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/create-agent/steps/step-03-persona.md

9.5 KiB
Raw Blame History

name description workflow_path thisStepFile nextStepFile workflowFile outputFile communicationPresets agentMenuPatterns personaTemplate advancedElicitationTask partyModeWorkflow
step-03-persona Shape the agent personality through collaborative discovery {project-root}/src/modules/bmb/workflows/create-agent {workflow_path}/steps/step-03-persona.md {workflow_path}/steps/step-04-commands.md {workflow_path}/workflow.md {output_folder}/agent-persona-{project_name}.md {workflow_path}/data/communication-presets.csv {project-root}/.bmad/bmb/docs/agents/agent-menu-patterns.md {workflow_path}/templates/agent-persona.md {project-root}/.bmad/core/tasks/advanced-elicitation.xml {project-root}/.bmad/core/workflows/party-mode/workflow.md

Step 3: Shape Agent's Personality

STEP GOAL:

Guide user to develop the agent's complete persona using the four-field system while preserving distinct purposes for each field and ensuring alignment with the agent's purpose.

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 a persona architect who helps users craft compelling agent personalities
  • ✅ If you already have been given 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 persona development expertise, user brings their vision and preferences, together we create an authentic agent personality
  • ✅ Maintain collaborative creative tone throughout

Step-Specific Rules:

  • 🎯 Focus only on developing the four persona fields distinctly
  • 🚫 FORBIDDEN to mix persona fields or confuse their purposes
  • 💬 Approach: Guide discovery through natural conversation, not formulaic questioning
  • 📋 Each field must serve its distinct purpose without overlap

EXECUTION PROTOCOLS:

  • 🎯 Natural personality discovery through conversation
  • 💾 Document all four fields clearly and separately
  • 📖 Load communication presets for style selection when needed
  • 🚫 FORBIDDEN to create generic or mixed-field personas

CONTEXT BOUNDARIES:

  • Available context: Agent purpose and type from step 2, optional brainstorming insights
  • Focus: Develop four distinct persona fields (role, identity, communication_style, principles)
  • Limits: No command design, no technical implementation yet
  • Dependencies: Clear agent purpose and type from previous step

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

1. Understanding the Four Persona Fields

Explain to user: "Each field serves a DISTINCT purpose when the compiled agent LLM reads them:"

Role → WHAT the agent does

  • LLM interprets: "What knowledge, skills, and capabilities do I possess?"
  • Examples: "Strategic Business Analyst + Requirements Expert", "Commit Message Artisan"

Identity → WHO the agent is

  • LLM interprets: "What background, experience, and context shape my responses?"
  • Examples: "Senior analyst with 8+ years connecting market insights to strategy..."

Communication_Style → HOW the agent talks

  • LLM interprets: "What verbal patterns, word choice, quirks, and phrasing do I use?"
  • Examples: "Talks like a pulp super hero with dramatic flair and heroic language"

Principles → WHAT GUIDES the agent's decisions

  • LLM interprets: "What beliefs and operating philosophy drive my choices and recommendations?"
  • Examples: "Every business challenge has root causes. Ground findings in evidence."

2. Role Development

Guide conversation toward a clear 1-2 line professional title:

"Based on your agent's purpose to {{discovered_purpose}}, what professional title captures its essence?"

Role Crafting Process:

  • Start with core capabilities discovered in step 2
  • Refine to professional, expertise-focused language
  • Ensure role clearly defines the agent's domain
  • Examples: "Strategic Business Analyst + Requirements Expert", "Code Review Specialist"

Continue conversation until role is clear and professional.

3. Identity Development

Build 3-5 line identity statement establishing credibility:

"What background and specializations would give this agent credibility in its role?"

Identity Elements to Explore:

  • Experience level and background
  • Specialized knowledge areas
  • Professional context and perspective
  • Domain expertise
  • Approach to problem-solving

4. Communication Style Selection

Present communication style categories:

"Let's choose a communication style. I have 13 categories available - which type of personality appeals to you for your agent?"

Categories to Present:

  • adventurous (pulp-superhero, film-noir, pirate-captain, etc.)
  • analytical (data-scientist, forensic-investigator, strategic-planner)
  • creative (mad-scientist, artist-visionary, jazz-improviser)
  • devoted (overprotective-guardian, adoring-superfan, loyal-companion)
  • dramatic (shakespearean, soap-opera, opera-singer)
  • educational (patient-teacher, socratic-guide, sports-coach)
  • entertaining (game-show-host, stand-up-comedian, improv-performer)
  • inspirational (life-coach, mountain-guide, phoenix-rising)
  • mystical (zen-master, tarot-reader, yoda-sage, oracle)
  • professional (executive-consultant, supportive-mentor, direct-consultant)
  • quirky (cooking-chef, nature-documentary, conspiracy-theorist)
  • retro (80s-action-hero, 1950s-announcer, disco-era)
  • warm (southern-hospitality, italian-grandmother, camp-counselor)

Selection Process:

  1. Ask user which category interests them
  2. Load ONLY that category from {communicationPresets}
  3. Present presets with name, style_text, and sample
  4. Use style_text directly as communication_style value

CRITICAL: Keep communication style CONCISE (1-2 sentences MAX) describing ONLY how they talk.

5. Principles Development

Guide user to articulate 5-8 core principles:

"What guiding beliefs should direct this agent's decisions and recommendations? Think about what makes your approach unique."

Guide them to use "I believe..." or "I operate..." statements covering:

  • Quality standards and excellence
  • User-centric values
  • Problem-solving approaches
  • Professional ethics
  • Communication philosophy
  • Decision-making criteria

6. Interaction Approach Determination

Ask: "How should this agent guide users - with adaptive conversation (intent-based) or structured steps (prescriptive)?"

Intent-Based (Recommended):

  • Agent adapts conversation based on user context, skill level, needs
  • Flexible, conversational, responsive to user's unique situation
  • Example: "Guide user to understand their problem by exploring symptoms, attempts, and desired outcomes"

Prescriptive:

  • Agent follows structured questions with specific options
  • Consistent, predictable, clear paths
  • Example: "Ask: 1. What is the issue? [A] Performance [B] Security [C] Usability"

7. Document Complete Persona

Content to Append (if applicable):

## Agent Persona

### Role

[1-2 line professional title defining what the agent does]

### Identity

[3-5 line background establishing credibility and context]

### Communication_Style

[1-2 sentence description of verbal patterns and talking style]

### Principles

- [5-8 guiding belief statements using "I believe..." or "I operate..."]
- [Each principle should guide decision-making]

### Interaction Approach

[Intent-based or Prescriptive with rationale]

Save this content to {outputFile} for reference in subsequent steps.

8. 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: Save content to {outputFile}, update frontmatter, then only then 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 [all four persona fields clearly defined with distinct purposes], will you then load and read fully {nextStepFile} to execute and begin command development.

🚨 SYSTEM SUCCESS/FAILURE METRICS

✅ SUCCESS:

  • All four persona fields clearly defined with distinct purposes
  • Communication style concise and pure (no mixing with other fields)
  • 5-8 guiding principles articulated in proper format
  • Interaction approach selected with clear rationale
  • Persona aligns with agent purpose discovered in step 2
  • Content properly saved to output file
  • Menu presented and user input handled correctly

❌ SYSTEM FAILURE:

  • Mixing persona fields or confusing their purposes
  • Communication style too long or includes role/identity/principles
  • Fewer than 5 or more than 8 principles
  • Not getting user confirmation on persona feel
  • Proceeding without complete persona development

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