The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied gene symbols and protein identifiers to generate reports and query biological databases. There is a lack of explicit sanitization or boundary markers for these inputs when they are interpolated into the generated markdown reports. Additionally, user input is used to construct output filenames, creating a surface for potential path issues.
Ingestion points: gene_list, protein_id, and pathway_keyword parameters in python_implementation.py.
Boundary markers: No delimiters are used to separate user data from instructions in the report output.
Capability inventory: The skill can write files to the local filesystem and is instructed to execute Python code for data analysis.
Sanitization: No input validation or path normalization is performed on the biological identifiers supplied by the user.
[COMMAND_EXECUTION]: In the SKILL.md instructions, the agent is directed to 'write and run Python code via Bash' to perform statistical calculations and data processing. While this is a high-privilege capability, it is consistent with the skill's primary purpose of performing complex systems biology computations.

tooluniverse-systems-biology