detecting-performance-regressions

Warn

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts scripts/create_github_comment.py and scripts/generate_report.py implement a generate_script method that writes a template string to a .sh file and then applies executable permissions via chmod 0o755. Since the content of this script can be provided via command-line arguments, this allows for the creation of executable files from arbitrary strings.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection due to its handling of untrusted CI/CD performance data.
  • Ingestion points: scripts/analyze_metrics.py traverses and analyzes files from external CI/CD build environments.
  • Boundary markers: No delimiters or instructions are used to ensure the agent ignores malicious instructions embedded within the performance data.
  • Capability inventory: The skill possesses the capability to write files and create executable scripts.
  • Sanitization: There is no evidence of data sanitization or validation of the metric inputs before they are utilized in reporting or script generation logic.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 18, 2026, 10:33 AM