academic-research-hub

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the download of academic papers in PDF format from trusted, well-known repositories such as arXiv and Semantic Scholar using the requests library.
  • [EXTERNAL_DOWNLOADS]: The documentation references clawhub.ai for the installation of the OpenClawCLI, which is a native resource associated with the platform vendor.
  • [PROMPT_INJECTION]: The skill processes external data (academic abstracts and titles) which could serve as a vector for indirect prompt injection, although this is a standard risk for research-focused agents.
  • Ingestion points: Research metadata is ingested from external APIs in scripts/research.py.
  • Boundary markers: The script formats results using structured text and JSON, but it does not implement specific delimiters or 'ignore' instructions for the agent context.
  • Capability inventory: The skill includes file-writing capabilities to save search results and PDFs, and network operations to call academic APIs.
  • Sanitization: Basic filename cleaning is performed for file system safety, and text is truncated for display purposes, but no semantic sanitization of retrieved abstracts is present.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 12:25 PM