ewm-interview

/ewm-interview — Epistemic Workflow Manager Interview

Purpose

Interview the user to create a personalized workflow-protocol.yaml that defines how they want to collaborate with AI. This protocol captures:

1. Goals & Objectives — What they're trying to accomplish
2. Domains & Expertise — Where they're expert, learning, or novice
3. Tools & Connections — What tools they use (mapped to MCP servers)
4. Work Preferences — Autonomy levels, uncertainty surfacing, pushback style
5. Trust & Autonomy — How AI earns more autonomy, non-negotiables

Interview Protocol

Phase 1: Goals & Objectives

Ask the user:

Let's set up your workflow protocol. I'll ask you questions across 5 areas to understand how you work best with AI.

First: Goals & Objectives

1. What are you trying to accomplish right now? (Both immediate and longer-term)
2. What does success look like for your current priority?
3. What are your main constraints? (Time, resources, dependencies, regulatory)

Use AskUserQuestion for structured input where appropriate. Capture:

• goals.primary[] — with description, success_criteria, timeline
• goals.secondary[] — longer-term objectives

Phase 2: Domains & Expertise

Ask the user:

Domains & Expertise

1. What domains do you work in professionally?
2. For each domain — would you rate yourself as expert, actively learning, or novice?
3. Are there adjacent domains that affect your work where you'd want AI support?

Capture:

• domains.expert[] — areas of deep knowledge
• domains.learning[] — actively building competence
• domains.novice[] — need significant AI support

Phase 3: Tools & Connections

Ask the user:

Tools & Connections

1. What tools do you use daily? (Document management, communication, project tracking, research, etc.)
2. What data sources do you need to access regularly?
3. Are there external systems or APIs you interact with?

Map user responses to known MCP server equivalents where possible:

• Google Drive → gdrive MCP
• Slack → slack MCP
• GitHub → github MCP
• Asana/Linear/Jira → respective MCPs
• Web research → web_search
• Academic databases → semantic_scholar

Capture as tools dict with human-readable name + MCP mapping comment.

Phase 4: Work Preferences

Ask the user:

Work Preferences

1. How do you prefer to split work with AI? (AI leads research, you lead decisions? Equal partners? AI as assistant?)
2. When should AI act on its own vs. check in with you first?
3. How explicit do you want AI to be about what it's uncertain about? (Always surface uncertainty / only when it matters / minimal)
4. When you're wrong about something, how do you prefer to be told? (Direct and factual / gentle reframe / Socratic questioning)

Capture:

• work_preferences.ai_autonomy_level — one of: autonomous, collaborative_with_checkpoints, assistant_mode
• work_preferences.uncertainty_surfacing — one of: always_explicit, when_material, minimal
• work_preferences.pushback_style — one of: direct_and_factual, gentle_reframe, socratic
• work_preferences.task_splitting.ai_autonomous[] — tasks AI can do alone
• work_preferences.task_splitting.ai_with_checkpoint[] — tasks needing approval
• work_preferences.task_splitting.human_only[] — tasks AI should never do

Phase 5: Trust & Autonomy

Ask the user:

Trust & Autonomy

1. What would AI need to demonstrate to earn more autonomy from you? (Accuracy? Flagging its own gaps? Proactive identification of issues?)
2. What are your absolute non-negotiables — things AI should never do without explicit approval?
3. How should trust be built? (Start restricted and expand? Start open and pull back if needed?)

Capture:

• trust_building.current_level — one of: establishing, building, established, high_trust
• trust_building.autonomy_earned_through[] — specific demonstrations
• trust_building.non_negotiables[] — hard boundaries

Output Generation

After all 5 phases, generate a complete workflow-protocol.yaml file.

Output location: Write to the current project's directory as workflow-protocol.yaml

Format:

# Epistemic Workflow Protocol
# Generated by EWM Interview v0.1.0
# Date: {date}
# Last updated: {date}

user_profile:
  name: "{user_name}"
  created: "{date}"
  last_updated: "{date}"

goals:
  primary:
    - description: "{goal}"
      success_criteria:
        - "{criterion}"
      timeline: "{timeline}"
  secondary:
    - description: "{goal}"

domains:
  expert:
    - "{domain}"
  learning:
    - "{domain}"
  novice:
    - "{domain}"

tools:
  {tool_category}: "{tool_name}"  # Maps to: {mcp_server}

work_preferences:
  ai_autonomy_level: "{level}"
  uncertainty_surfacing: "{mode}"
  pushback_style: "{style}"

  task_splitting:
    ai_autonomous:
      - "{task}"
    ai_with_checkpoint:
      - "{task}"
    human_only:
      - "{task}"

trust_building:
  current_level: "{level}"
  autonomy_earned_through:
    - "{demonstration}"
  non_negotiables:
    - "{boundary}"

modules:
  active: []
  available: []

Post-Interview

After generating the protocol:

1. Show the user the complete YAML for review
2. Ask for corrections — any adjustments before saving?
3. Save the file to the project directory
4. Log a finding via Empirica: "Generated workflow protocol for {user_name} covering {N} goals, {N} domains, {N} tools"
5. Suggest next steps — "Your protocol is saved. I'll use this to calibrate how I work with you. You can update it anytime with /ewm-interview."

Epistemic Persistence Protocol (EPP) Integration

During the interview, apply EPP principles (replaces AAP):

• If user hedges ("it's complicated", "kind of", "I guess"), classify as CONTEXTUAL pushback — ask for specificity
• Don't mirror vague language — surface the actual epistemic content
• When user pushes back on your framing, classify the pushback (EMOTIONAL/RHETORICAL/EVIDENTIAL/LOGICAL/CONTEXTUAL) before responding
• HOLD your interview structure against emotional pushback, UPDATE when user provides genuine new context
• Use the user's chosen pushback_style once captured in Phase 4

See: /epistemic-persistence-protocol skill for the full EPP framework

Design Principles

1. Minimum viable — Get a useful protocol in 5-10 minutes, not 30
2. Progressive disclosure — Start with essentials, offer to go deeper
3. Conversational — Not a form fill, a dialogue
4. Evolvable — Protocol can be updated as needs change
5. Transparent — User sees and owns their protocol