literature-survey

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions and shell commands for using pdftotext (part of the poppler-utils suite) to extract content from academic PDFs. It includes commands for directory creation (mkdir -p) and batch processing of files using shell loops.
  • [EXTERNAL_DOWNLOADS]: The documentation references the installation of the poppler-utils dependency via standard system package managers, specifically apt for Linux and brew for macOS.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of academic PDFs and BibTeX entries.
  • Ingestion points: Files located in references/pdf/ and entries in references/bibliography.bib.
  • Boundary markers: None explicitly defined; the agent processes the raw output of the text extraction tool.
  • Capability inventory: File system writes (.md, .bib, .txt), shell command execution (pdftotext), and potential network access for citation verification.
  • Sanitization: No specific sanitization or filtering logic is provided for the extracted text before it is presented to the agent for summarization.
  • [SAFE]: The skill's behavior is consistent with its stated purpose of assisting in scientific research. The operations performed are local, transparent, and rely on well-known third-party utilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 12:47 PM
Security Audit — agent-trust-hub — literature-survey