block rp-why: Three Dimensions of AI Collaboration
Overview
The rp-why skill is a self-reflection framework that helps AI practitioners measure and improve their collaboration practice. It tracks three dimensions:
- DOK (Depth of Knowledge) - Cognitive complexity of human prompts, scored 1.0-4.0 using Webb's framework. "Adjusted" accounts for compression of intent.
- TM (Tool Maturity) - Intentional orchestration of AI tools, tracked through the Orchestra Model (Tiers 1-6).
- ADT (Agentic Delegation Trust) - The relationship between tool sophistication and cognitive depth. Derived from the TM x DOK matrix.
The intersection of these dimensions produces diagnostic zones that reveal the health of the collaboration practice.
How to Use This Skill
Installation
npx skills add https://github.com/block/agent-skills --skill rp-why
Available Commands
| Command | What It Does |
|---|---|
/rp-why init |
Generate a baseline from your session history |
/rp-why baseline |
Same as init |
/rp-why current |
Analyze the current session |
/rp-why compare |
Compare current session to your baseline |
/rp-why overall |
Full longitudinal analysis across all sessions |
/rp-why token-spend |
Daily token spend breakdown across all sessions |
Alternative: Natural Language
You can also ask naturally:
You: Analyze my AI collaboration patterns
You: What's my DOK distribution for this session?
You: How does this session compare to my baseline?
You: Give me the full rp-why longitudinal report
When to Use
- First time: Run
/rp-why init(or/rp-why baseline) to establish your starting point - End of session: Run
/rp-why currentto reflect on your work - Weekly: Run
/rp-why compareto track progress against baseline - Monthly: Run
/rp-why overallfor the full growth picture
The Three Dimensions
Dimension 1: DOK (Depth of Knowledge)
Measures the cognitive complexity of human prompts. Scored 1.0-4.0 using Webb's Depth of Knowledge framework.
| Level | Name | Description | Prompt Indicators |
|---|---|---|---|
| 1 | Recall & Reproduction | Simple factual prompts | "What is X?" "Show me the syntax for Y" |
| 2 | Application of Skills & Concepts | Applying learned skills to solve a problem | "Build this component" "Fix this error using pattern X" |
| 3 | Strategic Thinking | Reasoning across multiple concepts to plan, analyze, or design | "Design a system..." "Analyze trade-offs..." "What if..." |
| 4 | Extended Thinking | Creating something entirely new - frameworks, cross-disciplinary synthesis | "Research and synthesize..." "Create a framework..." |
Adjusted DOK accounts for compression of intent - when short prompts carry complex meaning due to established context. A compressed "proceed" that triggers a multi-step architectural deployment is not DOK 1.
DOK 3+4 % is the primary growth signal. It measures what proportion of your work operates at strategic or extended thinking levels.
Compression % tracks how often short directives carry complex intent. Compression emerges as trust deepens between practitioner and agent.
Dimension 2: TM (Tool Maturity) - Orchestra Model
Measures intentional orchestration - how deliberately you coordinate AI tools, agents, and workflows. Only counts actions the user deliberately initiated.
| Tier | Name | Description |
|---|---|---|
| 1 | Solo | Human works alone. AI reviews after. |
| 2 | Duet | Back-and-forth conversation. Human prompts, AI responds, human edits. |
| 3 | Ensemble | Human provides meaningful body of work. Evaluates holistically. |
| 4 | Chamber | Human delegates work streams. Sub-agents introduced. Orchestration required. |
| 5 | Symphony | Multiple AI interactions coordinated toward unified goal. Minimal intervention. |
| 6 | Virtuoso | Flow state. Human and AI synthesized. Optimal DOK, ADT, and TM. |
Dimension 3: ADT (Agentic Delegation Trust) - Diagnostic Zones
Measures the gap between Tool Complexity and Human Cognitive Depth. The TM x DOK matrix produces six zones:
| Zone | Description | Signal |
|---|---|---|
| Frontier | TM and DOK matched and growing together | Operating at the productive edge |
| Growing | Approaching a match between tool sophistication and cognitive depth | Building toward effective use |
| Expected | Tool usage and cognitive depth appropriate for current level | Healthy starting position |
| Thinking Ahead | Cognitive depth exceeds tool sophistication | Opportunity to adopt more powerful orchestration |
| Underutilizing | Tool sophistication exceeds cognitive depth | Opportunity to deepen the questions being asked |
| Overpowered | Significant mismatch between tool complexity and task depth | Resources spent without proportional cognitive return |
Diagnostic Zone Matrix
DOK 1 DOK 2 DOK 3 DOK 4
(Recall) (Application) (Strategic) (Extended)
+-----------+--------------+-------------+-------------+
Tier 5-6 | Overpowered| Underutil- | Frontier | Frontier |
(Symphony/ | | izing | | |
Virtuoso) | | | | |
+-----------+--------------+-------------+-------------+
Tier 3-4 | Overpowered| Expected | Growing | Frontier |
(Ensemble/ | | | | |
Chamber) | | | | |
+-----------+--------------+-------------+-------------+
Tier 1-2 | Expected | Growing | Thinking | Thinking |
(Solo/ | | | Ahead | Ahead |
Duet) | | | | |
+-----------+--------------+-------------+-------------+
Report Formats
/rp-why init or /rp-why baseline
Establishes the starting point. Analyzes all available session history.
Report structure:
+==================================================================+
| rp-why . BASELINE |
+==================================================================+
DATA SUMMARY
--------------------------------------------------------------------
Period: [start] - [end] ([N] days)
Sessions: [N]
Prompts: [N] classified
THREE DIMENSIONS
--------------------------------------------------------------------
DOK (Depth of Knowledge) [score]
TM (Tool Maturity) Tier [N] . [Name]
ADT (Delegation Trust) [Zone]
DIAGNOSTIC ZONE: [Zone Name]
--------------------------------------------------------------------
[Zone description from the six-zone model]
DOK DISTRIBUTION
--------------------------------------------------------------------
DOK 1 (Recall): [bar] [%]
DOK 2 (Application): [bar] [%]
DOK 3 (Strategic): [bar] [%]
DOK 4 (Extended): [bar] [%]
DOK 3+4: [%]
Compression: [%]
GROWTH TARGETS
--------------------------------------------------------------------
DOK 3+4 target: [%]
Compression target: [%]
Next TM tier: Tier [N] . [Name] ([what it means])
--------------------------------------------------------------------
Baseline saved to: ~/.config/goose/rp-why-baseline.json
Run /rp-why current after sessions to track progress.
/rp-why current
Analyzes the active session. The quick-check mirror.
Report structure:
+==================================================================+
| rp-why . CURRENT SESSION |
+==================================================================+
SESSION SNAPSHOT
--------------------------------------------------------------------
Date: [date]
Conversations: [N]
Prompts: [N] classified
THREE DIMENSIONS
--------------------------------------------------------------------
DOK (Adjusted) [score]
TM (Tool Maturity) Tier [N] . [Name]
ADT (Delegation) [Zone]
DIAGNOSTIC ZONE: [Zone Name]
--------------------------------------------------------------------
[Zone description]
DOK DISTRIBUTION
--------------------------------------------------------------------
DOK 1 (Recall): [bar] [%]
DOK 2 (Application): [bar] [%]
DOK 3 (Strategic): [bar] [%]
DOK 4 (Extended): [bar] [%]
DOK 3+4: [%] Compression: [%]
PEAK MOMENT
--------------------------------------------------------------------
"[highest DOK prompt text, truncated]" - DOK [N] . [Level Name]
GROWTH NUDGE
--------------------------------------------------------------------
[Contextual nudge based on diagnostic zone, not generic]
> [Reflection question tailored to current zone]
/rp-why compare
Delta report. Shows movement from baseline to now.
Report structure:
+==================================================================+
| rp-why . COMPARE |
+==================================================================+
COMPARING: Today ([date]) vs. Baseline ([baseline period])
THREE DIMENSIONS Baseline Today Delta
--------------------------------------------------------------------
DOK (Adjusted) [score] [score] [+/-%]
TM (Tool Maturity) Tier [N] Tier [N] [+/-N]
ADT (Delegation Trust) [Zone] [Zone] [arrow]
DIAGNOSTIC ZONE: [Baseline Zone] -> [Current Zone]
--------------------------------------------------------------------
DOK DISTRIBUTION Baseline Today Delta
--------------------------------------------------------------------
DOK 1 (Recall) [%] [%] [+/-pp]
DOK 2 (Application) [%] [%] [+/-pp]
DOK 3 (Strategic) [%] [%] [+/-pp]
DOK 4 (Extended) [%] [%] [+/-pp]
DOK 3+4 [%] [%] [+/-pp]
Compression [%] [%] [+/-pp or "emerged"]
TRAJECTORY
--------------------------------------------------------------------
Direction: [arrow] [Improving/Stable/Declining]
Signal: [1-2 sentence interpretation of what the delta means]
WHAT SHIFTED
--------------------------------------------------------------------
* [Bullet interpretations of the most meaningful changes]
* [Focus on what the numbers mean for the collaboration practice]
* [Connect to diagnostic zone movement]
--------------------------------------------------------------------
Run /rp-why overall for full longitudinal analysis.
/rp-why overall
Full longitudinal report. The complete growth picture.
Report structure:
+==================================================================+
| rp-why . OVERALL |
+==================================================================+
FULL DATASET: [start] - [end] ([N] days)
Sessions: [N] | Prompts: [N] | Conversations: [N]
CURRENT STANDINGS
--------------------------------------------------------------------
DOK (Adjusted, full mean) [score]
TM (Tool Maturity) Tier [N] . [Name]
ADT (Delegation Trust) [Zone]
DOK 3+4: [%] Compression: [%] Floor: [score]
ROLLING AVERAGE (Last 10 Sessions vs. First 10)
--------------------------------------------------------------------
First 10 Last 10 Delta
Adjusted DOK [score] [score] [+/-%]
DOK 3+4 % [%] [%] [+/-pp]
Compression % [%] [%] [emerged/+/-pp]
PHASE ANALYSIS
--------------------------------------------------------------------
Phase Dates Sessions DOK DOK3+4 Comp TM
----- ----- -------- --- ------ ---- --
[Auto-detected phases based on DOK trajectory shifts]
TRAJECTORY
--------------------------------------------------------------------
Peak DOK: [score] ([date] - [context])
Floor: [score] ([interpretation])
Direction: [arrow] [Narrative of the growth arc]
DOK DISTRIBUTION (Full Dataset)
--------------------------------------------------------------------
DOK 1 (Recall): [bar] [%]
DOK 2 (Application): [bar] [%]
DOK 3 (Strategic): [bar] [%]
DOK 4 (Extended): [bar] [%]
GROWTH STORY
--------------------------------------------------------------------
[2-3 sentence narrative interpreting the full arc. Not prescriptive.
Describes what the data reveals about how the collaboration practice
has evolved.]
--------------------------------------------------------------------
Data source: ~/.local/share/goose/sessions/sessions.db
Methodology: DOK keyword classification, compression detection
(short prompt + high response ratio + established context),
adjusted DOK (+1 level for compressed prompts, capped at 4).
Backward Compatibility
Reading Previous Baselines
If an existing baseline file uses the v3 format (Gas Town stages, quadrant terminology), the skill translates automatically:
| v3 Field | v4 Equivalent |
|---|---|
estimated_stage: 5 |
tm_tier: 3 (Ensemble) - Stage 5 maps to Tier 3 for new users |
quadrant: "Underutilizing" |
adt_zone: "Underutilizing" (zone name preserved) |
quadrant: "Frontier" |
adt_zone: "Frontier" |
quadrant: "Thinking Ahead" |
adt_zone: "Thinking Ahead" |
quadrant: "Learning Zone" |
adt_zone: "Expected" |
average_dok_score |
dok_adjusted (treated as raw if no compression data) |
growth_targets.dok_target |
growth_targets.dok_3_4_pct |
Gas Town to Orchestra Mapping
For users familiar with the Gas Town stages:
| Gas Town Stage | Orchestra Tier | Name |
|---|---|---|
| 1-2 (Observer/Curious) | Tier 1 | Solo |
| 3 (Copilot) | Tier 2 | Duet |
| 4 (Chat IDE) | Tier 3 | Ensemble |
| 5 (CLI Agent) | Tier 4 | Chamber |
| 6 (Multi-Agent) | Tier 5 | Symphony |
| 7-8 (Agentic/Full) | Tier 6 | Virtuoso |
Metrics Reference
| Metric | What It Measures | Range | Growth Signal |
|---|---|---|---|
| DOK (Adjusted) | Cognitive complexity of prompts, accounting for compression | 1.0 - 4.0 | Rising mean |
| DOK 3+4 % | Proportion of strategic/extended thinking | 0 - 100% | Increasing |
| Compression % | Short directives carrying complex intent | 0 - 100% | Emerging, then growing |
| Floor | Lowest DOK on routine days | 1.0 - 4.0 | Rising (baseline becomes unreproducible) |
| TM Tier | Orchestra Model tier | 1 - 6 | Advancing |
| ADT Zone | Diagnostic zone from TM x DOK | 6 zones | Moving toward Frontier |
Growth Nudges by Zone
Frontier
- "Operating at the productive edge. Document what works for others."
- "The collaboration is matched. Look for opportunities to extend into new domains."
Growing
- "Approaching a match. Keep pushing DOK 3+ work and the zone will shift."
- "Consider: what's one workflow you could delegate more fully?"
Expected
- "Healthy starting position. Growth comes from asking 'why' before implementing."
- "Try framing one task as a design decision rather than an execution request."
Thinking Ahead
- "Cognitive depth exceeds tool sophistication. Time to adopt more powerful orchestration."
- "Your thinking is ready for Tier [N+1]. Explore sub-agents or multi-step delegation."
Underutilizing
- "Powerful tools deserve powerful questions. Before each prompt: can this be more strategic?"
- "Batch simple queries. Reserve the agent for work that requires reasoning."
Overpowered
- "Significant mismatch. Consider whether this task needs an autonomous agent."
- "Opportunity: redirect this tool toward a problem that requires analysis or design."
Attribution
- Depth of Knowledge (DOK): Norman Webb (1997). Webb, N. L. Criteria for alignment of expectations and assessments in mathematics and science education.
- Orchestra Model (TM): Dakota Fabro (2026). Measuring Cognitive Complexity in Human-AI Collaboration.
- Three Dimensions Framework: Dakota Fabro (2026). rp-why longitudinal dataset, Block Builder Fellowship.
- Gas Town Stages (v1-v3 foundation): Steve Yegge, "Welcome to Gas Town" (January 2026).
Version History
| Version | Date | Changes |
|---|---|---|
| 4.0 | 2026-06 | Three Dimensions model (DOK + TM + ADT), Orchestra Tiers, diagnostic zones, compression tracking, /rp-why baseline alias, /rp-why overall report, phase analysis, backward compatibility with v3 baselines |
| 3.0 | 2026-02 | Quadrant visualization, growth nudges, reflection prompts |
| 2.x | 2026-01 | Integration matrix, target profiles, baseline comparison |
| 1.x | 2025-12 | Initial Gas Town stages, basic DOK tracking |