The Agent Skills Directory

[SAFE]: The skill is a legitimate scientific tool that implements well-known medicinal chemistry rules and structural alerts. No malicious code or patterns were detected in the scripts or documentation.
[DATA_EXFILTRATION]: Analysis of scripts/filter_molecules.py shows that the skill only interacts with the local file system for reading molecular data and writing results. There are no network calls or external data transmissions.
[REMOTE_CODE_EXECUTION]: The skill does not contain any remote code execution patterns or downloads of untrusted scripts. All external library references are to standard, well-known scientific packages.
[CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or private credentials were found within the skill files or metadata.
[PROMPT_INJECTION]: The skill handles molecular property metadata in scripts/filter_molecules.py. Although this constitutes an ingestion surface for external data, the risk is negligible as it is restricted to standard chemoinformatics parsing and does not influence agent behavior or safety constraints.

medchem