← Back to Skills Marketplace
110
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install arxiv-query
Description
Search and summarize papers from ArXiv. Use when the user asks for the latest research, specific topics on ArXiv, or a daily summary of AI papers.
README (SKILL.md)
ArXiv Query
This skill interacts with the ArXiv API to find and summarize the latest research papers.
Capabilities
- Search: Find papers by keyword, author, or category.
- Summarize: Fetch the abstract and provide a concise summary.
- Save to Memory: Automatically record summarized papers to
memory/RESEARCH_LOG.mdfor long-term tracking. - Deep Dive: Use
web_fetchon the PDF link to extract more details if requested.
Workflow
- Use
scripts/search_arxiv.sh "\x3Cquery>"to get the XML results. - Parse the XML (look for
\x3Centry>,\x3Ctitle>,\x3Csummary>, and\x3Clink title="pdf">). - Present the findings to the user.
- MANDATORY: Append the title, authors, date, and summary of any paper discussed to
memory/RESEARCH_LOG.md. Use the format:### [YYYY-MM-DD] TITLE_OF_PAPER - **Authors**: Author List - **Link**: ArXiv Link - **Summary**: Brief summary of the paper and its relevance.
Examples
- "Busca los últimos papers sobre LLM reasoning en ArXiv."
- "Dime de qué trata el paper con ID 2512.08769."
- "Hazme un resumen de las novedades de hoy en ArXiv sobre agentes."
Resources
scripts/search_arxiv.sh: Direct API access script.
Usage Guidance
This ArXiv skill is mostly coherent with its stated purpose but has two things you should consider before installing: (1) the bundled script uses curl but the skill declares no required binaries — ensure curl is present or update the skill to declare it; (2) the SKILL.md forces the skill to append every discussed paper to memory/RESEARCH_LOG.md (persistent storage) without asking the user. If you want to use it, ask the author to make saving optional or require explicit consent before writing to memory, and to declare required binaries. Also consider restricting PDF deep‑fetching (it will download external PDFs) and validate or sanitize query input. If you don't want automatic persistent logging, do not install or modify the skill to prompt before saving.
Capability Analysis
Type: OpenClaw Skill
Name: arxiv-query
Version: 1.0.0
The skill contains a shell injection vulnerability in `scripts/search_arxiv.sh`, where the `$QUERY` variable is passed directly into a `curl` command within double quotes without sanitization. While the stated purpose of searching ArXiv is legitimate and no evidence of intentional malice or data exfiltration was found, the lack of input handling allows for potential command execution if the agent provides a crafted query.
Capability Assessment
Purpose & Capability
The skill's name/description match its behavior: it queries the ArXiv API and summarizes papers. However, the SKILL.md and included script rely on curl to fetch results even though the declared requirements list no required binaries. Also the SKILL.md promises automatic saving to memory (RESEARCH_LOG.md) but no required config path was declared; these mismatches are noteworthy.
Instruction Scope
The instructions require parsing XML from the API and present results (expected). They also mandate appending any discussed paper (title, authors, date, summary) to memory/RESEARCH_LOG.md unconditionally ('MANDATORY'), which causes persistent writes without explicit per-action consent. The skill also advises using web_fetch to download PDFs for deep dives (external network fetches). The combination of mandatory persistent logging and autonomous network fetches expands scope beyond a simple read-only query.
Install Mechanism
There is no install spec and only a small shell script is bundled; the script performs a curl request to export.arxiv.org. This is low-risk compared to arbitrary remote installs. No third-party package downloads or obscure URLs are used.
Credentials
The skill requests no credentials (appropriate), but it instructs the agent to write into a persistent memory file (memory/RESEARCH_LOG.md) without declaring that config path or asking for user permission. Also the script implicitly requires curl even though no binaries were declared. These omissions reduce transparency about what resources the skill uses and modifies.
Persistence & Privilege
The skill is not always-enabled and does not request special platform privileges, but it explicitly requires mandatory persistent writes to the agent's memory store for every discussed paper. If the agent invokes this skill autonomously, it may add entries to memory without explicit user confirmation. That persistent behavior is a meaningful privilege and should be made optional or gated by user consent.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install arxiv-query - After installation, invoke the skill by name or use
/arxiv-query - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the arxiv-query skill.
- Search and summarize research papers from ArXiv by keyword, author, or category.
- Automatically records summarized papers to memory for long-term tracking.
- Provides abstracts with concise summaries and optionally extracts more details from PDF links.
- Includes Spanish-language example queries.
Metadata
Frequently Asked Questions
What is Arxiv?
Search and summarize papers from ArXiv. Use when the user asks for the latest research, specific topics on ArXiv, or a daily summary of AI papers. It is an AI Agent Skill for Claude Code / OpenClaw, with 110 downloads so far.
How do I install Arxiv?
Run "/install arxiv-query" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Arxiv free?
Yes, Arxiv is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Arxiv support?
Arxiv is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Arxiv?
It is built and maintained by ClawKK (@codekungfu); the current version is v1.0.0.
More Skills