The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to load business context from a local file (.agents/product-marketing-context.md). While this represents a data ingestion surface for external instructions, the skill does not possess high-risk capabilities (such as arbitrary command execution, network operations, or file system writes) that could be exploited via this vector. The ingestion is used for legitimate contextualization of marketing strategy.
[DATA_EXPOSURE]: Code snippets for Segment and analytics platforms use standard placeholders (e.g., user_123, user@example.com). No hardcoded credentials, API keys, or sensitive system file paths are accessed or exposed.
[COMMAND_EXECUTION]: The skill does not contain or execute any shell commands, scripts, or system-level operations. It focuses on providing documentation and generating implementation code for the user to review and deploy manually.
[REMOTE_CODE_EXECUTION]: No external dependencies, package managers (npm, pip), or remote scripts are downloaded or executed. All logic is contained within the markdown instructions.

analytics-tracking