The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documentation describes routines for querying and processing data from the Google Ads API (e.g., campaign names, performance metrics, and budgets). This creates a vulnerability surface where an attacker with control over the external Google Ads account could inject malicious instructions into metadata or campaign fields to influence the agent.
Ingestion points: Results from API search queries such as googleAds:search and resource listings via treg connections resources (SKILL.md).
Boundary markers: The skill does not define specific delimiters or 'ignore' instructions for the agent when handling results from the external API.
Capability inventory: The skill utilizes the treg CLI tool to perform network requests, list connections, and execute mutations like creating campaigns or updating budgets (SKILL.md).
Sanitization: There is no mention of sanitization or validation protocols for the incoming API data before it is processed by the agent.
[COMMAND_EXECUTION]: The instructions direct the agent to interact with the environment through shell commands using the treg tool (e.g., treg call, treg connections ls). While these commands are central to the skill's stated purpose of managing Google Ads, they represent a high-privilege execution capability.

google-ads