recipe-backup-sheet-as-csv
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
gwsbinary with various subcommands (sheets spreadsheets get,drive files export,sheets +read) to interact with Google Workspace. These commands are used to retrieve metadata and file content. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external Google Sheets, which creates an attack surface for indirect prompt injection if a spreadsheet contains instructions meant to override agent behavior.
- Ingestion points: The
gws sheets +readandgws drive files exportcommands inSKILL.mdfetch content from remote Google Sheets. - Boundary markers: Absent. The instructions do not define clear delimiters or warnings to treat the spreadsheet content as untrusted data.
- Capability inventory: The skill uses the
gwstool to read, get, and export file data. - Sanitization: Absent. There is no evidence of content filtering or validation before the data is processed by the agent.
Audit Metadata