The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses system2 within its test suite (tests/run_tests.R) to execute the workflow and validation scripts. These calls are restricted to the skill's own internal scripts and utilize a validation function to ensure output directories remain within the designated test folder.
[DATA_EXPOSURE]: The workflow performs file I/O strictly for reading user-provided expression matrices and writing results to a user-defined output directory. No access to sensitive system paths or credentials was detected.
[DYNAMIC_EXECUTION]: The R scripts use source() to load modular components (utils.R, io.R, etc.) from the skill's own directory. The search path for these modules is dynamically resolved at runtime based on the location of the main script, which is a standard and safe practice for modular R scripts.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external data in CSV or TSV format. It implements strong validation by strictly converting data columns to numeric types and checking for non-finite values (Inf, NaN). While it processes feature identifiers as strings, these are not used to influence control flow or build dynamic commands, effectively mitigating injection risks.
[EXTERNAL_DOWNLOADS]: The skill declares a dependency on the well-known R package optparse in the DESCRIPTION file and references its installation from standard R repositories. No unauthorized or untrusted remote code downloads were identified.

gene-protein-expression-matrix-normalization