Release20260324
v1.0.0Academic literature discovery and citation network analysis. Multi-source search across arXiv, DBLP, Semantic Scholar, and Google Scholar. Build citation net...
⭐ 0· 60·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (OpenPaperGraph: search, PDF parsing, citation graphs, Zotero import, LLM summaries) matches the included code (search/services for arXiv, S2, DBLP, Google Scholar, PDF parsing, Zotero, graph builder, server). Required binaries/env are optional and appropriate for the claimed features.
Instruction Scope
SKILL.md instructs installing a small set of Python packages, running the CLI from SKILL_DIR, and using local PDF files and graph JSON exports. It does not instruct reading or exfiltrating unrelated system files. Allowed tools (Read/Write/Edit/Bash) are consistent with the need to read PDFs, write JSON, and run the CLI.
Install Mechanism
There is no registry install spec, but an included install.sh performs pip installs (httpx, pymupdf, scholarly) and creates a symlink under ~/.claude/commands/opg for global install. Installing via pip and creating a user-level symlink is expected for this kind of skill, but running install.sh will modify the user environment; review it before executing and prefer a virtualenv.
Credentials
No required environment variables are declared. SKILL.md documents many optional LLM provider keys and a recommended S2_API_KEY for rate limits. Those optional keys are reasonable for the optional LLM summarization and S2 rate-limit avoidance; they are not required for core functionality.
Persistence & Privilege
The installer can register the skill globally by creating a symlink in ~/.claude/commands/opg (user-level change). always:false and no elevated system privileges are requested. Autonomous invocation (default) is allowed by platform norms; combined with local file writes and an optional local HTTP server (serve command), the skill will persist graph JSON files and may run a local server — expected for the feature set.
Assessment
This project is internally consistent and implements the advertised features. Before installing: (1) inspect install.sh (it will pip install packages and may create a symlink in ~/.claude if you choose global install); prefer using a Python virtualenv to avoid system-wide changes; (2) you do not need to set any API keys to use core functionality, but only provide LLM or S2 keys if you understand the implications (these keys grant the tool access to external services); (3) the tool may scrape Google Scholar (via the scholarly package) which can be slow, rate-limited, or against site terms — expect possible blocking; (4) the CLI writes graph JSON files to disk and can start a local HTTP server (serve) — do not expose that server to untrusted networks; (5) if you need higher assurance, review services/llm_client.py and networking code for any unexpected external endpoints before providing credentials or running the tool.Like a lobster shell, security has layers — review code before you run it.
latestvk97e99da3q6nhdwkefwcep79vn83k01x
License
MIT-0
Free to use, modify, and redistribute. No attribution required.