The Agent Skills Directory

[SAFE]: The skill provides instructions for a legitimate scientific library (pyopenms) used in proteomics and metabolomics. It refers to official documentation and uses standard package installation methods.- [EXTERNAL_DOWNLOADS]: Installs the 'pyopenms' Python package from a standard package registry.- [INDIRECT_PROMPT_INJECTION]: The skill processes external mass spectrometry data files which is a potential attack surface. 1. Ingestion points: File loading methods like MzMLFile.load() and FASTAFile.load() in SKILL.md and multiple reference files (file_io.md, identification.md, metabolomics.md). 2. Boundary markers: No boundary markers or 'ignore' instructions are present for data ingestion. 3. Capability inventory: The skill performs local data processing, file I/O, and mathematical transformations. No network exfiltration or arbitrary command execution capabilities were found. 4. Sanitization: Data parsing and validation are delegated to the underlying pyopenms C++ library.

pyopenms