tooluniverse-vaccine-design
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a comprehensive and legitimate computational pipeline for designing vaccine candidates. It utilizes established bioinformatics services (IEDB, UniProt, BVBRC, AlphaFold) to predict immune responses and population coverage.
- [COMMAND_EXECUTION]: The skill invokes a bundled Python utility,
scripts/population_coverage.py, to calculate population-level HLA coverage. This execution is scoped to the skill's package and represents a standard method for extending agent capabilities with custom logic. - [INDIRECT_PROMPT_INJECTION]: The
population_coverage.pyscript accepts user-provided file paths for allele lists and frequency tables. \n - Ingestion points: The script reads data from files specified by the
--alleles-fileand--freq-filecommand-line arguments inSKILL.md. \n - Boundary markers: No explicit delimiters are used in the prompt interpolation; however, the script expects structured data (alleles or frequency tables). \n
- Capability inventory: The script
scripts/population_coverage.pyreads the specified local files and prints a JSON-formatted summary of parsed bioinformatics data to the console for the agent's consumption. \n - Sanitization: The script performs regex-based splitting and strictly attempts to convert frequency values to floating-point numbers, which limits the potential for processing non-numeric malicious payloads.
Audit Metadata