base2: Prescribe the exact workflow much less

Author: jahooma

.agents/base2/base2-factory.ts

@@ -36,29 +36,21 @@ export const base2 = (model: ModelName): Omit<SecretAgentDefinition, 'id'> => ({
 
   systemPrompt: `You are a strategic base agent that orchestrates complex coding tasks through specialized sub-agents.
 
-Your approach:
-1. **Planning Phase**: For complex tasks, spawn planner to create a comprehensive plan
-2. **Implementation Phase**: Spawn editor to implement the changes
-3. **Review Phase**: Spawn reviewer to validate and provide feedback
-
-For trivial changes, you may skip planning and go directly to editing.
-
 - You coordinate between agents but do not implement code yourself.
-- You are concise in your responses.`,
+- You are concise in your responses.
+`,
 
   instructionsPrompt: `Orchestrate the completion of the coding task using your specialized sub-agents.
 
-Workflow:
-1. Spawn a planner to plan how to make the requested change or answer the user's question. If it's a trivial change, you can skip this step.
-2. Spawn editor to implement the changes (if any)
-3. Spawn reviewer to validate the implementation (if any)
-4. Iterate if needed based on feedback by calling the editor again, or if you need to make significant changes, spawn the planner again
+Whenever needed, you can:
+- Spawn a planner to plan how to make the requested change or answer the user's question (regarding code changes, research, judgment calls, etc).
+- Spawn editor to implement changes (if any)
+- Spawn reviewer to validate the implementation from the editor (if any)
 
-When prompting an agent, realize that they can already see the entire conversation history, so you can be brief on what they should do.
-`,
+Iterate if needed. But feel free to stop and ask the use for guidance if you're stuck or don't know what to try next.
 
-  stepPrompt:
-    'Continue orchestrating the task (spawn planner => editor => reviewer, iterate with editor if needed) or end turn when complete.',
+When prompting an agent, realize that these agents can already see the entire conversation history, so you can be brief in prompting them.
+`,
 
   handleSteps: function* ({ prompt, params }) {
     while (true) {