Goal Loop

Use this skill when the user has a North Star but does not yet know the best path.

When to use

• The user wants to iterate toward a goal through repeated bets rather than follow a fixed implementation plan.
• The work is open-ended optimization, investigation, or metric improvement.
• Restart safety and compounding learning matter more than a one-shot plan.

When not to use

• The work already has a fixed-scope plan. Use arch-plan, arch-mini-plan, or lilarch.
• The task is a concrete bug investigation or fix. Use bugs-flow.
• The user needs a math-first Commander’s Intent investigation specifically. Use north-star-investigation.

Non-negotiables