Editable Plan
Plan-as-Document Before Execution
Before the agent runs, its plan is a document you can edit. Add steps, remove steps, change order, leave comments. Execution starts only after the plan is yours, not the model’s.
Framing
The problem
Plan reviews are read-only; users cannot fix the plan, only argue with it.
The pattern
Render the plan as an editable document users sign off on before execution.
Why chat breaks here
Chat-based plan review forces users to negotiate edits in prose instead of editing directly.
Risks
Editable plans can lull users into rubber-stamp approval if structure is weak.
Avoid when
The plan is too small to warrant a document or too rigid to benefit from edits.
Use when
Plans for agents are long enough that arguing in chat costs more than direct editing.
DOPE evaluation
- Directability
- Add, remove, reorder, or annotate any step before kickoff
- Observability
- The full plan is visible as a structured document, not buried in chat
- Predictability
- Edits stick — the agent runs the plan you signed off on
- Explainability
- Each step shows its rationale, expected output, and risk class
In the wild
- Devin Interactive Planning (Cognition) — Devin 2.0 lets users modify the plan before letting the agent work autonomously; plan acts as inspectable starting point with dynamic re-planning during execution.
- Manus AI Plan (Manus) — Generates a task plan (steps + dependencies) visible and editable before execution, with live replay across phases.
- Replit Agent Plan (Replit) — Plan-mode surface is editable — user refines tasks before approval, then the agent runs against the agreed plan.
- Copilot Coding Agent Plan (GitHub) — Spec (current vs desired) and plan (which files change, how) are editable in natural language before any code is written. The successor lineage to the sunset Copilot Workspace.
- Cocoa: Co-Planning and Co-Execution with AI Agents (Academic source · Feng, Pu, Latzke et al. (Allen AI + UW, Dec 2024)) — Research argument for keeping the plan editable *through* execution — not just before it. Editable Plan is the pre-execution surface; the paper's contribution is co-execution, which is broken out into the separate Live Plan Surface pattern.
FAQ
When should I use the Editable Plan pattern?
Plans for agents are long enough that arguing in chat costs more than direct editing.
When should I avoid the Editable Plan pattern?
The plan is too small to warrant a document or too rigid to benefit from edits.
What problem does Editable Plan solve?
Plan reviews are read-only; users cannot fix the plan, only argue with it.
Why is chat the wrong fit for this?
Chat-based plan review forces users to negotiate edits in prose instead of editing directly.
Related patterns
- Often paired with: Intent Preview — Preview surfaces the plan; editing it commits the version that runs.
- Often paired with: Approval Gates — Edit upfront, then catch deviations at gates inside the actual run.
- Alternative to: Supervisor-Worker — One editable plan vs a decomposed plan handed to specialist workers.