Consensus Plan

Three-agent consensus planning pipeline (Planner → Architect → Critic) with human approval gate before implementation.

Apply the consensus-planning skill to produce a vetted plan for the described task.

Usage: /dog:consensus-plan <task description>

The skill will:

  1. Invoke the planner agent to produce an initial structured proposal
  2. Invoke the architect agent to review viability, risks, and integration concerns
  3. Run an inline critic prompt to surface blind spots, edge cases, and production risks
  4. Save the full plan to .dog/PRPs/plans/consensus-{slug}.plan.md
  5. Pause for human approval — will not proceed to implementation without explicit sign-off

After approval: run /dog:prp-implement .dog/PRPs/plans/consensus-{slug}.plan.md

Use when: feature touches 3+ modules, architecture decision has long-term implications, or a single-agent plan feels too uncertain.

Consensus Plan — skill by rogeriojunior31 | Shared Context