using-superpowers

To enforce a plan-first workflow for coding agents, you need a mandatory execution protocol that checks for relevant skills and prioritizes planning before any code generation. This prevents the agent from rationalizing skipped workflows or jumping straight into implementation.

AI agents skip planning and jump straight into code due to inconsistent or undisciplined behavior lacking a mandatory skill-aware execution hierarchy. Forcing the agent to adopt a structured, plan-first protocol ensures it invokes brainstorming and planning skills before responding to complex requests.

The best way to automate TDD and YAGNI discipline in agentic workflows is activating a mandatory skill-aware execution protocol at the start of your session. This automatically invokes relevant engineering skills, ensuring implementation follows structured testing and design principles rather than ad-hoc coding.

Yes, mandatory skill invocation requires the presence of the Skill tool or an equivalent platform-specific activation mechanism to load and execute operational instructions. This tool ensures the agent checks for relevant skills before responding, preventing skipped engineering workflows.

Yes, you can override the automated planning workflow because the execution protocol establishes a clear hierarchy where user instructions take precedence. Direct user instructions are prioritized first, followed by Superpowers skills, and finally the default system prompt.

You should not use a mandatory plan-driven execution protocol for simple task-based interactions that do not require agentic discipline or adherence to predefined engineering workflows. The protocol is designed for complex coding projects where skipping brainstorming, planning, or TDD principles would be detrimental.