The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/export_coding.py executes subprocess.check_call to install the openpyxl library from the public PyPI registry. This action modifies the local environment by running shell commands at runtime.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its handling of external, untrusted research data.
Ingestion points: The skill reads interview records and qualitative data from local file paths provided by the user as described in SKILL.md.
Boundary markers: The instructions do not define clear delimiters or use "ignore embedded instructions" warnings when processing the contents of these files.
Capability inventory: The skill has the ability to write multiple Markdown files to the workspace (coding_GT_*.md) and execute a Python script (export_coding.py) which performs environment modifications.
Sanitization: There is no evidence of sanitization or filtering of the content within the processed qualitative files prior to analysis or inclusion in generated reports.

grounded-coding