performance-profiling
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external CSV data through the
performance_profiling_diagnostics.pyscript. The script uses thepandaslibrary to read and process numeric performance metrics. There is a low risk of indirect prompt injection because the agent reads the resulting JSON summary. However, the script strictly validates required columns and performs numeric conversions, which significantly limits the attack surface for instruction injection via data files.
Audit Metadata