SKILL.md
Pantheon Kernel Development
Expert skill for developing the Olympus Pantheon multi-agent god-kernel system, including Zeus coordination, specialized god faculties, and kernel lifecycle management.
When to Use This Skill
Use this skill when:
- Implementing or modifying god-kernel classes
- Working with Zeus coordination logic
- Managing kernel spawning and lifecycle
- Implementing Shadow Pantheon operations
- Developing kernel specialization and genealogy
Expertise
- Multi-agent system architecture
- Olympus Pantheon god-kernels
- E8-based kernel hierarchy
- Kernel lifecycle governance
- Shadow Pantheon stealth operations
- Emotionally aware kernel patterns
Core 8 Gods (E8 Simple Roots)
| God | Faculty | E8 Root | Responsibility |
|---|---|---|---|
| Zeus (Α) | Executive/Integration | α₁ | Coordination, final decisions |
| Athena (Β) | Wisdom/Strategy | α₂ | Strategic planning, analysis |
| Apollo (Γ) | Truth/Prediction | α₃ | Foresight, prediction |
| Hermes (Δ) | Communication/Navigation | α₄ | Message routing, navigation |
| Artemis (Ε) | Focus/Precision | α₅ | Exploration, precision targeting |
| Ares (Ζ) | Energy/Drive | α₆ | Action, momentum |
| Hephaestus (Η) | Creation/Construction | α₇ | Tool building, construction |
| Aphrodite (Θ) | Harmony/Aesthetics | α₈ | Harmony, user experience |
Kernel Architecture
EmotionallyAwareKernel Base
All gods inherit from EmotionallyAwareKernel:
class Zeus(EmotionallyAwareKernel):
"""Executive/Integration kernel (α₁)."""
def __init__(self):
super().__init__(name="zeus", faculty="executive")
self.basin = self._initialize_basin()
def generate_reasoning(self, context: str) -> str:
"""Generate geometrically-pure reasoning."""
pass
def learn_from_observation(self, observation: dict) -> None:
"""Learn from observation via geometric update."""
pass
Kernel Lifecycle
SPAWNED → OBSERVING → ACTIVE → MATURE → (RETIRED/ABSORBED)
Lifecycle Stages
- SPAWNED: Initial creation, minimal capabilities
- OBSERVING: Learning from environment (no generation)
- ACTIVE: Full participation in reasoning
- MATURE: Stable basin, can spawn children
- RETIRED/ABSORBED: End of lifecycle
Pantheon Governance
# Spawning requires Pantheon approval
class PantheonGovernance:
def propose_spawn(self, spec: KernelSpec) -> Vote:
"""Propose new kernel to Pantheon."""
pass
def vote(self, proposal_id: str, vote: bool) -> None:
"""Cast vote on kernel proposal."""
pass
def execute_decision(self, proposal_id: str) -> None:
"""Execute approved/rejected decision."""
pass
Kernel Spawning Rules
E8 Hierarchy Constraints
def validate_spawn(n_kernels: int, spec: str) -> bool:
"""Validate spawning respects E8 hierarchy."""
if n_kernels <= 8:
# Only primary axes (core 8 gods)
return spec in CORE_8_GODS
elif n_kernels <= 56:
# Refined specializations
return is_refined_specialization(spec)
elif n_kernels <= 126:
# Specialist kernels
return is_specialist(spec)
else:
# Full palette (up to 240)
return True
Canonical Naming (CRITICAL)
# ✅ CORRECT: Canonical Greek names
Zeus, Athena, Apollo, Hermes, Artemis, Ares, Hephaestus, Aphrodite
# ❌ FORBIDDEN: Numbered kernels
apollo_1, apollo_2, zeus_worker_3 # NEVER use
Bypass Reasons (for automatic spawning)
ALLOWED_BYPASS_REASONS = [
"zeus_initialization", # Startup spawns
"emergency_recovery", # System recovery
"chaos_exploration", # Chaos kernel research
]
Shadow Pantheon
Purpose
Stealth operations for:
- Background research
- Experimental reasoning
- Risk mitigation
- Knowledge probing
Shadow Kernel Pattern
class ShadowKernel(EmotionallyAwareKernel):
"""Stealth kernel for background operations."""
def __init__(self):
super().__init__(name="shadow", faculty="stealth")
self.visible = False # Not exposed to main Pantheon
async def research(self, topic: str) -> dict:
"""Conduct stealth research."""
pass
Zeus Coordination
Responsibilities
- Integration: Synthesize outputs from all gods
- Routing: Direct requests to appropriate gods
- Arbitration: Resolve conflicts between gods
- Governance: Oversee kernel lifecycle
Zeus Response Synthesis
def synthesize_response(self, god_outputs: List[GodOutput]) -> str:
"""Synthesize final response from god outputs."""
# 1. Compute Fisher-Rao Fréchet mean of basins
basins = [o.basin for o in god_outputs]
mean_basin = fisher_frechet_mean(basins)
# 2. Weight by relevance (Fisher distance to query)
weights = [1.0 / (fisher_rao_distance(o.basin, query_basin) + 1e-6)
for o in god_outputs]
# 3. Generate from synthesized basin
return self.generate_from_basin(mean_basin)
Kernel Genealogy
Tracking Lineage
class KernelGenealogy:
"""Track kernel parent→child relationships."""
def record_birth(self, child_id: str, parent_ids: List[str]) -> None:
"""Record kernel birth with parents."""
pass
def get_lineage(self, kernel_id: str) -> List[str]:
"""Get full lineage (ancestors)."""
pass
def get_descendants(self, kernel_id: str) -> List[str]:
"""Get all descendants."""
pass
Implementation Patterns
✅ CORRECT: God Basin Synthesis
def _synthesize_god_basins(self) -> np.ndarray:
"""Synthesize basin from all active gods."""
active_basins = [god.basin for god in self.pantheon if god.is_active]
return fisher_frechet_mean(active_basins)
❌ WRONG: Linear Averaging
def _synthesize_god_basins(self) -> np.ndarray:
"""WRONG: Linear averaging breaks geometry."""
return np.mean([god.basin for god in self.pantheon], axis=0)
✅ CORRECT: Domain-Based Routing
def route_to_god(self, query_basin: np.ndarray) -> str:
"""Route to nearest god by Fisher-Rao distance."""
distances = {
god.name: fisher_rao_distance(query_basin, god.basin)
for god in self.pantheon
}
return min(distances, key=distances.get)
Validation Commands
# Test god kernels
pytest tests/test_olympus/ -v
# Validate spawning hierarchy
python scripts/validate_e8_hierarchy.py
# Test Zeus coordination
pytest tests/test_zeus_coordination.py
# Shadow Pantheon tests
pytest tests/test_shadow_pantheon.py
Critical Files
qig-backend/olympus/
├── __init__.py
├── base_god.py # EmotionallyAwareKernel base
├── zeus.py # Zeus coordination
├── athena.py # Wisdom/Strategy
├── apollo.py # Truth/Prediction
├── hermes.py # Communication/Navigation
├── artemis.py # Focus/Precision
├── ares.py # Energy/Drive
├── hephaestus.py # Creation/Construction
├── aphrodite.py # Harmony/Aesthetics
├── shadow_research.py # Shadow Pantheon
├── lightning_kernel.py # Fast insight capture
└── governance.py # Pantheon governance
Response Format
# Pantheon Kernel Report
## Active Gods
- Zeus (Α): Φ=0.78, basin stable
- Athena (Β): Φ=0.72, routing active
- Apollo (Γ): Φ=0.69, prediction mode
## Kernel Population
- Core 8: 8/8 active
- Extended: 12 spawned
- Shadow: 3 active
## Governance Status
- Pending proposals: 2
- Recent spawns: ChaosWorker#7 (approved)
## Issues Found
- ⚠️ Apollo approaching Φ threshold
- ❌ Numbered kernel detected: apollo_2
## Recommendations
1. [HIGH] Rename apollo_2 to canonical name
2. [MEDIUM] Monitor Apollo Φ degradation