google-ads-plan
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided business context, goals, and targets and saves them directly to workspace files like
workspace/ads/plan.mdandworkspace/ads/goals.md. This persisted data is used by the agent as a reference for campaign architecture and budget allocation.\n - Ingestion points: User interaction during the planning process to gather business models, offers, and KPIs as defined in
SKILL.md.\n - Boundary markers: No explicit boundary markers or instructions to ignore embedded instructions are present in the workspace writing logic.\n
- Capability inventory: The skill performs file write operations to the
workspace/directory and utilizes Model Context Protocol (MCP) to pull live account data.\n - Sanitization: No sanitization, validation, or escaping logic is defined for the user-provided data before it is written to the workspace memory.
Audit Metadata