The Agent Skills Directory

[SAFE]: The skill uses a highly structured workflow that requires explicit user approval before proceeding to subsequent phases, significantly reducing the risk of autonomous malicious behavior.
[COMMAND_EXECUTION]: The skill utilizes shell commands for standard development tasks, such as mkdir for directory creation and tsc, lint, and test for code verification. These actions are appropriate for the intended use case of software development.
[DATA_EXPOSURE]: The skill manages specifications within a local ./.specs/ directory. No patterns of sensitive file access (e.g., credentials, SSH keys) or external data exfiltration were detected.
[REMOTE_CODE_EXECUTION]: The skill explicitly instructs the agent to propose commit messages rather than executing git commands directly, maintaining user control over the repository state.

sdd-lite