The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute an external Python script, recalc.py, which is not included in the skill's source files. This represents a dependency on an external execution component assumed to be present in the platform environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes financial data from external sources such as SEC filings, analyst reports, and web search results. This data is used to build complex models without specified boundary markers, creating a surface for indirect prompt injection where instructions hidden in financial text could influence the agent.
Ingestion points: Financial data is sourced from MCP servers (e.g., Daloopa), user-provided research, and web fetching tools as described in SKILL.md.
Boundary markers: The skill lacks instructions for using delimiters or protective prompts to segregate untrusted financial data from the agent's core instructions.
Capability inventory: The skill possesses capabilities to write to the file system (Excel generation) and execute local scripts (scripts/validate_dcf.py and the external recalc.py).
Sanitization: There are no documented procedures for sanitizing or validating the content of external financial documents before they are processed by the agent.

dcf-model