The Agent Skills Directory

[SAFE]: No security issues detected. The skill provides legitimate infrastructure-as-code patterns for authentication and identity management.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill demonstrates secure practices by using Kubernetes secretKeyRef and AWS Secrets Manager (aws secretsmanager put-secret-value) to handle administrative credentials and client secrets. It does not hardcode any real secrets, using only placeholders and platform-native secret references (e.g., <+secrets.getValue(...)>).\n- [COMMAND_EXECUTION]: The Bash scripts utilize standard DevOps utilities such as curl, jq, and the aws CLI to interact with Keycloak APIs and cloud services. These operations are restricted to the intended purpose of configuring OIDC clients and importing realms.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes environment variables and Harness pipeline variables (e.g., <+service.name>). While these are interpolated into shell commands, they are sourced from the platform's internal configuration environment rather than untrusted user-controlled inputs.

harness-keycloak-auth