The Agent Skills Directory

[SAFE]: The skill is entirely instructional, providing a structured framework for financial analysis. All identified patterns, including code snippets for Office JS and Python, are directly related to the legitimate purpose of spreadsheet generation.
[DYNAMIC_EXECUTION]: The skill contains logic for the agent to generate code using Python's openpyxl library and Microsoft's Office JS API. These instructions are used for automating standard financial modeling tasks and do not involve untrusted code execution or remote payloads.
[INDIRECT_PROMPT_INJECTION]: The skill describes an attack surface by ingesting external financial data from MCP sources (e.g., FactSet, S&P Kensho) and local Excel examples. However, it explicitly mitigates risks by instructing the agent to maintain strict audit trails, use verified institutional sources, and verify all data points step-by-step with the user.

comps-analysis