recipe-bulk-download-folder

Pass

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from Google Drive, which acts as an untrusted external source.
  • Ingestion points: File metadata (retrieved via gws drive files list) and file content (retrieved via gws drive files get and export) enter the agent's context.
  • Boundary markers: None identified; the instructions do not define delimiters or warnings to ignore instructions embedded within the downloaded files.
  • Capability inventory: The skill utilizes the gws CLI to interact with the file system and network for listing and downloading files in SKILL.md.
  • Sanitization: No sanitization or validation of external content is specified before the agent handles the data.
  • [COMMAND_EXECUTION]: The skill uses shell commands to interact with Google Workspace via a command-line interface.
  • Evidence: Executes gws drive files list, gws drive files get, and gws drive files export with user-provided identifiers (FOLDER_ID, FILE_ID).
  • [EXTERNAL_DOWNLOADS]: The skill retrieves data and exports documents from Google Drive.
  • Evidence: Uses the gws tool to fetch media and convert Google Docs to PDF from Google's official services.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 18, 2026, 12:25 PM