infrastructure-as-code

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill is largely benign Terraform IaC guidance but includes instructions that could expose sensitive plan/state details (e.g., posting full terraform plans to GitHub PR comments and sending plan contents to Slack), which weakens security boundaries.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The repository contains runtime downloads of the Terraform binary (e.g. "https://releases.hashicorp.com/terraform/${TF_VERSION}/terraform_${TF_VERSION}_${os}_${arch}.zip" and "https://releases.hashicorp.com/terraform/${TF_VERSION}/terraform_${TF_VERSION}_linux_amd64.zip") that are fetched via curl and installed/executed by CI/scripts, so remote code is retrieved and run at runtime.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill includes a cron example that writes a job file to /etc/cron.d (lines 295-298), which instructs placing/modifying files in a system directory that requires root/sudo privileges, so it encourages modifying the machine's system state.