tooluniverse-organic-chemistry
Pass
Audited by Gen Agent Trust Hub on May 23, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions focus entirely on teaching chemical reasoning strategies. No override markers, bypass attempts, or role-play injections were detected.
- [DATA_EXFILTRATION]: No network operations (such as curl, wget, or requests) or sensitive file path access (such as .ssh or .aws) are present in the scripts or instructions. All calculations are performed locally.
- [REMOTE_CODE_EXECUTION]: The skill does not download or execute code from external sources. All provided scripts use standard Python libraries for mathematical and chemical calculations.
- [COMMAND_EXECUTION]: The skill uses local Python scripts as computational aids for tasks like molecular weight verification and combustion analysis. These scripts do not utilize dangerous functions like eval(), exec(), or os.system() on user-provided input.
- [OBFUSCATION]: No obfuscated content, encoded strings, or hidden characters were found in the skill files or metadata.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied chemical notations (formulas, SMILES strings) as data for local scripts. These scripts use specific regex-based parsers that do not interpret data as instructions, maintaining a safe boundary.
Audit Metadata