opencite

Pass

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on uvx and uv pip to install and run the opencite CLI tool, which handles paper searches and metadata management.
  • [EXTERNAL_DOWNLOADS]: The skill performs network requests to fetch metadata and PDF documents from reputable academic services, including Semantic Scholar, PubMed, OpenAlex, and arXiv.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it retrieves and processes content from external academic papers that are then provided to the agent for synthesis.
  • Ingestion points: Untrusted data enters the context from external API search results and from PDF-to-markdown conversions of scholarly papers.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded content are defined in the workflow for synthesizing paper content.
  • Capability inventory: The skill possesses capabilities for shell command execution (uvx), network access to academic domains, and local file system operations (reading and writing PDFs, JSON, and Markdown).
  • Sanitization: No specific evidence of content validation or sanitization of the text extracted from papers is provided prior to agent processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 17, 2026, 04:36 PM
Security Audit — agent-trust-hub — opencite