The Agent Skills Directory

[SAFE]: The skill's primary function is to resolve design decisions through structured user interaction. It explicitly mandates that the agent must not act on synthesized plans until the user provides confirmation, establishing a robust human-in-the-loop safety control.
[SAFE]: The instructions emphasize looking up factual information in the codebase rather than making assumptions. While codebase reading is an ingestion point for external data, the requirement for user approval of any resulting plan mitigates the risk of indirect prompt injection.
[SAFE]: The skill encourages the use of platform-native structured input tools (e.g., AskUserQuestion, request_user_input) for all user-facing questions, ensuring interactions are handled through established and managed communication channels.
[SAFE]: No malicious patterns, including prompt injection, credential exfiltration, remote code execution, or persistence mechanisms, were found in the skill metadata or instructions.

shape