glycoengineering

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the 'glycoshield' Python package via pip. This is a legitimate scientific tool developed for glycan shielding analysis in molecular dynamics simulations.
  • [DATA_EXFILTRATION]: The skill uses the 'requests' library to query established public bioinformatics databases, including GlyConnect (hosted by the SIB Swiss Institute of Bioinformatics) and GlyTouCan (a glycan structure repository). These operations are standard for retrieving site-specific glycosylation data and do not involve the transmission of sensitive user credentials or system files.
  • [COMMAND_EXECUTION]: The skill includes documentation for using the 'glycoshield' command-line interface. The provided examples demonstrate standard scientific usage for analyzing protein topology and trajectory files.
  • [PROMPT_INJECTION]: No attempts to override agent behavior, bypass safety guardrails, or extract internal system prompts were found in the instructions or code documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 08:21 AM
Security Audit — agent-trust-hub — glycoengineering