The Agent Skills Directory

[SAFE]: The skill instructions are purely analytical and provide a methodology for evaluating sales data provided by the user.
[COMMAND_EXECUTION]: The skill references a local Python script scripts/funnel_leak.py for performing calculations. Analysis of the script shows it uses only Python standard libraries (argparse, re, sys) and contains no network, file system, or dynamic execution capabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied sales counts and objections. Ingestion occurs via user input described in the 'What you need' section. The capabilities of the skill are limited to basic arithmetic and text-based mapping against reference files, presenting no significant risk. No sanitization is explicitly mentioned, but the data is processed into structured tables.

offer-diagnosis