What problem does it solve? When users ask to modify an existing plan, AI assistants often drift into creating entirely new plans, adding unrelated features, or jumping into implementation. This Skill enforces disciplined plan revision behavior so changes stay focused on the original plan. ## Core Features & Use Cases - Revision Tag Enforcement: Ensures all plan modifications use <PLAN REVISION> tags instead of <PLAN> tags, keeping revision output distinct from new plans. - Drift Prevention: Blocks new-task drift, context loss, and implementation drift by constraining revisions to the existing plan's scope and approach. - Structured Change Tracking: Requires a clear statement of what changed, updated file lists, and updated change lists referencing the previous plan. - Use Case: A user says "Revise the plan to add caching and remove Redis." The Skill produces a focused <PLAN REVISION> block listing only those changes, without inventing new features or starting implementation. ## Quick Start Ask the assistant to revise the current plan, for example: "Revise the plan to use PostgreSQL instead of MongoDB."