The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for untrusted data via the input.csv file processed by scripts/etf_structure_liquidity_diagnostics.py. However, the risk is mitigated as the script enforces strict type conversion using pd.to_numeric and pd.to_datetime with error coercion, preventing the agent from interpreting raw text as instructions. The resulting output is a restricted JSON artifact containing only statistical aggregates.
[SAFE]: The Python script is self-contained and focuses exclusively on mathematical diagnostics. It does not utilize any network-enabled libraries, nor does it access sensitive system directories or credentials. The workflow follows standard quantitative research practices with explicit controls and reproducible outputs.

etf-structure-liquidity