The Agent Skills Directory

[SAFE]: No security issues detected. The skill is designed for B2B strategy and positioning. It utilizes local project files for state persistence and context, which is standard behavior for development-focused AI skills.
[PROMPT_INJECTION]: The skill utilizes the $ARGUMENTS variable at the beginning of the persona prompt. This represents a controlled ingestion of user input into the system instructions for the purpose of identifying the project name, which does not pose a security risk in this context.
[COMMAND_EXECUTION]: The skill instructions describe writing structured JSON data to the local file system in the data/gtm/ directory. These file operations are strictly scoped to the project's data directory and align with the skill's functional requirements for persisting ICP profiles and messaging frameworks.

gtm-icp