google-drive
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the
@googleworkspace/clipackage via NPM and mentions binary releases on GitHub. These references target the official Google Workspace community-maintained tools, which are considered well-known service resources. - [COMMAND_EXECUTION]: The skill uses
curl,jq, and thegwsCLI tool within shell blocks to interact with the Google Drive API. These commands are used to read metadata, download files, and perform file operations like moving or renaming. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by downloading and exporting external data (e.g., Google Docs as Markdown or Sheets as CSV) that the agent then processes.
- Ingestion points: File export recipes in
SKILL.md(e.g.,export?mimeType=text/markdown). - Boundary markers: Absent; there are no specific delimiters or instructions to ignore embedded commands within the downloaded data.
- Capability inventory: The skill has access to the
Bashtool, enabling network requests (curl) and local file operations. - Sanitization: While the skill uses
jq --argto safely handle shell arguments for filenames, it does not provide methods to sanitize or filter the actual content of the files before the agent reads them.
Audit Metadata