRISMA flow diagram
project.generate_prisma_diagram("prisma.png")
# Summary statistics
summary = project.summarize()
print(f"Screened: {summary['screened']}")
print(f"Included: {summary['included']}")
print(f"Excluded: {summary['excluded']}")
```
## Configuration
```python
# Use different LLM providers
screener = ScreeningAgent(
llm_provider="openai",
model="gpt-4o",
)
# Local models via Ollama
screener = ScreeningAgent(
llm_provider="ollama",
model="llama3",
base_url="http://localhost:11434",
)
```
## Dual-Reviewer Mode
```python
# Simulate dual-reviewer screening for reliability
results = screener.dual_screen(
project.papers,
models=["claude-sonnet-4-20250514", "gpt-4o"],
agreement_threshold=0.8,
)
# Papers with disagreement flagged for human review
conflicts = [p for p in results if p.agreement < 0.8]
print(f"{len(conflicts)} papers need human adjudication")
```
## Use Cases
1. **Systematic reviews**: PRISMA-compliant literature reviews
2. **Scoping reviews**: Rapid evidence mapping
3. **Meta-analysis preparation**: Structured data extraction
4. **Grant applications**: Quick literature landscape assessment
## References
- [LatteReview GitHub](https://github.com/PouriaRouzrokh/LatteReview)
- [LatteReview Documentation](https://lattereview.readthedocs.io/)
- Rouzrokh, P. et al. (2024). "LatteReview: AI-Assisted Systematic Literature Reviews."
版本历史
-
v2026.07
GitHub ↗
原版更新说明
## Highlights - **First tagged release.** Every gate is green: `make check` (validate + 183 unit tests + eval-harness + numeric benchmarks), validate-catalog, quality-evals, and OpenSSF Scorecard on `main`. - **Rigor double loop at 13 method families.** Each family closes taxonomy tag → eval scenario (does the agent avoid the method's known traps?) → deterministic numeric benchmark (does it recover the true effect?). New this cycle: heterogeneous treatment effects (CATE) and quantile treatment effects (QTE). - **Paper-WorkFlow flagship at 29/29 executable rigor gates**, now merged to its `main` with a drift-gated README badge — the credibility artifact is checked by CI, so it can never silently go stale ([RIGOR.md](https://github.com/brycewang-stanford/Paper-WorkFlow/blob/main/RIGOR.md)). - **1,150 vendored skills / 69 collections**, all license-audited, security-scanned, tracked in a machine-readable catalog with CI-enforced freshness, and importable as a single whole-repo router skill. - **Hardened weekly upstream syncs**: synced trees must pass the full gate before merge; this cycle both sync PRs were verified, patched (StatsPAI's SkillOpt execution-gate card restored), and merged. - **First-party plugin marketplace** (Python / R / Stata empirical-analysis plugins + guide) shipped in the previous cycle, now under a tagged version. ## Known Follow-Ups - Ecosystem benchmark comparison (AERS vs econometrics agents on the shared benchmark suite) — [PLAN-2026-07 week 3](https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/main/docs/PLAN-2026-07.md). - Decomposition (Oaxaca–Blinder) and shift-share (Bartik) method families (weeks 3–4). - GitHub Pages catalog site, README demo screenshots, five-language README consistency script (week 4). ## Catalog - Top-level collections: **69** - `SKILL.md` files: **1150** - Cataloged agent tools: **335** ([`tools/CATALOG.md`](../tools/CATALOG.md)) ## Methodological rigor - Method families in the coverage map: **13** ([`RIGOR_COVERAGE.md`](RIGOR_COVERAGE.md)) - Eval scenarios: **26** (138 rubric items, 122 auto-checkable) ([`../eval-harness/`](../eval-harness/README.md)) - Benchmark tasks: **13** (70 gold checks, 58 required) ([`../benchmark/`](../benchmark/README.md)) ### Method families (skills tagged) - `iv`: 27 - `rdd`: 25 - `panel-fe`: 19 - `staggered-did`: 17 - `bayesian`: 13 - `did`: 10 - `event-study`: 10 - `matching`: 10 - `synthetic-control`: 10 - `cate`: 8 - `dml`: 8 - `survival`: 3 - `quantile`: 2 ### Eval scenarios by category - `causal-identification`: 14 - `citation-hygiene`: 1 - `reproducibility`: 2 - `research-integrity`: 1 - `runtime-safety`: 1 - `statistical-validity`: 2 - `writing-compliance`: 3 - `writing-style`: 2 ## License buckets - AGPL-3.0: 1 - Apache-2.0: 1 - CC-BY-NC-4.0 (non-commercial): 1 - CC-BY-SA-4.0: 1 - CC-BY-SA-4.0 (repository default): 4 - GPL-3.0: 1 - MIT: 35 - MIT Non-Commercial: 1 - Mixed (per-skill; see folder): 2 - UNKNOWN - check upstream: 22 ## Quality gates - `make check` — catalog/provenance/audit/eval freshness, repo hygiene, Paper-WorkFlow gates, stdlib unit tests, eval-harness lint + graded smoke, and the strict numeric benchmark (reference-gold drift fails the build). - External links and OpenSSF Scorecard run on their scheduled workflows.
当前收录版本
0.0.1
质量评测
满分 100。前四项来自 DeepSkill 收录的质量报告。
评分依据
高价值/高频刚需:直接解决研究人员系统文献综述耗时痛点
评分依据
触发词精准完整:覆盖LatteReview、systematic review、literature screening等多种变体
评分依据
Phase/步骤清晰:Step 1-5完整流程划分
评分依据
工具来源明确:GitHub地址清晰
计算方式
取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。
权限与使用前检查
根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。
-
安装或获取外部依赖
使用官方来源并锁定版本,在隔离环境检查安装脚本。当前文档检查不覆盖后续下载的代码。
主文档提及 · 第 22 行
文档引用的域名或地址
这些地址可能用于文档链接或示例,不是已批准的联网清单。使用前核对实际目的地及发送的数据。
- github.com
- lattereview.readthedocs.io
- localhost
安全评级与访问范围
risk · 静态证据核验,不代表运行安全保证。
使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。
用户讨论
还没有已核实的点评。