backing-claims

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust workflow for academic integrity. It includes a Python validator (scripts/check_provenance.py) that strictly checks for documentation requirements in BibTeX files rather than performing dangerous operations.
  • [COMMAND_EXECUTION]: The skill utilizes a CLI tool called scholar (and its alias scholarcli). Analysis of the commands shows they are scoped to literature searching, metadata retrieval, and local file management (BibTeX and PDF caching). The use of scholar search, scholar rq, and scholar verify are standard for the intended purpose of the skill.
  • [EXTERNAL_DOWNLOADS]: The skill describes downloading academic papers and metadata from reputable sources such as Semantic Scholar, OpenAlex, arXiv, and Crossref. These are well-known academic services. The scholar pdf open command downloads files to a local cache, which is a standard feature for research agents.
  • [DATA_EXFILTRATION]: No patterns of data exfiltration were detected. The network operations are limited to academic database queries and paper downloads. The skill explicitly warns against citing from memory and requires recording the search path (FOUND-VIA), which improves transparency.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 07:10 AM
Security Audit — agent-trust-hub — backing-claims