PMC Harvest — 技能 — openclaw中文资讯站

技能详情（站内镜像，无评论）

Fetch and retrieve full-text or abstracts of open-access articles from PubMed Central by journal, year, or PMCID using NCBI's public APIs without an API key.

开发与 DevOps

许可证：MIT-0

MIT-0 ·免费使用、修改和重新分发。无需归因。

版本：v1.0.0

统计：⭐ 2 · 577 · 2 current installs · 2 all-time installs

⭐ 2

安装量（当前） 2

🛡 VirusTotal ：良性 · OpenClaw ：良性

Package：angusthefuzz/pmc-harvest

安全扫描（ClawHub）

VirusTotal ：良性
OpenClaw ：良性

OpenClaw 评估

The skill's code, instructions, and requirements are consistent with its stated purpose of harvesting open-access PMC content via NCBI public APIs and do not request unrelated credentials or suspicious network endpoints.

目的

The name/description (PMC harvesting) matches the code and SKILL.md: the package only requires node, uses NCBI E-utilities and OAI-PMH endpoints, implements search, summaries, and full-text retrieval. No unrelated services, credentials, or binaries are requested.

说明范围

SKILL.md and the CLI instruct only running the included Node scripts and using NCBI endpoints. The runtime instructions do not ask the agent to read unrelated files, environment variables, or post data to third-party endpoints. All network calls in the code go to ncbi.nlm.nih.gov / pmc.ncbi.nlm.nih.gov.

安装机制

There is no install spec (instruction-only), and the included code is plain Node.js source that uses only built-in https/zlib modules. No external downloads or archives, package registry installs, or untrusted URLs are present.

证书

The skill requires only the 'node' binary and requests no environment variables, keys, or config paths. This is proportionate for a Node CLI that calls public HTTP APIs.

持久

The skill does not request always:true, does not modify other skills or system-wide settings, and contains no code that persistently installs components or stores credentials. It performs transient network requests only when invoked.

综合结论

This skill appears to do exactly what it claims: search PMC and fetch open-access full-text via NCBI's public APIs. Before installing, ensure you have Node available and that your environment allows outbound HTTPS to ncbi.nlm.nih.gov. If you plan large-scale harvesting, respect NCBI rate limits and usage policies (the code implements a 3 req/sec delay but large batches should be staggered and run off-peak). In an enterprise environment, verify…

安装（复制给龙虾 AI）

将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。

请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「PMC Harvest」。简介：Fetch and retrieve full-text or abstracts of open-access articles from PubMed C…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/angusthefuzz/pmc-harvest/SKILL.md
（来源：yingzhi8.cn 技能库）

SKILL.md

打开原始 SKILL.md（GitHub raw）

---
name: pmc-harvest
description: Fetch articles from PubMed Central using NCBI APIs. Search journals, retrieve full text via OAI-PMH, batch harvest for RAG pipelines. No API key required.
version: 1.0.0
author: Ania
metadata:
  clawdbot:
    emoji: "📚"
    requires:
      bins: ["node"]
---

# PMC Harvest

Fetch full-text articles from PubMed Central using official NCBI APIs.

## Features

- **E-utilities search** — Find articles by journal, year, query
- **OAI-PMH full text** — Retrieve complete article XML (open access only)
- **Batch harvesting** — Process multiple journals at once
- **Abstract fetch** — Lightweight retrieval for review queues
- **No API key required** — Uses public NCBI APIs (rate-limited)

## Usage

```bash
# Search a journal
node {baseDir}/scripts/pmc-harvest.js --search "J Stroke[journal]" --year 2025

# Fetch full text for a specific article
node {baseDir}/scripts/pmc-harvest.js --fetch PMC12345678

# Batch harvest from multiple journals
node {baseDir}/scripts/pmc-harvest.js --harvest journals.json --year 2025

# Test with known journals
node {baseDir}/scripts/pmc-harvest.js --test
```

## Options

| Flag | Description |
|------|-------------|
| `--search <query>` | PMC search query (use journal[name] format) |
| `--year <year>` | Filter by publication year |
| `--max <n>` | Max results (default: 100) |
| `--fetch <pmcid>` | Fetch full text for specific PMCID |
| `--harvest <file>` | Batch harvest from JSON journal list |
| `--test` | Run test with sample journals |

## Programmatic API

```javascript
const pmc = require('{baseDir}/lib/api.js');

// Search
const { count, pmcids } = await pmc.searchJournal('"J Stroke"[journal]', { year: 2025 });

// Get summaries
const summaries = await pmc.getSummaries(pmcids);

// Fetch full text
const { available, xml, reason } = await pmc.fetchFullText('PMC12345678');

// Parse JATS XML
const { title, abstract, body } = pmc.parseJATS(xml);

// Fetch abstract only (lightweight)
const { title, abstract } = await pmc.fetchAbstract('PMC12345678');
```

## Journal Query Examples

```javascript
const queries = {
  'Stroke': '"Stroke"[journal]',
  'Journal of Stroke': '"J Stroke"[journal]',
  'Stroke & Vascular Neurology': '"Stroke Vasc Neurol"[journal]',
  'European Stroke Journal': '"Eur Stroke J"[journal]',
  'BMC Neurology': '"BMC Neurol"[journal]'
};
```

## Limitations

- **OAI-PMH only returns open-access articles** — restricted content unavailable
- **Rate limits** — ~3 requests/second without API key
- **Peak hours** — NCBI recommends avoiding 5AM-9PM ET for large batches

## API Reference

This skill wraps NCBI's official APIs:

- **E-utilities**: `https://eutils.ncbi.nlm.nih.gov/entrez/eutils`
  - `esearch.fcgi` — Search PMC
  - `esummary.fcgi` — Get article metadata
- **OAI-PMH**: `https://pmc.ncbi.nlm.nih.gov/api/oai/v1/mh`
  - `GetRecord` — Fetch full text XML

Full docs: https://www.ncbi.nlm.nih.gov/books/NBK25501/