ads-landing
Pass
Audited by Gen Agent Trust Hub on Apr 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes GAQL queries through the
runScripttool to access Google Ads account data, such as ad group performance and conversion metrics.\n- [DATA_EXFILTRATION]: The skill usesWebFetchto connect to external landing page URLs and the Google PageSpeed Insights API. While the API call targets a well-known and trusted service, the fetching of arbitrary landing pages constitutes network activity to non-whitelisted domains.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection attack surface as it retrieves and interprets untrusted HTML content from external sources.\n - Ingestion points: External HTML content is ingested via
WebFetchin Phase 2 of the skill execution.\n - Boundary markers: There are no explicit markers or instructions provided to delimit or ignore instructions that may be embedded within the external content.\n
- Capability inventory: The agent can execute Ads API queries (
runScript), perform network fetches (WebFetch), and record data to local history files.\n - Sanitization: No specific logic is defined to sanitize or filter the fetched HTML text before the agent processes it to identify headlines and call-to-action details.
Audit Metadata