using-superpowers

To ensure agent skill discovery, enforce a mandatory pre-response skill check using the platform's skill tool before answering or asking clarifying questions. This structured process prevents ad-hoc actions by loading applicable domain-specific skills at conversation start or before planning steps.

Agent workflow priority rules establish a strict order: user instructions take the highest priority, followed by superpowers skills, and finally the default system prompt. This hierarchy ensures explicit user commands override predefined skill behaviors and avoid operational conflicts.

Mandatory skill invocation works across platforms like Claude Code, Copilot CLI, and Gemini through platform adaptation and tool mapping. These mappings provide the necessary guidance so skill discovery and loading mechanisms operate reliably across different agent environments.

Trigger a skill discovery check at conversation start and before entering planning or action steps in agent workflows. This timing ensures domain-specific skills that may change behavior are loaded before the agent produces any response or executes a task.

Skipping pre-response skill discovery causes agents to miss opportunities to apply structured processes, platform-specific tools, or established checklists. This leads to ad-hoc or unsafe actions, bypassing specialized behaviors that should guide the agent's task execution.

Yes, the agent workflow supports on-demand skill content loading. This allows the agent to discover applicable skills initially and load the actual content dynamically when needed during planning or action steps, optimizing context usage.