The Agent Skills Directory

[SAFE]: No security issues or malicious patterns were identified across the 21 files analyzed. The skill follows best practices for R-based tools, including atomic file writing and resource timeout controls.- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through the processing of user-provided CSV/TSV files. However, it implements strong numeric validation and schema checking, reducing the risk of malicious data influencing the agent's behavior.
Ingestion points: scripts/io.R reads expression matrices and group files.
Boundary markers: The skill treats inputs as numeric data for statistical modeling; while no natural language delimiters are present, the data is never interpreted as instructions.
Capability inventory: File system access for writing CSV results and PDF plots (scripts/runtime_utils.R, scripts/plotting.R).
Sanitization: load_expression_matrix and load_group_data verify data types, ID uniqueness, and matrix dimensions to ensure only valid numeric omics data is processed.- [PROMPT_INJECTION]: Instructions in SKILL.md and the reference documentation are purely functional and do not attempt to bypass agent safety filters or override system instructions.- [DATA_EXFILTRATION]: No network capabilities or hardcoded credentials were found. The skill does not access sensitive system files like SSH keys or environment variables.- [DYNAMIC_EXECUTION]: The R scripts use standard modularization via source() for local files. No unsafe eval() or system() calls on untrusted user input were identified.

lasso-logistics-analysis