generate-profile

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes several shell commands to manage the SGLang server lifecycle, such as sglang serve, python3 -m sglang.test.run_eval, and pkill. These are necessary for the stated purpose of benchmarking and profiling the server.
  • [EXTERNAL_DOWNLOADS]: The skill references https://ui.perfetto.dev/ as a tool for viewing the generated profiling traces. This is a well-known, trusted web-based tool for performance analysis.
  • [DATA_EXFILTRATION]: No data exfiltration patterns were detected. The skill interacts with the local server via 127.0.0.1 and saves trace data to the local /tmp directory.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 01:19 AM
Security Audit — agent-trust-hub — generate-profile