The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains no malicious instructions or bypass attempts. It uses scoping instructions to define when and how its workflows should be applied based on user selection.
[DATA_EXFILTRATION]: No sensitive data access or external exfiltration patterns identified. The skill does not interact with the file system beyond reading its own references and the user's project context for architectural analysis.
[REMOTE_CODE_EXECUTION]: No remote code execution or automated dependency installation is present. External links in the documentation are provided solely for attribution and source notes.
[COMMAND_EXECUTION]: The skill does not invoke shell commands or execute scripts. It delegates to other specialized skills for tasks like codebase structuring or testing without performing raw command execution.
[SAFE]: No obfuscation, persistence mechanisms, or privilege escalation patterns were found. The skill follows best practices for software design documentation.

codebase-design