The Agent Skills Directory

[SAFE]: The skill documents legitimate Pulumi development workflows, project scaffolding, and resource management. It actively promotes security best practices, such as centralizing secrets in Pulumi ESC, using OIDC for cloud provider authentication, and enabling encryption on storage resources.
[DATA_EXPOSURE_AND_EXFILTRATION]: No evidence of hardcoded credentials or unauthorized data access. The instructions explicitly warn against committing secrets to version control and demonstrate the use of pulumi.Config.requireSecret() for secure handling.
[REMOTE_CODE_EXECUTION]: No suspicious remote script execution patterns (e.g., curl | bash) were found. References to external packages and repositories are limited to standard Pulumi providers and example project templates.
[OBFUSCATION]: The content is presented in clear text without any hidden encoding, zero-width characters, or homoglyph attacks.
[INDIRECT_PROMPT_INJECTION]: While the skill generates infrastructure code based on user input, it includes a 'Deployment Workflow with Validation' section. This mandates using pulumi preview and manual verification of changes before deployment, which serves as a critical human-in-the-loop security control against unexpected resource creation.

pulumi-typescript