The Agent Skills Directory

[SAFE]: The skill acts as a procedural guide for manual or agent-assisted code refactoring and does not include any automated scripts, network requests, or access to sensitive local files.- [COMMAND_EXECUTION]: While the skill references shell commands for testing and building (e.g., <test-the-shared-package>), these are placeholders designed for the agent to resolve based on the specific project context (such as instructions in a CLAUDE.md file). No hardcoded or malicious commands are included.- [PROMPT_INJECTION]: The instructions are consistently focused on refactoring logic and do not contain patterns intended to override agent safety guidelines or bypass operational constraints.- [DATA_EXFILTRATION]: There are no patterns involving data collection or transmission to external domains. The metadata source link points to a repository associated with the author and is not used for functional network operations.

refactor-extract