cnki-export

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a local Python script (push_to_zotero.py) via the shell to process citation data and communicate with the Zotero desktop application.\n- [EXTERNAL_DOWNLOADS]: The bundled Python script performs network requests to download PDF documents from academic servers using URLs parsed from CNKI search results or detail pages.\n- [INDIRECT_PROMPT_INJECTION]: The skill extracts and processes untrusted academic metadata and citation strings from the CNKI website.\n
  • Ingestion points: Browser content extracted via mcp__chrome-devtools__evaluate_script in SKILL.md.\n
  • Boundary markers: None present in the instructions to prevent the agent from potentially following instructions embedded in paper titles or abstracts if processed as part of a larger workflow.\n
  • Capability inventory: Shell command execution (to run the Python script), local network access (to Zotero API), and remote network access (for PDF downloads).\n
  • Sanitization: The Python script includes basic HTML tag stripping using regular expressions for the ELEARNING citation format.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 02:57 AM
Security Audit — agent-trust-hub — cnki-export