Skills
skills/chaterm/terminal-skills/benchmarking/Gen Agent Trust Hub

benchmarking

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
  • [CREDENTIALS_UNSAFE] (HIGH): Hardcoded credential placeholders used in CLI examples. Evidence: The MySQL benchmark sections in SKILL.md include --mysql-password=pass. Providing passwords as command-line arguments is a security risk as they appear in plaintext in command history files and system process listings.
  • [COMMAND_EXECUTION] (MEDIUM): Instructions for software installation and system-intensive operations. Evidence: SKILL.md contains commands to install software via apt and yum, and performs operations like sysbench fileio --file-total-size=10G prepare which can have significant performance impacts if run on sensitive environments.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 10:56 AM