The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is instructed to ingest and process external data sources, including market data, trading requirements, and historical strategies, which could serve as a vector for malicious instructions.
Ingestion points: The skill instructions in SKILL.md require querying context managers for requirements and analyzing external market opportunities and datasets.
Boundary markers: There is no specification in the instructions to use delimiters or ignore embedded instructions within the ingested financial data.
Capability inventory: The skill has access to powerful tools including Read, Write, Bash, and python, along with scientific libraries such as numpy, pandas, quantlib, zipline, and backtrader.
Sanitization: The development workflow lacks explicit steps for validating or sanitizing external market data before it is used in model development or execution.

quant-analyst