venue-templates
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The helper script
scripts/validate_format.pyexecutes system utilitiespdfinfoandpdffonts(part of the Poppler suite) using thesubprocess.runmethod. This functionality is used to programmatically inspect PDF metadata and verify that fonts are correctly embedded. The implementation follows security best practices by using list-based arguments and resolved file paths, which prevents shell injection vulnerabilities. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it is designed to process untrusted data from external sources to verify academic formatting compliance.
- Ingestion points: User-provided PDF files are inspected by
scripts/validate_format.py, and instructions inSKILL.mdand thereferences/folder direct the agent to visit external academic websites (e.g., nature.com, icml.cc). - Boundary markers: The skill recommends capturing verified data in a specific "compliance note" format, though it does not implement formal "ignore embedded instructions" delimiters for ingested web content.
- Capability inventory: The skill's primary capabilities include local file manipulation and executing Poppler utilities via
scripts/validate_format.py. - Sanitization: Python helpers use resolved paths and avoid shell invocation, mitigating risks associated with malicious file names or path traversal.
- [EXTERNAL_DOWNLOADS]: The skill documentation encourages the agent to visit official academic and scientific platforms (such as Nature Portfolio, Cell Press, NeurIPS, and arXiv) to retrieve current formatting guidelines and citation information. These network operations target well-known and reputable organizations and are strictly for retrieving research-related configuration and metadata, which is consistent with the skill's primary purpose.
Audit Metadata