jinko-task-literature-search

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests publication metadata, including titles and abstracts, from external sources like PubMed and Crossref. While this provides a surface for indirect prompt injection, the skill processes this data into a structured JSON shortlist and does not execute instructions contained within the text. Ingestion points occur in scripts/literature_search.py and scripts/publication_download.py.- [DYNAMIC_EXECUTION]: The skill utilizes importlib.import_module to lazily load the requests library in scripts/common.py and scripts/pmc_open_access.py. This pattern is used for standard dependency management to ensure the library is available only when network operations are initiated, which is a common and safe practice in this context.- [EXTERNAL_DOWNLOADS]: The skill connects to multiple well-known and trusted scientific data services, including NCBI (eutils.ncbi.nlm.nih.gov), Crossref (api.crossref.org), NIH iCite (icite.od.nih.gov), OpenAlex (api.openalex.org), and the PMC Open Access bucket on Amazon S3. These downloads are essential for the skill's primary purpose of literature discovery and are directed at legitimate, well-known services.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 11:53 AM
Security Audit — agent-trust-hub — jinko-task-literature-search