The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes R code via the Rscript command to perform QCA, as defined in its core functionality and allowed-tools configuration in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill requires the installation and use of the external R package 'QCA', a standard academic library for this type of analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied CSV files, which serves as a potential attack surface if the data contains malicious instructions designed to influence the agent's interpretation of the output fit parameters (inclS, PRI, etc.).\n
Ingestion points: Data enters the system via read.csv(args[1]) in the template provided in references/qca_in_r.md.\n
Boundary markers: No explicit boundary markers or prompt-level directives to ignore embedded instructions are present in the provided templates.\n
Capability inventory: The skill has the capability to execute shell commands (Bash(Rscript)) and perform file system reads.\n
Sanitization: No specific sanitization or validation of the input data fields is implemented prior to processing in R.
[DYNAMIC_EXECUTION]: The skill is designed to generate a temporary R script (qca_run.R) at runtime based on the provided template and user parameters, which is then executed by the system.

alterlab-qca