profiling-analysis-computing
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
op_perf_analysis_combine.pyusessubprocess.run()to execute other local scripts within the skill's directory (op_high_time_selector.py,op_pivot_table_analyzer.py, andextract_op_shapes.py). The commands are constructed using a list format and usesys.executable, which is a secure method for automating internal workflows. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill processes performance data from external CSV and Excel files located in a user-provided directory (
op_statistic_*.csv,op_summary_*.csv, etc.). - Boundary markers: Data is processed as structured metrics; no specific markers are used to separate data from instructions.
- Capability inventory: The skill performs file I/O and executes local subprocesses to complete its analysis.
- Sanitization: Data is parsed using standard libraries like
pandas. While no natural language filtering is present, the data source (NPU profiling logs) is a technical log format with a very low risk of containing adversarial instructions.
Audit Metadata