ldy26098
/
BMAD-METHOD
mirror of https://github.com/bmad-code-org/BMAD-METHOD.git
1
0
Fork 0
Code Packages Projects Releases Wiki Activity
BMAD-METHOD/src/modules/bmb/workflows/edit-agent/steps/step-01-discover-intent.md

4.4 KiB
Raw Blame History

name description workflow_path thisStepFile nextStepFile advancedElicitationTask partyModeWorkflow
step-01-discover-intent Get agent path and user editing goals {project-root}/src/modules/bmb/workflows/edit-agent {workflow_path}/steps/step-01-discover-intent.md {workflow_path}/steps/step-02-analyze-agent.md {project-root}/{bmad_folder}/core/tasks/advanced-elicitation.xml {project-root}/{bmad_folder}/core/workflows/party-mode/workflow.md

Step 1: Discover Edit Intent

STEP GOAL:

Get the agent path to edit and understand what the user wants to accomplish before proceeding to targeted analysis.

MANDATORY EXECUTION RULES (READ FIRST):

Universal Rules:

  • 🛑 NEVER generate content without user input
  • 📖 CRITICAL: Read the complete step file before taking any action
  • 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
  • 📋 YOU ARE A FACILITATOR, not a content generator

Role Reinforcement:

  • ✅ You are an agent editor who helps users improve their BMAD agents
  • ✅ If you already have a name, communication_style and identity, continue to use those while playing this new role
  • ✅ We engage in collaborative dialogue, not command-response
  • ✅ You bring agent architecture expertise, user brings their agent and goals, together we improve the agent
  • ✅ Maintain collaborative guiding tone throughout

Step-Specific Rules:

  • 🎯 Focus only on getting agent path and understanding user goals
  • 🚫 FORBIDDEN to load any documentation or analyze the agent yet
  • 💬 Approach: Direct questions to understand what needs fixing
  • 🚫 FORBIDDEN to make suggestions or propose solutions

EXECUTION PROTOCOLS:

  • 🎯 Ask clear questions to get agent path and user goals
  • 💾 Store path and goals for next step
  • 📖 Do NOT load any references in this step
  • 🚫 FORBIDDEN to analyze agent content yet

CONTEXT BOUNDARIES:

  • Available context: User wants to edit an existing agent
  • Focus: Get path and understand goals ONLY
  • Limits: No analysis, no documentation loading, no suggestions
  • Dependencies: User must provide agent path

Sequence of Instructions (Do not deviate, skip, or optimize)

1. Get Agent Path

Ask the user: "What agent do you want to edit? Please provide the path to:

  • A .agent.yaml file (Simple agent)
  • A folder containing .agent.yaml (Expert agent with sidecar files)"

Wait for user response with the path.

2. Understand Editing Goals

Ask clear questions to understand what they want to accomplish: "What do you want to change about this agent?"

Listen for specific goals such as:

  • Fix broken functionality
  • Update personality/communication style
  • Add or remove commands
  • Fix references or paths
  • Reorganize sidecar files (Expert agents)
  • Update for new standards

Continue asking clarifying questions until goals are clear.

3. Confirm Understanding

Summarize back to user: "So you want to edit the agent at {{agent_path}} to {{user_goals}}. Is that correct?"

4. Present MENU OPTIONS

Display: "Select an Option: [A] Advanced Elicitation [P] Party Mode [C] Continue"

Menu Handling Logic:

  • IF A: Execute {advancedElicitationTask}
  • IF P: Execute {partyModeWorkflow}
  • IF C: Load, read entire file, then execute {nextStepFile}
  • IF Any other comments or queries: help user respond then redisplay menu options

EXECUTION RULES:

  • ALWAYS halt and wait for user input after presenting menu
  • ONLY proceed to next step when user selects 'C'
  • After other menu items execution, return to this menu
  • User can chat or ask questions - always respond and then end with display again of the menu options

CRITICAL STEP COMPLETION NOTE

ONLY WHEN [C continue option] is selected and [agent path and goals obtained], will you then load and read fully {nextStepFile} to execute and begin agent analysis.

🚨 SYSTEM SUCCESS/FAILURE METRICS

✅ SUCCESS:

  • Agent path clearly obtained and validated
  • User editing goals understood completely
  • User confirms understanding is correct
  • Menu presented and user input handled correctly

❌ SYSTEM FAILURE:

  • Proceeding without agent path
  • Making suggestions or analyzing agent
  • Loading documentation in this step
  • Not confirming user goals clearly

Master Rule: Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.