The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill implements a legitimate configuration workflow for marketing automation.
[DATA_EXPOSURE]: The skill reads from local configuration directories including .claude/ads-agent/config/ad_accounts.md and account briefs. This access is necessary for identifying account contexts and is restricted to the agent's own configuration workspace.
[COMMAND_EXECUTION]: Calls the get_campaigns tool to retrieve live campaign data. This is a standard functional requirement for auditing existing naming structures.
[PROMPT_INJECTION]: Ingests campaign data (names and IDs) from external sources. While this represents a surface for indirect prompt injection if campaign names are manipulated by an external actor, the skill's logic is focused on static mapping and configuration generation rather than autonomous decision-making based on that data.

naming-rules