deep-research

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill is configured to interact with well-known and trusted academic indexing services, including Semantic Scholar (api.semanticscholar.org), Crossref (api.crossref.org), and OpenAlex (api.openalex.org). These connections are used strictly for retrieving bibliographic metadata and verifying the existence of research papers cited in reports.
  • [COMMAND_EXECUTION]: The timeline_extraction_agent contains instructions to execute the pdftotext shell command to extract publication dates from the cover pages of local research PDFs. This is a standard functional requirement for automated research workflows.
  • [COMMAND_EXECUTION]: The instructions reference several internal verification scripts (e.g., scripts/check_pipeline_integrity.py and scripts/temporal_integrity_audit.py) intended to validate the consistency and temporal logic of the generated research materials.
  • [SAFE]: The skill employs a structured 'Three-Layer Citation' system using HTML comments (e.g., <!--ref:slug-->) to maintain metadata integrity without affecting the human-readable markdown output, demonstrating a sophisticated approach to data tracking and verification.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 02:03 PM
Security Audit — agent-trust-hub — deep-research