dd-gcp-integration
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill reads
.envand.env.localfiles to retrieve Datadog API and application keys. These credentials are transmitted to official Datadog API endpoints (e.g.,datadoghq.comand its regional variants) for authentication and verification. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the GCP environment and user input, creating a surface where malicious project or folder names could influence agent behavior during automation. 1. Ingestion points: Results from
gclouddiscovery commands and direct user input for resource scoping. 2. Boundary markers: Absent; resource identifiers are interpolated directly into configuration templates. 3. Capability inventory: Execution ofterraform,curl, andgcloudsubprocesses. 4. Sanitization: No explicit validation or escaping of discovered GCP metadata is performed before generation. - [DYNAMIC_EXECUTION]: The skill generates Terraform HCL templates at runtime by incorporating discovered GCP metadata and user-provided IDs, which are then applied to the user's infrastructure.
- [COMMAND_EXECUTION]: Subprocess execution is used to invoke
terraform,gcloud, andcurlto discover environment state, provision cloud resources, and verify integration status.
Audit Metadata