The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from SEC filings, analyst reports, and web searches to populate the DCF model. This represents an attack surface where malicious content in these documents could attempt to influence the agent's behavior.
Ingestion points: Data retrieval from MCP servers and web search/fetch as described in Step 1 of the DCF Process Workflow.
Boundary markers: The skill uses a structured step-by-step verification process where the agent must show raw inputs and projections to the user for confirmation before proceeding.
Capability inventory: The agent has capabilities for network operations (fetching data via yfinance/requests) and file system operations (writing Excel files via openpyxl/xlsx tool).
Sanitization: Mitigated by mandatory step-by-step human review of all retrieved data and generated projections.
[DYNAMIC_EXECUTION]: The skill instructs the agent to programmatically generate Excel formulas for sensitivity analysis tables using loops in Python (openpyxl) or Office JS. This is a standard and expected requirement for building functional financial models and is contained within the context of the Excel file creation.
[EXTERNAL_DOWNLOADS]: The skill utilizes well-known and trusted Python packages (openpyxl, yfinance, requests) for its core functionality. These are standard tools for financial data analysis and spreadsheet manipulation.

dcf-model