The Agent Skills Directory

[SAFE]: The skill utilizes read-only tools including Read, Grep, and Glob to analyze the local codebase. This is limited to gathering technical context for the decision-making process and does not involve any data exfiltration or sensitive file access.
[SAFE]: The skill employs the Agent tool to delegate tasks to sub-agents acting as advocates. These sub-agents are explicitly instructed not to edit any files and are restricted to information gathering, which aligns with the principal skill's read-only scope.
[SAFE]: No patterns of prompt injection, obfuscation, or credential exposure were detected. The skill's instructions focus on logical reasoning, context gathering, and objective synthesis of technical trade-offs.

decide