The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions specify that the agent should read from .claude/product-marketing-context.md if it exists. This ingestion of external data creates a surface for indirect prompt injection where instructions hidden in the context file could influence the agent's behavior.
Ingestion points: The local file .claude/product-marketing-context.md is accessed at the start of the task to gather campaign context.
Boundary markers: The skill does not provide any specific delimiters or instructions for the agent to ignore potentially malicious commands embedded within the context file.
Capability inventory: The skill is intended to be used with tools for high-impact advertising platforms such as Google Ads, Meta Ads, and LinkedIn Ads, which could be misconfigured if the agent follows injected instructions.
Sanitization: No sanitization or validation logic is specified for the information ingested from the context file.

paid-ads