identity-access-management

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill includes multiple examples that embed plaintext passwords and tokens directly into commands and data (e.g., --password "TempP@ss123!", CSV password fields, and Authorization headers using token variables), which would require the agent to accept and/or emit secret values verbatim and therefore poses an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs fetching and executing a remote installer from a public URL (bash <(curl -s -S -L https://gam-shortn.appspot.com/gam-install) in the "Install GAM" step), which is untrusted third-party content that the workflow expects to be retrieved and executed and therefore can materially influence behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill includes a runtime installation command that fetches and executes a remote script via bash <(curl ...) from https://gam-shortn.appspot.com/gam-install, which executes remote code and is a required dependency for the Google Workspace flow.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (low risk: 0.30). The skill primarily instructs configuration of remote IdPs via APIs (Okta, Google Workspace, Azure AD) and does not ask the agent to create local OS accounts or explicitly request sudo, but it does include local actions (installing CLIs via curl|bash, modifying crontab, writing to /var/log) that change the host state, so it poses a small but non-zero risk.