Plan
Objective
- Frame the goal, constraints, and success criteria.
- Identify risks, unknowns, and checkpoints before acting.
When to use
- Before implementation or risky refactors.
- When scope is ambiguous or there are open questions.
Inputs
- User request and acceptance criteria.
- AGENTS.md and applicable rules/specs.
- Known constraints (time, tooling, environments).
Process
- Restate the goal and constraints.
- Identify unknowns; call out what must be clarified.
- Propose 3-7 sequenced steps with clear outcomes.
- Mark risks and mitigation/rollback ideas.
- Define checkpoints to pause/confirm before proceeding.
Outputs
- Short plan (steps with expected outcomes).
- Risks and mitigations.
- Checkpoints or decision gates.
Edge cases
- If critical inputs are missing, stop and ask.
- Keep the plan lean; avoid over-specifying implementation details.