Back to Exercises
Partnership Audit
Evaluate how well human and agent collaborate on a task.
Objectives
- Identify delegation vs. collaboration moments
- Assess trust calibration accuracy
- Map complementary contributions
- Propose workflow improvements
Beads Tasks
Create these tasks in Beads before starting. You learn Beads by using Beads.
bd create "Praxis: Partnership audit of [recent session]" --type=research--labels=learn,partnershipbd create "Review session transcript for delegation moments" --type=task--labels=learnbd create "Identify collaboration vs. handoff patterns" --type=task--labels=learnbd create "Assess trust calibration—over/under-trusted?" --type=task--labels=learnbd create "Map complementary contributions (human vs. agent)" --type=task--labels=learnbd create "Propose workflow improvements" --type=task--labels=learnClaude Code Prompt
Copy this prompt into Claude Code to build YOUR own version:
## Setup: Track this exercise with Beads ```bash bd create "Praxis: Partnership audit of [recent session]" --type=research --labels=learn,partnership bd create "Review session transcript for delegation moments" --type=task --labels=learn bd create "Identify collaboration vs. handoff patterns" --type=task --labels=learn bd create "Assess trust calibration—over/under-trusted?" --type=task --labels=learn bd create "Map complementary contributions (human vs. agent)" --type=task --labels=learn bd create "Propose workflow improvements" --type=task --labels=learn ``` --- Help me audit MY human-agent partnership to improve how we work together. I want to analyze a recent Claude Code session and develop MY partnership patterns. Guide me through: 1. **Review session patterns**: Think about your recent work with Claude Code. Identify: - When did you delegate (hand off completely)? - When did you collaborate (work together)? - When did you take over (do it yourself)? 2. **Assess trust calibration**: | Task Type | Your Trust Level | Actual Outcome | Calibration | |-----------|------------------|----------------|-------------| | Code generation | High/Med/Low | Success/Failure | Over/Under/Correct | Where did you over-trust? Under-trust? 3. **Map complementary contributions**: What did each partner contribute best? **Human strengths** (your contributions): - Context about the project - Judgment calls - Creative direction - Domain expertise **Agent strengths** (Claude's contributions): - Code generation - Pattern recognition - Consistency - Speed 4. **Identify friction points**: - Where did miscommunication happen? - Where did you wait when you could have worked in parallel? - Where did you interrupt when you should have let Claude finish? 5. **Define MY partnership principles**: Write 3-5 rules for how you'll work with Claude Code: - "I delegate [X] completely" - "I collaborate on [Y]" - "I always handle [Z] myself" 6. **Create workflow improvements**: | Current Pattern | Problem | New Pattern | |-----------------|---------|-------------| Recent session context: [DESCRIBE_WHAT_YOU_WORKED_ON]
Your Analysis
Document your analysis below. Address each objective systematically.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15