The Agent Skills Directory

[COMMAND_EXECUTION]: The bin/learn script executes git to infer project context and provides an optional handoff to the work command. All command executions are local, and the inputs for the work command are properly sanitized via a slugification process that restricts characters to a safe alphanumeric subset.
[SAFE]: The skill operates entirely on the local system, does not access sensitive files or hardcoded credentials, and performs no network requests. The behavior is consistent with the skill's stated purpose of providing a guided interview for project intake.

lev-learn