The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the gws binary to interact with the Google Workspace Admin API.- [INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface where high-privilege administrative actions could be triggered by malicious instructions embedded in data processed by the agent.
Ingestion points: Method parameters and flags for Workspace resources (users, groups, devices) as described in SKILL.md.
Boundary markers: The skill refers to an external file (../gws-shared/SKILL.md) for security rules, but provides no local boundary markers or instructions to ignore data-embedded commands.
Capability inventory: Exposes highly sensitive capabilities including users.makeAdmin (promoting users to Super Administrator), twoStepVerification.turnOff (disabling MFA), verificationCodes.generate/list (accessing backup codes), and remote device wiping via mobiledevices.action and chromeosdevices.action.
Sanitization: No sanitization or parameter validation is described within the skill instructions.

gws-admin