The Agent Skills Directory

Indirect Prompt Injection Surface: The skill is designed to ingest large volumes of data from external sources, including SEC filings, company websites, news articles, and financial transcripts. There is a potential consideration that if these external sources were to contain hidden instructions or malicious content, the agent could inadvertently process them. The skill uses detailed verification protocols and structured tasks which provide some procedural boundaries, but explicitly sanitizing or using specific delimiters for external content is recommended.
Dynamic Script Generation: For the 'Chart Generation' task, the skill provides templates for generating Python code using libraries such as Matplotlib and Seaborn. This represents a dynamic execution pattern where the agent generates and runs scripts locally. While this is the intended functionality for creating professional financial visualizations, users should ensure the agent's execution environment is appropriately sandboxed.
External Data Retrieval: The skill directs the agent to retrieve data from well-known financial services and government repositories, specifically the SEC EDGAR system and financial news platforms. These are established services used for legitimate research purposes, and the skill provides direct URLs to these trusted platforms.
Package Dependencies: The skill identifies several standard data science and visualization packages for installation, including pandas, numpy, and matplotlib. These are well-maintained, industry-standard libraries, and their use is consistent with the skill's stated purpose of financial modeling and reporting.

initiating-coverage