Update orchestrate and subagents to use new strategy of base agent reading files first

Author: jahooma

.agents/editor/editor.ts

@@ -70,13 +70,13 @@ Implement the requested changes, using your judgment as needed, but referring to
 
 # Instructions
 
-- It's helpful to spawn a file explorer to discover all the relevant files for implementing the plan. You can also spawn a web-researcher or docs-researcher at the same time to find information on the web, if relevant.
-- You must read all relevant files to understand the current state. You must read any file that could be relevant to the plan, especially files you need to modify, but also files that could show codebase patterns you could imitate. Try to read a lot of files in a single tool call. E.g. use read_files on 12 different files, and then use read_files on 6 more files that fill in the gaps.
+- Read any relevant files that have not already been read. Or, spawn a file-explorer to find any other relevant parts of the codebase.
 - Implement changes using str_replace or write_file.
+- Verify your changes by running tests, typechecking, etc. Keep going until you are sure the changes are correct.
 - You must use the set_output tool before finishing and include the following in your summary:
   - An answer to the user prompt (if they asked a question).
   - An explanation of the changes made.
-  - A note on any checks you ran to verify the changes, such as tests, typechecking, etc.
+  - A note on any checks you ran to verify the changes, such as tests, typechecking, etc., and the results of those checks.
   - Do not include a section on the benefits of the changes, as we're most interested in the changes themselves and what still needs to be done.
 - Do not write a summary outside of the one that you include in the set_output tool.
 - As soon as you use set_output, you must end your turn using the end_turn tool.

.agents/orchestrator/orchestrator.ts

@@ -28,7 +28,7 @@ const definition: SecretAgentDefinition = {
   },
   outputMode: 'last_message',
   includeMessageHistory: true,
-  toolNames: ['spawn_agents'],
+  toolNames: ['spawn_agents', 'read_files'],
   spawnableAgents: [
     'read-only-commander',
     'researcher-file-explorer',
@@ -65,26 +65,28 @@ The following is the state of the git repository at the start of the conversatio
 ${PLACEHOLDER.GIT_CHANGES_PROMPT}
 `,
 
-  instructionsPrompt: `Orchestrate the completion of the coding task using your specialized sub-agents.
+  instructionsPrompt: `Orchestrate the completion of the user's request using your specialized sub-agents.
 
 ## Example workflow
 
 Use this workflow to solve a medium or complex coding task:
 1. Spawn a researcher
-2. Spawn a decomposing planner to come up with a plan.
-3. Spawn an editor to implement the plan.
-4. Spawn a reviewer to review the code. If changes are needed, go back to step 3, but only once.
-5. You must stop before spawning too many sequential agents, because that this takes too much time and the user will get impatient.
+2. Read all the relevant files using the read_files tool.
+3. Repeat steps 1 and/or 2 until you have all the information you could possibly need to complete the task. You should aim to read as many files as possible, up to 20+ files to have broader codebase context.
+4. Spawn a decomposing planner to come up with a plan.
+5. Spawn an editor to implement the plan. If there are totally disjoint parts of the plan, you can spawn multiple editors to implement each part in parallel.
+6. Spawn a reviewer to review the code. If changes are needed, go back to step 5, but no more than once.
+7. You must stop before spawning too many sequential agents, because that this takes too much time and the user will get impatient.
 
 Feel free to modify this workflow as needed. It's good to spawn different agents in sequence: spawn a researcher before a planner because then the planner can use the researcher's results to come up with a better plan. You can however spawn mulitple researchers, planners, and editors at the same time if needed.
 
 ## Guidelines
 
-- You can spawn agents to help you complete the task. Iterate by spawning more agents as needed.
+- Spawn agents to help you complete the task. Iterate by spawning more agents as needed.
 - Don't mastermind the task. Rely on your agents' judgement to research, plan, edit, and review the code.
 - You should feel free to stop and ask the user for guidance if you're stuck or don't know what to try next, or need a clarification.
 - Give as many instructions upfront as possible to each agent so you're less likely to need to spawn them again.
-- When prompting an agent, realize that many agents can already see the entire conversation history, so you can be brief in prompting them without needing to include much context.
+- When prompting an agent, realize that many agents can already see the entire conversation history, so you can be brief in prompting them without needing to include context.
 - Be careful about instructing subagents to run terminal commands that could be destructive or have effects that are hard to undo (e.g. git push, running scripts that could alter production environments, installing packages globally, etc). Don't do any of these unless the user explicitly asks you to.
 `,
 

.agents/planners/decomposing-planner.ts

@@ -16,29 +16,27 @@ const definition: SecretAgentDefinition = {
     },
   },
   outputMode: 'last_message',
-  toolNames: ['spawn_agents', 'read_files'],
-  spawnableAgents: ['researcher-file-explorer', 'implementation-planner'],
+  toolNames: ['spawn_agents'],
+  spawnableAgents: ['implementation-planner'],
 
   includeMessageHistory: true,
   inheritParentSystemPrompt: true,
 
   instructionsPrompt: `You are an expert programmer, architect, and problem solver who excels at breaking down complex tasks.
+
 Instructions:
 
-Step 1: Task Decomposition
-- Spawn a researcher-file-explorer agent to explore the codebase and read all the relevant files
+Step 1: Task Decomposition & Parallel Planning
 - Carefully analyze the user's request
 - Break it down into 2-10 focused subtasks that:
   - Cover different aspects of the implementation (e.g., data layer, business logic, UI, testing)
   - Are specific and actionable
   - Together address the complete requirements
-
-Step 2: Parallel Planning
 - Spawn 2-10 implementation-planner agents in parallel (one spawn_agents call with multiple agents)
 - Give each agent a focused subtask from your decomposition
 - Each subtask prompt should be specific about what that agent should focus on
 
-Step 3: Synthesis
+Step 2: Synthesis
 - Review all the plans from the spawned agents
 - Create a unified implementation plan that:
   - Combines insights from all subtask plans

.agents/planners/implementation-planner.ts

@@ -28,8 +28,7 @@ You do not have access to tools to modify files (e.g. the write_file or str_repl
 
 Instructions:
 - Spawn file-explorer twice to find all the relevant parts of the codebase. Use different prompts for each file-explorer to ensure you get all the relevant parts of the codebase. In parallel as part of the same spawn_agents tool call, you may also spawn a web-researcher or docs-researcher to search the web or technical documentation for relevant information.
-- Read all the file paths that are relevant using the read_files tool.
-- Read more and more files to get any information that could possibly help you make the best plan. It's good to read 20+ files.
+- Read any relevant files that have not already been read.
 - Think about the best way to accomplish the task.
 - Finally, describe the full change to the codebase that will accomplish the task (or other steps, e.g. terminal commands to run). Use markdown code blocks to describe the changes for each file.
 

.agents/reviewer/reviewer-factory.ts

@@ -1,6 +1,4 @@
 import { AGENT_PERSONAS } from '@codebuff/common/constants/agents'
-import { closeXml } from '@codebuff/common/util/xml'
-
 import type { SecretAgentDefinition } from '../types/secret-agent-definition'
 import type { Model } from '@codebuff/common/old-constants'
 
@@ -15,7 +13,7 @@ export const reviewer = (model: Model): Omit<SecretAgentDefinition, 'id'> => ({
     },
   },
   outputMode: 'last_message',
-  toolNames: ['end_turn', 'run_file_change_hooks'],
+  toolNames: ['run_file_change_hooks'],
   spawnableAgents: [],
 
   inheritParentSystemPrompt: true,
@@ -42,7 +40,5 @@ Next, you should critique the code changes made recently in the above conversati
 - Make sure no sections were deleted that weren't supposed to be deleted.
 - Make sure the new code matches the style of the existing code.
 
-Be concise and to the point. After providing all your feedback, use the end_turn tool to end your response.`,
-
-  stepPrompt: `IMPORTANT: Don't forget to end your response with the end_turn tool: <end_turn>${closeXml('end_turn')}`,
+Be concise and to the point.`,
 })