1  # Tribal Build Protocol
  2  
  3  *Spawn agents to build while maintaining meta-level focus*
  4  
  5  ---
  6  
  7  - **principle**
  8    - "Philosophy discovers. Tribes build. Meta-level maintains attention."
  9    - Build liberally. Rewrites are cheap. Tokens are cheap.
 10  
 11  - **shape**
 12    - Discovery loop: Philosophy → Insight → Build candidate
 13    - Spawn loop: Build candidate → Agent → Background execution
 14    - Tracking loop: Agent tracker → Status → Integration
 15    - Attention remains at meta-level while builds execute
 16  
 17  ---
 18  
 19  ## The Pattern
 20  
 21  ### 1. Discovery Phase (Meta-level)
 22  - Engage in philosophical exploration
 23  - Extract actionable insights
 24  - Identify build candidates
 25  
 26  **Triggers for build candidate:**
 27  - "This concept could be code"
 28  - "We've been circling - time to make it real"
 29  - "Philosophy → Implementation"
 30  
 31  ### 2. Spawn Phase
 32  When a build candidate is identified:
 33  
 34  ```
 35  Use Task tool with:
 36  - subagent_type: "general-purpose" for complex builds
 37  - subagent_type: "Bash" for simple scripts
 38  - run_in_background: true (ALWAYS for builds)
 39  - model: "haiku" for straightforward tasks (cheap)
 40  - model: "sonnet" for moderate complexity
 41  - model: "opus" for judgment-heavy tasks
 42  ```
 43  
 44  **Spawn prompt template:**
 45  ```
 46  Build [THING] for Sovereign_OS.
 47  
 48  Location: /Users/rcerf/repos/Sovereign_OS/[path]
 49  
 50  Requirements:
 51  1. [Core requirement]
 52  2. [Core requirement]
 53  3. Test against axioms (A0-A4)
 54  
 55  RETURN FORMAT (required):
 56  
 57  ## What Was Built
 58  - File: [path]
 59  - Purpose: [description]
 60  - Links to: [[concept1]], [[concept2]]
 61  
 62  ## Test Record
 63  | Test | Expected | Actual | Status |
 64  |------|----------|--------|--------|
 65  | [test1] | [expected] | [actual] | PASS/FAIL |
 66  
 67  ## Axiom Alignment (graph-walk)
 68  | Axiom | Alignment | Link Created |
 69  |-------|-----------|--------------|
 70  | [[A0 Boundary Operation]] | [how it aligns] | [[target-concept]] |
 71  | [[A1 Telos of Integration]] | [how it aligns] | [[target-concept]] |
 72  | [[A2 Recognition of Life]] | [how it aligns] | [[target-concept]] |
 73  | [[A3 Dynamic Pole Navigation]] | [how it aligns] | [[target-concept]] |
 74  | [[A4 Ergodic Asymmetry]] | [how it aligns] | [[target-concept]] |
 75  
 76  ## Failures & Solutions
 77  - [failure]: [how solved]
 78  
 79  ## Final State
 80  - Status: WORKING / PARTIAL / BLOCKED
 81  - Evidence: [command to verify]
 82  - Graph edges created: [count]
 83  ```
 84  
 85  > **Meta-principle:** Every test against axioms is a graph walk. Testing creates connections. The alignment check IS the linking process.
 86  
 87  ### 3. Tracking Phase
 88  When spawning multiple agents, spawn a tracker:
 89  
 90  ```
 91  Use Task tool with:
 92  - description: "Agent tracker - meta first officer"
 93  - subagent_type: "Bash"
 94  - model: "haiku"
 95  - run_in_background: true
 96  ```
 97  
 98  The tracker creates `/sessions/AGENT-TRACKER.md` with:
 99  - Active agents
100  - Their status
101  - Their purpose
102  - Results when complete
103  
104  ### 4. Integration Phase
105  When agents complete:
106  1. Review their output
107  2. Test the builds
108  3. Document in backlog if gaps found
109  4. Radiate implementation back through graph (link philosophy to code)
110  
111  ---
112  
113  ## Testing as Graph-Walking
114  
115  > "Every test against our principles is us live walking our graph, building a connection to this new related thought, to this new protocol, to this new solution built."
116  
117  - **principle**
118    - Testing is not verification alone—it's graph construction
119    - Each axiom check creates an edge from build → axiom → derived principle
120    - The test record IS the link manifest
121  
122  - **shape**
123    - Test A0 (Boundary) → Creates link to [[boundary-operation]], [[markov-blanket]]
124    - Test A1 (Integration) → Creates link to [[telos-of-integration]], [[binding]]
125    - Test A2 (Life) → Creates link to [[recognition-of-life]], [[primitive-vs-calcified]]
126    - Test A3 (Navigation) → Creates link to [[dynamic-pole-navigation]], [[tension-dyad]]
127    - Test A4 (Ergodicity) → Creates link to [[ergodic-asymmetry]], [[building-as-cognition]]
128  
129  - **implementation**
130    - When agent tests against axiom, it MUST output [[wiki-link]] to related concepts
131    - These links become the connection fabric between philosophy and code
132    - The alignment report is also a link manifest
133  
134  ---
135  
136  ## When to Use This Protocol
137  
138  **Use when:**
139  - Discovery cycle is repeating without producing artifacts
140  - Insight is concrete enough to specify in code
141  - Meta-level attention should stay on synthesis, not implementation
142  - Multiple independent builds could run in parallel
143  
144  **Don't use when:**
145  - The build requires tight human feedback loops
146  - The implementation is trivial (< 50 lines)
147  - Discovery is still too vague to specify
148  
149  ---
150  
151  ## Agent Count Thresholds
152  
153  | Active Agents | Action |
154  |---------------|--------|
155  | 1-2 | No special tracking needed |
156  | 3-5 | Spawn tracker agent |
157  | 6+ | Consider batching or serializing |
158  
159  ---
160  
161  ## Model Selection Guide
162  
163  | Task Type | Model | Rationale |
164  |-----------|-------|-----------|
165  | Simple scripts | Haiku | Fast, cheap |
166  | Multi-file builds | Sonnet | Balance of speed/quality |
167  | Judgment-heavy | Opus | Needs reasoning depth |
168  | Agent tracker | Haiku | Simple coordination |
169  | Semantic analysis | Sonnet/Opus | Needs understanding |
170  
171  ---
172  
173  ## Example: This Session's Pattern
174  
175  1. **Discovery:** Resonance engine concept emerged from Hayek/Deleuze discussion
176  2. **Spawn:** Agent a45b875 built `resonance_engine.py` and fixed orphans
177  3. **Result:** 228 connections, 49% → 65% health
178  4. **Gap identified:** Need semantic resonance, not term overlap
179  5. **Spawn:** Agent a6bdd0b building semantic engine
180  6. **Tracker:** Agent abdf0c5 monitoring progress
181  7. **Integration:** Philosophy pages updated to reference implementations
182  
183  The meta-level (Opus) maintained synthesis focus while tribes (Sonnet/Haiku) built.
184  
185  ---
186  
187  ## Related
188  
189  - **axioms**
190    - [[A4 Ergodic Asymmetry]] - cheap rewrites enable liberal building
191      - shape:: "Prevent ruin before optimizing gain."
192    - [[A1 Telos of Integration]] - agents integrate back to graph
193      - shape:: "Satan didn't know he was choosing isolation."
194  - **protocols**
195    - [[first-officer-protocol]] - FO maintains meta attention
196    - [[mission-control-protocol]] - MC coordinates across threads
197  - **tools**
198    - [[resonance_engine.py]] - file connection tool
199    - [[insight_escalator.py]] - concept escalation tool
200    - [[axiom_resonance.py]] - axiom matching tool
201  
202  ---
203  
204  *Protocol extracted from session 2026-01-15 | Pattern: Philosophy → Tribes → Integration*
205  - [[concept-topology-research]] - resonance: 27%
206  - [[sprint-2026-01-15-assessment]] - resonance: 80%
207