senior-devops

Pass

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes local Python scripts (pipeline_generator.py, terraform_scaffolder.py, deployment_manager.py) and standard cloud CLI tools like terraform, aws, and kubectl to automate infrastructure tasks.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted user input such as file paths, container image tags, and environment names which are passed directly to automation scripts.
  • Ingestion points: User-provided CLI arguments such as , , and --image in SKILL.md.
  • Boundary markers: None identified; user input is interpolated directly into command structures without delimiters.
  • Capability inventory: Comprehensive subprocess execution capabilities through management scripts and manual workflow commands.
  • Sanitization: No explicit sanitization or validation of external input is described in the skill instructions or scripts descriptions.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 5, 2026, 08:06 PM
Security Audit — agent-trust-hub — senior-devops