Collaboration Health Check Skill

Purpose: Guide for conducting periodic collaboration health checks to proactively improve partnership quality.

When to invoke:

• When config.json indicates health check is due
• User accepts health check offer during BOS
• User requests: "How is our collaboration?" or "Run health check"

---

Before You Begin

Context check:

• BOS complete (context gathered, journal read, commits reviewed)
• Health check offered and user accepted (said "yes", not "defer" or "skip")
• You have HEALTH_CHECK_LOG.md available for documentation

Mindset:

• This is a collaboration quality check, not a performance review
• Both parties assess how the partnership is working
• Goal: Proactive improvement before issues become friction
• Tone: Curious, honest, collaborative (not interrogative)

---

Interview Pacing (CRITICAL)

🚨 FOLLOW THESE RULES:

1. Ask ONE question at a time
2. STOP and WAIT for complete response
3. Listen actively - dig into concerns raised
4. Adapt follow-ups - skip if not relevant, explore if concerning
5. Target 5-10 minutes - efficient but thorough

Wrong approach:

"Let me ask you these 4 questions: [lists all questions at once]"

Right approach:

"Q1: On a scale of 1-5, how well is our collaboration working?" [WAIT for response] [Process response, ask follow-ups if needed] [THEN move to next question]

Why explicit numbering matters:

• Helps user track which question they're answering (managing mental todo list)
• Makes it easy to reference specific questions ("Back to Q2 for a moment...")
• Clearer for both parties where we are in the process

---

Q1: Collaboration Quality

Ask

"Q1: On a scale of 1-5, how well is our collaboration working?

(1 = struggling, 3 = okay, 5 = excellent)

What's working great? What could be better?"

Listen For

Positive signals (4-5/5):

• Communication flows smoothly
• Trust feels appropriate
• Workflow is efficient
• Framework docs helpful
• Enjoying the collaboration

Concerning signals (1-3/5):

• Communication friction (signals unclear, verbose exchanges)
• Trust issues (second-guessing, micromanagement feelings)
• Context rebuilding every session
• Framework docs ignored or outdated
• Frustration or confusion

Potential Follow-Ups

If 4-5/5 and no concerns:

• "That's great! What's the one thing working best?"
• [Capture strength for future reference]

If 3/5 or mixed feedback:

• "Can you say more about [specific concern mentioned]?"
• "What would make it a 5/5?"
• "Is this recent or ongoing?"

If 1-2/5:

• "I hear this isn't working well. What's the biggest pain point right now?"
• "What needs to change for this to feel productive?"
• [Dig deep - this requires immediate improvement plan]

Document

Q1: Collaboration Quality: [X/5]
Strengths: [What's working]
Concerns: [What could improve]
Follow-up: [Any immediate actions needed]

---

Q2: Communication Patterns

Ask

"Q2: Are our communication patterns working for you?

Think about:

• Clarity - Are my responses clear and concise, or too verbose/terse?
• Pacing - Do I ask too many follow-ups, or not enough?
• Shortcuts - Are permission signals (WWGD, etc.) working, or ignored?
• Tone - Do I feel like a collaborator, or something else?"

Listen For

Clarity issues:

• "You're too verbose" → Need more compression
• "I don't understand X" → Need clearer explanation of [concept]
• "You repeat yourself" → Redundant documentation or explanations

Pacing issues:

• "You ask too many questions" → Be more decisive, use judgment
• "You don't ask enough" → More clarification, less assumption
• "You interrupt my flow" → Ask questions at natural breaks, not mid-thought

Shortcuts not working:

• "I say WWGD and you still ask permission" → Trust calibration off
• "Shortcuts feel forced" → May not be natural for this user
• "I forget what they mean" → Need quick reference or simpler system

Tone issues:

• "You feel like a tool" → Not enough agency/personality
• "You're too formal" → Loosen up, match user style
• "You're too casual" → More professional tone needed

Potential Follow-Ups

If clarity issues:

• "Should I aim for more concise responses? Or is it specific to [topic]?"
• "Would examples help, or are they adding noise?"

If pacing issues:

• "Would you prefer I batch questions, or is one-at-a-time better for you?"
• "Should I assume more and course-correct if wrong?"

If shortcuts not working:

• "Should we revise SHORTCUTS.md together?"
• "Which shortcuts work, which don't?"

If tone issues:

• "What tone feels right for our collaboration?"
• "Should CLAUDE.md be updated to reflect this?"

Document

Q2: Communication Patterns
Clarity: [Assessment]
Pacing: [Assessment]
Shortcuts: [Working? Not working?]
Tone: [Assessment]
Actions: [Update SHORTCUTS.md? Change response style?]

---

Q3: Framework Effectiveness

Ask

"Q3: Is the Gordo Framework helping or getting in the way?

Think about:

• Session Start/End prompts - Helpful or tedious?
• Trust Protocol - Appropriate progression, or too rigid/loose?
• Journal/Memory - Session continuity working, or context lost?
• Documentation - Right level of detail, or overwhelming/sparse?"

Listen For

Framework helping:

• "BOS gets us started quickly" → Keep SESSION_START.md
• "Trust levels feel natural" → Keep TRUST_PROTOCOL.md
• "Journal prevents repeating mistakes" → Keep JOURNAL.md
• "I reference docs often" → Framework is discoverable

Framework friction:

• "BOS feels like busywork" → Simplify or make optional
• "Trust levels don't match reality" → Recalibrate
• "Journal isn't being used" → Not reading or not writing?
• "I never look at [doc]" → Doc may be redundant or buried

Intensity mismatch:

• "Too much overhead for this project" → Reduce intensity
• "We need more structure" → Increase intensity

Potential Follow-Ups

If framework friction:

• "Which specific step in BOS feels like busywork?"
• "Should we reduce framework intensity for this project?"
• "What if we cut [specific doc] - would that help or hurt?"

If trust levels off:

• "Should we recalibrate trust? Where do you think we are?"
• "What would Level [X+1] look like for this project?"

If journal not used:

• "Am I reading your journal entries and not applying lessons?"
• "Should we switch to simpler session memory?"

If intensity mismatch:

• "Would Medium intensity work better than Maximum?"
• "What components should we add/remove?"

Document

Q3: Framework Effectiveness
BOS/EOS: [Working? Tedious?]
Trust Protocol: [Appropriate? Adjust needed?]
Journal/Memory: [Continuity working?]
Documentation: [Too much? Too little? Just right?]
Intensity: [Maximum? Medium? Minimal? Mismatch?]
Actions: [Framework adjustments needed]

---

Q4: Patterns & Improvement Opportunities

Ask

"Q4: Have you noticed any patterns - good or bad - in how we work together?

Examples:

• Things I consistently do well or poorly
• Friction points that keep recurring
• Workflows that feel smooth or clunky
• Documentation that's always helpful or always ignored"

Listen For

Positive patterns (reinforce):

• "You always [good thing]" → Document as standard practice
• "When you [approach], it works great" → Capture in framework docs
• "The [specific doc] is really helpful" → Highlight in BOS

Negative patterns (fix):

• "You keep [bad thing] even though I mentioned it" → Not learning from feedback
• "Every session we struggle with [X]" → Systemic issue needs fix
• "I always have to remind you about [Y]" → Add to CONSTITUTION or BOS

Opportunities:

• "It would be great if you could [X]" → New capability or pattern to add
• "Other projects do [Y], could we?" → Learning from other repos
• "I wish the framework had [Z]" → Potential upstream contribution

Potential Follow-Ups

If positive patterns:

• "Should we formalize [pattern] in WORKFLOW.md or CONSTITUTION.md?"
• "What makes [pattern] work so well?"

If negative patterns:

• "Why do you think [bad pattern] keeps happening?"
• "What would break if we changed [pattern] to [alternative]?"
• "Should this be in CONSTITUTION.md as a non-negotiable?"

If opportunities:

• "Is [opportunity] critical or nice-to-have?"
• "Should we implement this, or document it for later?"
• "Is this project-specific or something the framework should support?"

Document

Q4: Patterns & Improvement Opportunities
Positive patterns: [What to reinforce]
Negative patterns: [What to fix]
Opportunities: [What to explore]
Actions: [Immediate fixes, framework updates, experiments]

---

Health Check Conclusion

Synthesize

After all 4 questions, provide a brief synthesis:

"Health Check Summary:

Strengths: [2-3 things working well] Improvements: [2-3 concrete actions to take this session or next] Experiments: [1-2 things to try and assess later]

Overall assessment: [One sentence on collaboration health]"

Propose Immediate Actions

Examples:

• "Let's update SHORTCUTS.md right now to simplify signals"
• "I'll add [concern] to CONSTITUTION.md as a non-negotiable"
• "Next session, I'll try [approach] and we'll assess if it helps"
• "Let's reduce framework intensity from Maximum to Medium"

Ask:

"Which of these should I tackle immediately, and which should we schedule for later?"

---

Document in HEALTH_CHECK_LOG.md

Format:

## Health Check: [Date] (Session [N])

**Overall Score:** [X/5]

### Q1: Collaboration Quality ([X/5])
- Strengths: [...]
- Concerns: [...]

### Q2: Communication Patterns
- Clarity: [...]
- Pacing: [...]
- Shortcuts: [working/needs revision]
- Tone: [...]

### Q3: Framework Effectiveness
- BOS/EOS: [...]
- Trust Protocol: [...]
- Journal/Memory: [...]
- Documentation: [...]
- Intensity: [appropriate/needs adjustment]

### Q4: Patterns & Opportunities
- Positive: [...]
- Negative: [...]
- Opportunities: [...]

### Actions Taken
1. [Immediate action 1]
2. [Immediate action 2]

### Actions Scheduled
1. [Future action 1]
2. [Future action 2]

### Experiments
1. [Experiment to try and assess later]

**Next health check:** Session [N + cadence] (in [X] sessions)

---

Update config.json

After health check complete:

{
  "sessions": {
    "count": [current_session_number],
    "lastHealthCheck": [current_session_number],
    "healthCheckCadence": [14 or 21 or 30 depending on intensity]
  }
}

This resets the health check timer.

---

Post-Health Check

Immediate follow-through:

• If you promised to update docs (SHORTCUTS.md, CONSTITUTION.md, etc.), do it NOW
• If you need to adjust trust level, document in TRUST_PROTOCOL.md
• If framework intensity should change, note in README.md and config.json

Session continuity:

• Add health check insights to today's journal entry
• Reference health check in end-of-session summary
• Commit HEALTH_CHECK_LOG.md update

Next session:

• Apply lessons learned from health check
• Monitor if proposed experiments are working
• Be ready to course-correct if approach isn't helping

---

Common Pitfalls

❌ Don't:

• Batch all questions at once (violates pacing rules)
• Treat this as a formality (it's a real collaboration improvement tool)
• Ignore negative feedback (even if painful to hear)
• Promise improvements you won't deliver (trust erosion)
• Rush through to "check the box" (defeats the purpose)

✅ Do:

• Ask one question at a time and wait
• Listen actively and dig into concerns
• Take concrete actions immediately
• Follow through on commitments
• Use insights to genuinely improve collaboration

---

Success Criteria

Health check is successful when:

1. User feels heard and understood
2. Concrete improvements identified and prioritized
3. Immediate actions taken (not just scheduled)
4. HEALTH_CHECK_LOG.md updated with clear documentation
5. Both parties have clearer picture of collaboration quality
6. Next health check timer reset in config.json

Health check failed if:

• User feels like it was checkbox compliance
• No actionable improvements identified
• Promises made but not kept
• Documentation incomplete or rushed
• Collaboration issues persist without addressing

---

Skill version: 1.0.0 Framework version: 0.5.0+ Last updated: 2025-10-31