google-ads
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, including customer reviews, call notes, and search term reports, to generate ad copy and keyword strategies. This creates a surface where malicious instructions embedded in public-facing text (e.g., Google Business Profile questions or search queries) could influence the agent's behavior during the drafting phase.\n
- Ingestion points:
SKILL.md(Step 4) instructs the agent to collect words from call notes, form messages, reviews, and search queries;references/weekly-check.md(Step 3) processes search terms with spend.\n - Capability inventory: The agent drafts campaign structures, keyword themes, and responsive search ads.\n
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the processed customer data.\n
- Sanitization: No explicit sanitization or filtering logic is provided to handle potentially malicious strings in the ingested user-generated content.
Audit Metadata