Skip to main content
Back to Exercises
Partnership intermediate 30 min

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,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

Claude 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