The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes standard Python packages including pandas-ta, pandas, numpy, and httpx for data processing and API interaction.
[EXTERNAL_DOWNLOADS]: Market data is fetched from the Birdeye API (public-api.birdeye.so), which is a well-known service for cryptocurrency OHLCV data.
[DATA_EXFILTRATION]: No sensitive data exposure detected. The scripts use environment variables for API keys and restrict network communication to the official provider domain.
[PROMPT_INJECTION]: The skill processes structured OHLCV data from an external API, creating a theoretical surface for indirect prompt injection.
Ingestion points: Data enters via the fetch_ohlcv_birdeye function in the provided scripts.
Boundary markers: Not explicitly required as the data is used for numeric calculation rather than direct prompt construction.
Capability inventory: Data is processed using mathematical operations; no dangerous sinks such as eval, exec, or subprocess are used with external inputs.
Sanitization: API responses are parsed for specific keys and explicitly cast to floating-point numbers (.astype(float)), which sanitizes the input stream by removing non-numeric content.
[SAFE]: All core functionality is implemented using standard data science practices, and the code does not perform privileged operations or include persistence mechanisms.

pandas-ta