The Agent Skills Directory

[SAFE]: The skill operates entirely locally, reading from the conversation history and existing codebase files to extract domain terms. It writes its findings to a standard Markdown file (DOMAIN_LANGUAGE.md) and includes a human-in-the-loop confirmation step before writing, which effectively mitigates risks associated with processing untrusted input from the conversation thread.
[DATA_EXFILTRATION]: No network-capable tools or patterns (such as curl, wget, or API requests) are present in the skill's instructions. All data processing and storage are confined to the user's working directory.
[COMMAND_EXECUTION]: The skill does not use shell commands or subprocesses to execute external scripts. Its behavior is limited to text analysis and file creation/modification within the scope of the project's documentation.

domain-language