invoking-antigravity

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs saving and later writing back the OAuth token file (and describes copying/pasting authorization codes), which requires reading and reproducing secret tokens verbatim (copy/paste of credentials) and thus is direct secret handling/exfiltration risk.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs running a remote installer that is fetched and executed at runtime via "curl -fsSL https://antigravity.google/cli/install.sh | bash", which downloads and installs the required agy binary (and thus directly executes remote code the skill depends on).

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The prompt explicitly instructs bypassing OS keyring (by faking SSH env vars), writing durable OAuth tokens to disk and running the CLI with --dangerously-skip-permissions, which intentionally circumvents security checks and changes machine state/security posture.