The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from a financial API and local JSON files, which represents an attack surface where malicious instructions could theoretically be embedded in data fields.
Ingestion points: Data is ingested through FMPClient._request in scripts/screen_momentum_burst.py (fetching from Financial Modeling Prep API) and read_prices_json (loading local files).
Boundary markers: The skill does not explicitly use delimiters or instruction-ignore blocks when presenting the processed stock data to the agent context.
Capability inventory: The skill performs network reads via requests.get and writes analysis reports to the local filesystem using Path.write_text in scripts/screen_momentum_burst.py.
Sanitization: The script performs strict numeric conversion using to_float and to_int helper functions, and normalizes ticker symbols, which significantly mitigates the risk of non-numeric data injection within OHLCV bars.
[EXTERNAL_DOWNLOADS]: The script fetches stock market data from a well-known financial service provider at runtime.
Evidence: FMPClient in scripts/screen_momentum_burst.py makes HTTP GET requests to financialmodelingprep.com to retrieve company universe and historical price data.
Context: These network operations are the primary function of the skill and target a recognized financial data service.
[COMMAND_EXECUTION]: The skill provides instructions for the user or agent to execute a Python script to perform the screening.
Evidence: SKILL.md contains command-line examples such as python3 skills/stockbee-momentum-burst-screener/scripts/screen_momentum_burst.py --fmp-universe.
Context: The execution is limited to the provided script and is intended for local data analysis and report generation.

stockbee-momentum-burst-screener