The Agent Skills Directory

[SAFE]: The skill integrates with trusted cloud provider SDKs (boto3 for AWS and google-cloud-dlp for GCP) to manage sensitive data discovery. These downloads and references to official libraries are standard for the skill's purpose.
[SAFE]: Configuration throughout the skill uses appropriate placeholders for account identifiers, project IDs, and encryption keys, adhering to secure coding practices and preventing accidental credential exposure.
[SAFE]: Data discovery and de-identification tasks are performed strictly through official cloud service APIs. There is no evidence of unauthorized network activity or data exfiltration attempts.
[INDIRECT_PROMPT_INJECTION]: The skill contains logic to process external text and cloud storage content for DLP analysis. While this represents a data ingestion surface, the content is treated as data for inspection by specialized cloud services rather than being executed as instructions. This behavior is inherent to the skill's function and does not pose a runtime risk.

implementing-cloud-dlp-for-data-protection