The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute Python scripts for essential model maintenance, specifically recalc.py for formula recalculation (part of the platform's infrastructure) and a local script scripts/validate_dcf.py to check for logical errors and formula inconsistencies in the generated Excel files.
[EXTERNAL_DOWNLOADS]: The skill retrieves financial data from well-known external sources using the yfinance and requests libraries. This is a core functionality necessary for obtaining current stock prices, beta values, and historical financial statements for companies.
[PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection because it ingests untrusted data from the web.
Ingestion points: Untrusted data enters the agent's context through financial lookups via the yfinance library and general web searches specified in the data retrieval phase of SKILL.md.
Boundary markers: Absent. The instructions do not define delimiters or provide warnings for the agent to ignore instructions potentially embedded in external data.
Capability inventory: The skill possesses capabilities for writing Excel files (openpyxl), making network requests (requests), and executing local commands (recalc.py and validate_dcf.py).
Sanitization: Absent. The skill lacks instructions or logic to sanitize, escape, or validate external content before it is incorporated into the modeling process or final output.

dcf-model