analysis-macro-regime
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a 'pure compute' architecture. It processes local JSON files provided as input and performs mathematical analysis (statistical averages, standard deviations) without initiating network connections or shell commands.
- [COMMAND_EXECUTION]: Static analysis identified the use of
importlib.import_module()inscripts/regime_compose.py. This is used to load country-specific analysis modules (classify_us,classify_jp, etc.). The implementation is secure as it validates the module name against a hardcoded whitelist of supported countries (us,jp,tw,kr,cn) before loading, preventing arbitrary code execution or library hijacking. - [DATA_EXFILTRATION]: While the skill reads local files (macro indicator JSONs), it does not contain any code for network transmission (no use of
requests,urllib, or socket libraries). Data is consumed from local paths and output as structured JSON to standard output. - [CREDENTIALS_UNSAFE]: A scan of the 29 files (including YAML calibrations and Markdown references) found no hardcoded API keys, tokens, or private secrets. Dates and version numbers in the documentation are correctly identified as non-sensitive metadata.
Audit Metadata