The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external Excel files, which represents an attack surface for indirect prompt injection if the files contain malicious instructions meant to influence the agent. However, the skill primarily performs numeric operations and uses sanitization methods like numeric conversion with error coercion, making this a standard and safe data processing pattern.
Ingestion points: The skill ingests data via the df_processed variable and pd.read_excel(file_path) call in SKILL.md.
Boundary markers: No specific boundary markers or "ignore instructions" delimiters are present around the interpolated data.
Capability inventory: The skill has the capability to write files to the sandbox environment using to_excel and to_csv methods in SKILL.md.
Sanitization: The skill uses pd.to_numeric with errors='coerce' to sanitize the target column, which effectively filters out non-numeric injection attempts in the primary data column.

numeric-format-normalization