The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process experimental results from JSON and CSV files, which creates an ingestion surface for untrusted data if the source files are provided by an external party.
Ingestion points: The script scripts/results_to_table.py reads and parses data from files specified by the --input flag.
Boundary markers: The skill does not implement specific delimiters or instructions to ignore potential natural language commands embedded within the data cells.
Capability inventory: The skill scripts have the ability to read and write to the local filesystem as part of the data conversion process, but do not perform network operations or execute shell commands on the data contents.
Sanitization: The script includes standard LaTeX escaping for special characters like _, %, &, and # to ensure document integrity, although it does not filter for logical instructions within the input data.

table-generation