The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access detected. The skill follows standard development practices and includes human-in-the-loop gating.
[COMMAND_EXECUTION]: The skill utilizes pnpm checks to validate the implementation. This is a legitimate use of command execution within a development environment to ensure code quality and is not used for malicious purposes in this context.
[PROMPT_INJECTION]: The instructions establish strong behavioral boundaries by requiring the agent to "Implement strictly to the plan" and stating "Plan is the contract." These serve as safeguards against autonomous overreach or unintended actions.

implement