The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill uses the Financial Modeling Prep (FMP) API to fetch market index data. Access is managed through an API key stored in environment variables (FMP_API_KEY), which is a standard and secure practice. No sensitive user files or credentials from the local system are accessed.
[EXTERNAL_DOWNLOADS]: The skill requires the requests library for communicating with the FMP API. The primary data source is financialmodelingprep.com, which is a well-known and legitimate financial data provider. The workflow relies on these established services for its core functionality.
[COMMAND_EXECUTION]: The skill executes a local Python script (ftd_detector.py) to perform its analysis. The shell commands described in the documentation are restricted to running this script with the provided API key, with no signs of arbitrary command injection or dangerous system operations.
[DYNAMIC_EXECUTION]: All data processing is performed using standard Python logic without the use of eval(), exec(), or runtime code compilation. Reports are generated as static JSON and Markdown files for user review.
[INDIRECT_PROMPT_INJECTION]: The skill ingests numerical financial data (prices, volume, dates) from a remote API. While it presents this data to the agent in report form, the structured nature of the data (stock quotes and historical OHLCV) poses minimal risk for indirect injection attacks.

ftd-detector