compute
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a library of mathematical and technical analysis functions for quantitative finance. All code is self-contained, lacks external network dependencies, and performs data transformations consistent with its documentation.
- [DATA_EXFILTRATION]: No network access or data exfiltration mechanisms were detected. The skill reads market data and writes strategy labels in Parquet format, but these operations are restricted to the local filesystem within a research workspace context.
- [DYNAMIC_EXECUTION]: The code includes a Factor class in wrappers.py that executes provided callables. This is a standard programming pattern for extending the library's functionality and does not involve executing strings from untrusted sources or performing unsafe runtime compilation.
- [INDIRECT_PROMPT_INJECTION]: The skill reads external Parquet data in adjust.py and label_maker.py, which technically presents an ingestion surface. However, the data is parsed as numeric values and there is no mechanism to interpolate this data into agent prompts, mitigating the risk of injection attacks.
Audit Metadata