The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted external data, which could potentially contain malicious instructions targeted at the agent.
Ingestion points: The weekly_tasks and monthly_tasks sections in SKILL.md instruct the agent to review_search_terms and perform competitor_analysis by analyzing external ad copy and search queries.
Boundary markers: No specific delimiters or instructional guardrails are defined in the templates to differentiate between data and instructions when the agent processes search term reports.
Capability inventory: The skill utilizes tools such as google_ads_create and google_ads_report, providing the agent with the ability to modify campaign settings based on the data it processes.
Sanitization: There are no explicit sanitization or validation steps provided for the external strings retrieved from the Google Ads API before they are processed by the agent.

google-ads-manager