The Agent Skills Directory

[External Downloads] (MEDIUM): The skill requires the installation of the pyopenms Python package.
Evidence: Found uv uv pip install pyopenms in SKILL.md.
Risk: The package is downloaded from the public PyPI registry and is not from a source on the trusted organizations list, presenting a supply-chain risk. The command also contains a minor typo (uv uv).
[Command Execution] (LOW): The skill provides shell commands and Python code for package installation and data processing.
Evidence: Multiple code blocks across SKILL.md and reference files.
Risk: Commands are standard for the library's functionality and do not exhibit suspicious behavior such as arbitrary execution from untrusted sources.
[Indirect Prompt Injection] (LOW): The skill processes external mass spectrometry files (mzML, FASTA, etc.), which could theoretically contain malicious data intended to influence the agent.
Ingestion points: File loading functions like ms.MzMLFile().load(), ms.FASTAFile().load(), and ms.IdXMLFile().load() in references/file_io.md.
Boundary markers: None present in the code examples to delimit external data from agent instructions.
Capability inventory: File system access (read/write via store and load methods), data analysis, and result display to terminal.
Sanitization: No sanitization or validation logic for the content of the mass spectrometry files is provided in the snippets.
Risk: Since the skill primarily processes data for scientific analysis and does not make high-privilege decisions or external network calls based on the parsed file content, the risk is classified as low.

pyopenms