recipe-create-gmail-filter

Pass

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the gws (Google Workspace) CLI tool to perform operations such as listing labels, creating new labels, and configuring email filters. These actions involve direct interaction with the host's shell and external API services to modify account settings.
  • [INDIRECT_PROMPT_INJECTION]: This skill provides a surface for indirect prompt injection as it ingests untrusted data (such as sender email addresses or label names) and interpolates them into command-line arguments.
  • Ingestion points: Data enters the context via the 'criteria' and 'action' parameters in the filter creation step (e.g., from field).
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the provided steps.
  • Capability inventory: The skill utilizes the gws binary to execute shell commands that modify Gmail settings.
  • Sanitization: There is no evidence of input validation or escaping for the strings being passed to the gws command.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 18, 2026, 12:25 PM