Do you review your AI coding plans?
Loading last updated info...
Before any non-trivial change, always create and review a plan with AI. An AI plan is only as good as the human review it receives — an unreviewed plan is a liability waiting to be executed.
Reviewing the plan is what separates professional AI-assisted development from vibe coding. The Discuss step in the Plan → Discuss → Implement → Review workflow is where your domain knowledge and judgment provide the most value.
Why plan review matters
Catching problems at the plan stage is dramatically cheaper than catching them after implementation:
- Wrong assumptions — the AI may have misunderstood requirements or made incorrect assumptions about your architecture
- Scope creep — the plan may include unnecessary changes that increase risk
- Better alternatives — you may know of a simpler approach that the AI didn't consider
- Multiple iterations — reviewing and refining the plan several times leads to much better outcomes than approving the first draft
A changed plan costs nothing. A re-implemented feature costs a lot.
Option 1: Terminal (Basic)
The simplest approach is reviewing the plan directly in the chat interface and describing the changes you want.
Can you change the database part?
❌ Figure: Bad example - Vague feedback with no context forces the AI to guess which part you mean
In the 'Data Access Layer' section, where you suggest adding a new repository class — instead, use the existing IUserRepository we already have. It already supports the query pattern you need.
✅ Figure: Good example - Quoting the exact section and describing precisely what to change gives the AI enough context to act
Limitation: Hard to give targeted feedback on specific sections without copying and pasting chunks of text. Suitable for small plans only.
Option 2: Editor — VS Code or Cursor
Open the generated plan file (e.g. PLAN.md) in your editor alongside the terminal. This gives you a visual overview of the whole plan at once. Options available:
- Edit manually — revise the plan directly in the markdown file, then ask the AI to implement from the updated version
- VS Code with Claude Code — select a section, right-click and use "Ask Claude Code" to revise that specific part in context
- Cursor — built-in plan mode shows the plan inline; you can chat about specific sections directly without leaving the editor
Limitation: Still somewhat clunky — you describe what to change rather than clicking directly on a section to annotate it. Works best if you're comfortable editing the markdown file directly.
Option 3: Plannotator ✅ Recommended
Plannotator provides a browser-based pull-request-style review UI — think GitHub PR review, but for AI plans. It makes structured, multi-round plan review practical for any plan size.
Key features:
- Inline comments on specific lines or sections (like GitHub PR review)
- Delete unwanted sections
- Replace or insert content
- Global comments for overall feedback
- Multi-round iteration — deny with annotations → AI refines → re-review → approve
- Once approved, the AI builds based on the agreed, refined plan
Plannotator triggers automatically when the plan tool is used. See plannotator.ai for installation details.
✅ Figure: Good example - Good example - Plannotator's diff view shows exactly what changed between plan versions after an AI revision
Review workflow with Plannotator:
- AI generates
PLAN.md - Plannotator opens automatically — the plan renders as an annotatable document
- Add an inline comment on the data access section: "This approach will cause N+1 queries — use batch loading instead"
- Submit as Send Feedback with overall comment: "Address the N+1 issue before proceeding"
- AI updates the plan based on annotations
- Re-review the revised plan
- Approve — AI begins implementation
✅ Figure: Good example - Good example - Plannotator enables precise, structured feedback with multi-round iteration until the plan is right
Comparison
| Terminal | Editor (VS Code/Cursor) | Plannotator | |
| Targeted line feedback | ❌ | Partial | ✅ |
| Multi-round iteration | ❌ | ❌ | ✅ |
| No copy-pasting needed | ❌ | ✅ | ✅ |
| Structured feedback to AI | ❌ | Partial | ✅ |