The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to use the Bash tool to execute statistical scripts and specifications. This is a core functionality intended to 'run the grid' of analytical variations to test result stability.
[INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted external inputs, specifically user-provided scripts and their resulting data files.
Ingestion points: Path provided in the [script or results file] argument (SKILL.md).
Boundary markers: The instructions require the agent to 'pre-register' analytical choices and define the 'estimand' in words before execution, creating a conceptual boundary, though no technical sanitization of file content is specified.
Capability inventory: The skill uses Bash, Write, Edit, and Read tools to execute code and modify the 'specification-search ledger'.
Sanitization: There is no explicit logic for escaping or filtering instructions that might be embedded within the results files or code comments of the processed scripts.
[DYNAMIC_EXECUTION]: The skill involves dynamic generation and execution of code. It instructs the agent to create a grid of specifications (based on analytical 'forks' like measure definition and control sets) and execute each 'fit' individually at runtime using the Bash tool. This is the primary purpose of the skill and is handled within the scope of the provided tools.

challenge