oma-tf-infra
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes infrastructure-as-code (HCL) files, modules, and execution plans that may originate from untrusted sources or collaborative environments. This creates a surface for indirect prompt injection where malicious instructions could be embedded in metadata, comments, or data fields within the processed files.\n
- Ingestion points: The skill reads HCL files, modules, and plan output during the 'ACQUIRE' scene in
SKILL.mdand the 'Analyze' step inresources/execution-protocol.md.\n - Boundary markers: The instructions do not explicitly define boundary markers or instruct the agent to disregard instructions found within the ingested infrastructure data.\n
- Capability inventory: The agent has the capability to execute shell commands (
PROCESSscope) via theterraformCLI,checkov,trivy,opa, andinfracosttools.\n - Sanitization: There are no documented procedures for sanitizing or escaping ingested infrastructure code before it is processed by the agent or passed to execution tools.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of local system commands and CLI tools, including
terraform init,terraform plan,checkov,trivy, andopa. While these are standard tools for infrastructure management and security auditing, they are executed on code that the agent may have modified or ingested from external sources.
Audit Metadata