Pubmed Edirect

Search and retrieve literature from PubMed using NCBI's EDirect command-line tools. ⚠️ Advanced skill requiring manual installation.

Security Scan

The skill's requirements, instructions, and included scripts are consistent with installing and using NCBI's EDirect tools — nothing obvious or disproportionate to the stated purpose, though you should still review installer scripts and run in an isolated/test environment.

Purpose & Capability

The name/description match the actual requirements and behavior: the skill requires the EDirect binaries (esearch, efetch, elink, xtract, einfo, efilter) and provides documentation and scripts that invoke those tools. No unrelated binaries, services, or credentials are requested.

Instruction Scope

SKILL.md and the included docs/scripts confine instructions to installing, configuring, and running EDirect and to using OpenClaw's exec to run EDirect commands. The docs explicitly warn about reviewing installer scripts, avoiding piping remote scripts to shell, and recommend isolation. There are OpenClaw exec examples that run commands and save results to the agent workspace — expected for a local CLI integration. No instructions request reading unrelated system secrets or exfiltrating data to third-party endpoints.

Install Mechanism

There is no automated install spec — installation is manual and points to the official NCBI FTP/HTTP (ftp.ncbi.nlm.nih.gov / https://www.ncbi.nlm.nih.gov/books/NBK179288/). The docs explicitly instruct to download-and-review rather than piping to shell. This is a proportionate and lower-risk approach compared to blind downloads from untrusted servers.

Credentials

Only optional environment variables are documented (NCBI_API_KEY and NCBI_EMAIL) and they are appropriate for EDirect (rate-limiting and identification). The skill does not request unrelated secrets or config paths. The INSTALL.md does note that system package installation (Perl modules) may require sudo — that is a legitimate prerequisite for this toolchain but is a privilege escalation point you should manage consciously.

Persistence & Privilege

The skill does not request always:true or any ability to alter other skills or global agent configuration. It suggests adding EDirect to PATH and updating shell rc files only for the user, which is standard for CLI tools. OpenClaw integration examples create workspace files and cron jobs under the workspace; these are expected for automation but should be deliberately enabled by the user.

Assessment

This skill appears coherent and matches its stated purpose, but follow the documentation and good security hygiene before installing: review the official install-edirect.sh script you download (do not pipe remote scripts to bash), preferably test the installer in a container or VM, avoid running installers as root unless necessary, and audit the included scripts before running them. If you plan to use cron/automation in OpenClaw, limit workspace permissions and avoid storing sensitive credentials in shared shell rc files; store an NCBI API key only in a place you control and rotate it if needed. If you have low trust in the source of this registry entry, consider using a containerized ncbi/edirect image or perform installation on an isolated test environment first.

PubMed EDirect Skill

Search and retrieve literature from PubMed using NCBI's EDirect command-line tools.

⚠️ Security Advisory

Important: This skill requires installation of external command-line tools. The installation process involves:

1. External script execution: Downloading and executing installation scripts from the official NCBI FTP server
2. System modifications: Adding directories to your PATH environment variable
3. Permission requirements: May require installation of Perl modules and dependencies

Before installation, you must:

1. Review the installer script content after downloading
2. Confirm the source is trustworthy (official ftp.ncbi.nlm.nih.gov domain)
3. Validate in a test environment
4. Understand all commands that will be executed

Overview

This skill provides access to PubMed and other NCBI databases through the official EDirect (Entrez Direct) utilities. EDirect is a suite of programs that provide access to the NCBI's suite of interconnected databases (publication, sequence, structure, gene, variation, expression, etc.) from Unix terminals.

Note: This is a local installation skill – all tools run directly on your system without Docker or containerization. Follow the INSTALL.md guide for local setup.

Structure

The skill is organized into the following files:

• INSTALL.md - Installation and configuration guide
• BASICS.md - Basic usage and common commands
• ADVANCED.md - Advanced techniques and complex queries
• EXAMPLES.md - Practical usage examples
• REFERENCE.md - Quick reference (field qualifiers, formats, etc.)
• OPENCLAW_INTEGRATION.md - OpenClaw-specific usage guide
• scripts/ - Useful bash scripts for common tasks

Quick Start

1. Read the installation guide: Review INSTALL.md for secure installation steps
2. Manually install EDirect:

   # Step 1: Download the script
   wget -q https://ftp.ncbi.nlm.nih.gov/entrez/entrezdirect/install-edirect.sh
   
   # Step 2: Review content (important for security)
   less install-edirect.sh
   
   # Step 3: Execute installation
   ./install-edirect.sh
   

3. Verify installation:

   esearch -db pubmed -query "test" -retmax 1
   

4. Explore examples: Check EXAMPLES.md

Core Tools

The skill provides access to EDirect tools through OpenClaw's exec capability:

• esearch - Search databases
• efetch - Retrieve records
• elink - Find related records
• efilter - Filter results
• xtract - Extract data from XML
• einfo - Get database information

Databases Supported

EDirect supports numerous NCBI databases including:

• pubmed - Biomedical literature
• pmc - PubMed Central full-text articles
• gene - Gene information
• nuccore - Nucleotide sequences
• protein - Protein sequences
• mesh - Medical Subject Headings
• And many more...

Key Features

• Command-line access to NCBI databases
• Pipeline architecture using Unix pipes
• Structured data extraction with XML parsing
• Batch processing capabilities
• Cross-database linking between records

Getting Help

• Use -help with any EDirect command: esearch -help
• Consult the official documentation
• Check troubleshooting in installation guide

Included Scripts

The scripts/ directory contains ready-to-use bash scripts:

batch_fetch_abstracts.sh

Fetch abstracts for a list of PMIDs with error handling and rate limiting.

./scripts/batch_fetch_abstracts.sh pmids.txt abstracts/ 0.5

search_export_csv.sh

Search PubMed and export results to CSV with metadata.

./scripts/search_export_csv.sh "CRISPR [TIAB]" 100 results.csv

publication_trends.sh

Analyze publication trends over time with visualization.

./scripts/publication_trends.sh "machine learning" 2010 2023 trends.csv

Security Best Practices

1. Script Review

# Always download first and review scripts
wget -q SOURCE_URL -O script.sh
less script.sh  # or cat script.sh | head -50
# Execute only after review
./script.sh

2. Environment Isolation

• Running in Docker containers provides isolation
• Use virtual machines for testing
• Set up dedicated user accounts

3. Least Privilege

• Do not run as root
• Set appropriate file permissions
• Use dedicated directories for data

4. Network Controls

• Configure firewall rules
• Use proxies for controlled access
• Monitor network traffic

Notes

Important: This skill requires manual installation and configuration. All installation steps require explicit user confirmation and execution.

This skill provides command-line access to NCBI databases through local installation of EDirect tools.