The Agent Skills Directory

[SAFE]: No prompt injection, data exfiltration, or obfuscation patterns were detected. The skill instructions focus on logical task management and human alignment.
[SAFE]: The skill does not include external dependencies, remote code downloads, or unauthorized command execution. References to internal mechanisms like superpowers:finishing-a-development-branch and TodoWrite are standard for the platform.
[SAFE]: Although the skill processes external implementation plans (an indirect prompt injection surface), it explicitly mitigates this by instructing the agent to perform a "critical review" of the plan and stop immediately to seek human clarification if instructions are unclear or suspicious.

executing-plans