iac-security

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to read IaC and state files and include "Evidence: " and other concrete artifacts in findings, which will force the LLM to surface literal secret values (hardcoded keys, passwords, or state contents) verbatim in its output.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). SKILL runtime reads IaC source files provided to it via the target-file-or-directory argument (Terraform/CloudFormation/Pulumi/Bicep discovered with Glob and then evaluated), and those files are explicitly treated as untrusted free text.