microstructure-hft
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data via
scripts/microstructure_hft_diagnostics.py, which reads user-provided CSV files. This creates a surface for indirect prompt injection where malicious content in the data could attempt to influence the agent's interpretation or summary. - Ingestion points: The diagnostic script reads data from
input.csvusing thepandas.read_csvfunction. - Boundary markers: The instructions do not define clear delimiters or specific guidance for the agent to ignore instructions that might be embedded within the CSV data rows.
- Capability inventory: The skill's functionality is limited to statistical analysis and report generation; it does not contain network access, subprocess execution of arbitrary commands, or persistence mechanisms.
- Sanitization: The Python script uses
pd.to_numericandpd.to_datetimewith error coercion, which provides basic validation and sanitization of the data before it is summarized for the agent.
Audit Metadata