devops-engineer-agent

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill consists of natural language instructions and configuration templates for various AI agent platforms. It does not contain any executable scripts, external dependencies, or obfuscated content.
  • [PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection because it is designed to process untrusted external data such as CI/CD configurations, logs, and infrastructure manifests.
  • Ingestion points: Processing of external code, log files, and pull request data.
  • Boundary markers: None explicitly defined in the provided instructions.
  • Capability inventory: The provided Claude Code configuration template requests file system access and shell execution capabilities.
  • Sanitization: No explicit instructions are provided for sanitizing or escaping external content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 12:58 PM
Security Audit — agent-trust-hub — devops-engineer-agent