The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill relies on the workspace-mcp package, which is downloaded and executed at runtime using the uvx tool as specified in SKILL.md.
[COMMAND_EXECUTION]: The skill instructions define an execution mode that uses the system shell to run the uvx workspace-mcp --cli command for various workspace management tasks.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from multiple external sources while maintaining high-privilege capabilities.
Ingestion points: External data is ingested via tools defined in references/gmail.md (reading emails), references/drive.md (reading files), references/chat.md (reading chat messages), and references/search.md (processing search engine results).
Boundary markers: The provided instructions do not include explicit delimiters or instructions to the agent to disregard embedded commands within the external data being processed.
Capability inventory: The agent has access to sensitive tools including send_gmail_message (sending email), manage_drive_access (modifying file permissions), and run_script_function (executing Google Apps Script code).
Sanitization: The skill documentation does not specify any sanitization, escaping, or validation steps for content retrieved from external sources before it is interpreted by the agent.

managing-google-workspace