cluade-scientific-writer-wrapper

AI-powered scientific writing tool. Generate publication-ready papers, grant proposals, posters, and more with real-time research and verified citations.

Security Scan

The skill's stated purpose (AI scientific writing using Claude) mostly matches its requirements, but the runtime instructions ask you to pip-install an external package (unverified), reference an extra API key (OPENROUTER) that isn't declared, and instruct reading/writing local files—these mismatches and the unpinned external install elevate risk.

Purpose & Capability

Name/description align with requiring python3, pip, and an ANTHROPIC_API_KEY for Claude. The skill legitimately needs to read user data files and produce outputs locally for writing tasks. Minor mismatch: SKILL.md also mentions an optional OPENROUTER_API_KEY for research lookups, but that env var is not declared in the registry metadata.

Instruction Scope

The SKILL.md tells the agent to pip install the 'scientific-writer' package at runtime and to read data files from a local data/ folder and write outputs to writing_outputs/. It also suggests creating a .env file to store API keys. These instructions permit installing and executing third-party code and reading/writing arbitrary local files—appropriate for this tool but increasing exposure. The file/secret handling (echoing keys into .env) is a minor operational risk. The reference to OPENROUTER_API_KEY is an undeclared env var the agent may try to access.

Install Mechanism

There is no formal install spec in the registry; instead SKILL.md instructs users/agents to run 'pip install scientific-writer'. Installing from PyPI (or default pip sources) is a moderate risk because it fetches and executes external code without a pinned version, checksum, or verified source. The package.json points to a GitHub repo, but the registry did not include code to inspect. No download-from-odd-URL flags, but unverified pip installs deserve caution.

Credentials

Only ANTHROPIC_API_KEY is declared as required, which is proportional for a Claude-based writer. The SKILL.md's optional OPENROUTER_API_KEY is not declared in metadata — a mismatch. The instructions also encourage storing secrets in a .env file (local persistent secret storage), which may be undesirable depending on your environment.

Persistence & Privilege

The skill does not request always:true, does not claim to modify other skills, and has no system-level persistence declared. It will read and write local project files (data/ and writing_outputs/), which is expected for a writer tool.

What to consider before installing

This skill mostly does what it says, but exercise caution before installing or running it. The SKILL.md asks you (or the agent) to 'pip install scientific-writer'—that pulls code from an external package repository without a pinned version or checksum. Before installing, verify the package source (check the linked GitHub repo, review the code, and prefer a specific version). Be aware it will read files from data/ and write outputs to writing_outputs/, and it asks you to provide your ANTHROPIC_API_KEY (required) and optionally an OPENROUTER_API_KEY (undeclared in metadata). Avoid echoing secrets into plaintext .env files in shared environments. If you can't review the package code, consider running it in a sandboxed environment (VM or container) or asking the author/maintainer for a signed release or source tarball you can inspect.

Scientific Writer

AI-powered scientific writing tool that combines deep research with well-formatted outputs.

Installation

pip install scientific-writer

Prerequisites

• Python 3.10-3.12
• ANTHROPIC_API_KEY (required)
• OPENROUTER_API_KEY (optional for research lookup)

Set API keys:

export ANTHROPIC_API_KEY='your_key'
# or create .env file
echo "ANTHROPIC_API_KEY=your_key" > .env

Usage

As Python API

import asyncio
from scientific_writer import generate_paper

async def main():
    async for update in generate_paper(
        query="Create a Nature paper on CRISPR gene editing...",
        data_files=["editing_efficiency.csv", "western_blot.png"]
    ):
        if update["type"] == "progress":
            print(f"[{update['stage']}] {update['message']}")
        else:
            print(f"✓ PDF: {update['files']['pdf_final']}")

asyncio.run(main())

Via OpenClaw exec

# Run scientific writer
python3 -c "
import asyncio
from scientific_writer import generate_paper

async def main():
    async for update in generate_paper(
        query='Create a paper on your topic...',
        data_files=[]
    ):
        print(update)

asyncio.run(main())
"

Available Skills (when used as plugin)

• scientific-schematics - AI diagram generation (CONSORT, neural networks, pathways)
• research-lookup - Real-time literature search
• peer-review - Systematic manuscript evaluation
• citation-management - BibTeX and reference handling
• clinical-reports - Medical documentation
• research-grants - NSF, NIH, DOE proposal support
• scientific-slides - Research presentations
• latex-posters - Conference poster generation
• hypothesis-generation - Scientific hypothesis development

Output

• Scientific papers (Nature, Science, NeurIPS format)
• Grant proposals (NSF, NIH, DOE)
• Conference posters (LaTeX beamerposter)
• Literature reviews
• Clinical reports

Notes

• Requires ANTHROPIC_API_KEY for Claude to work
• Place data files in data/ folder (images → figures/, data → data/)
• Outputs saved to writing_outputs/