In Claude Code's Skill instructions, commands wrapped in backticks are pre-executed before the content is sent to the large language model, with the execution results substituted into the corresponding position, and the execution context is the current project's working directory.