cnki-export

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill interacts with the academic service kns.cnki.net (CNKI) to fetch citation metadata and download PDF documents.- [COMMAND_EXECUTION]: Executes a local Python script scripts/push_to_zotero.py to process paper data and interact with the Zotero API.- [DATA_EXFILTRATION]: Transmits extracted academic metadata to a local Zotero instance via the well-known Zotero Connector API at http://127.0.0.1:23119.- [PROMPT_INJECTION]: The skill is subject to indirect prompt injection because it processes content scraped from external academic web pages.
  • Ingestion points: Metadata is extracted from CNKI detail and search results pages via JavaScript evaluation in SKILL.md.
  • Boundary markers: None are present in the scraping scripts to delimit external content from instructions.
  • Capability inventory: The skill can perform network operations via fetch in the browser and urllib in Python, as well as execute local scripts.
  • Sanitization: The Python script scripts/push_to_zotero.py performs regex-based cleaning of HTML tags and whitespace on the ingested citation text.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 12:19 AM
Security Audit — agent-trust-hub — cnki-export