using-superpowers
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Prescriptive Procedural Instructions: The skill uses high-priority directives (e.g., "YOU ABSOLUTELY MUST", "not negotiable") to mandate the invocation of specialized skills before the agent provides a response. These instructions aim to override the agent's default task-handling logic to ensure it follows the framework's defined procedures.
- System Behavior Overrides: The content explicitly instructs the agent that framework-specific skills should override its default system prompt behavior where conflicts exist. To mitigate risk, the skill includes a clear hierarchy where explicit user instructions (found in files like CLAUDE.md or AGENTS.md) always maintain the highest priority.
- Tool and Platform Adaptation: The skill includes comprehensive documentation for adapting Claude-specific tool calls to other platforms like Gemini CLI, Copilot CLI, and Codex. These reference materials provide templates and mapping tables to ensure consistent tool usage across different execution environments.
Audit Metadata