clari-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill serves as a repository of documentation and code templates for cloud-based data orchestration. All provided scripts follow standard integration patterns for the targeted cloud platforms.
- [CREDENTIALS_UNSAFE]: The skill identifies the need for API keys but correctly avoids hardcoding them. It provides implementation examples using secure retrieval methods such as the AWS SSM Parameter Store (
boto3.client("ssm").get_parameter), Google Cloud Secret Manager (secretmanager.SecretManagerServiceClient()), and Apache Airflow Variables (Variable.get()). - [EXTERNAL_DOWNLOADS]: The code templates reference the following Python dependencies:
clari_client,boto3,functions-framework,google-cloud-secret-manager, andapache-airflow. These are either standard cloud SDKs or specialized clients relevant to the skill's stated purpose. The skill does not contain commands to download or execute scripts from unknown or untrusted remote servers. - [COMMAND_EXECUTION]: The skill requested access to cloud CLI tools via
Bash(gcloud:*)andBash(aws:*). These tools are necessary for the deployment tasks described in the skill and are appropriately scoped to the official cloud provider binaries. - [INDIRECT_PROMPT_INJECTION]: The skill defines templates that process external event data (e.g., in the AWS Lambda and Airflow handlers). While this creates an ingestion surface for untrusted data, the skill includes explicit guidance to "Return only aggregate job status to callers" and maintain data boundaries, reducing the risk of data leakage via injection.
Audit Metadata