feat: add L10 Proteus dynamic adaptive reasoning as default

README.md

@@ -25,7 +25,7 @@ tags:
 
 [![Solstice AI](https://img.shields.io/badge/Solstice--AI-Protium--1.0-blueviolet?style=for-the-badge&logo=openai)](https://huggingface.co/Solstice-AI)
 [![Model-Agnostic](https://img.shields.io/badge/Architecture-Model--Agnostic-success?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
-[![Levels](https://img.shields.io/badge/Cognitive%20Spectrum-L0%20to%20L9%20Solo-orange?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
+[![Levels](https://img.shields.io/badge/Cognitive%20Spectrum-L0%20to%20L10%20Dynamic-orange?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
 [![Swarm](https://img.shields.io/badge/Swarm-20--Agent%20%2B%20Council-red?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
 
 **Zero Neutrons. Zero Dead Weight. Pure Thrust.**
@@ -47,27 +47,29 @@ Unlike legacy reasoning templates that impose arbitrary token floors (e.g. David
 3. **Objective-Gated Early Exit**: The moment the target condition is satisfied and verified, `<think>` terminates immediately. Zero stall, zero padding, zero manufactured doubt.
 4. **Anti-MiG-25 / Delta-Only Correction**: If a contradiction or constraint violation occurs, the model isolates and repairs only the specific broken delta. It never spirals into circular loops or complete restarts.
 5. **Absolute Operational Aura**: Clean, decisive, unhedged output delivery with maximum information density and zero conversational preamble.
+6. **Zero Token Quotas or Word Floors**: Reasoning depth is guided 100% by problem complexity and verification gates, never an artificial token counter.
 
 ---
 
 ## 🎚️ The Cognitive Spectrum
 
-Protium-1.0 provides continuous elastic reasoning calibrated as typical output shares (never rigid token quotas):
+Protium-1.0 defaults to **`[L10]` Proteus (Dynamic Reasoning)**, which automatically shifts cognitive depth based on intrinsic prompt difficulty:
 
-| Level | Identifier | Archetype | Target Share | Ideal Workload |
+| Level | Identifier | Archetype | Operational Focus | Ideal Workload |
 |---|---|---|---|---|
-| **[L0]** | `mortal` / `regular` | Instant Execution | 0% (Thinking OFF) | Pure conversational chat, classification, summarization |
-| **[L1]** | `hermes` | Rapid Instinct | < 10% | Instant sanity check, syntax validation, arithmetic verify |
-| **[L2]** | `apollo` | Disciplined Deduction | 10% – 25% | Linear deduction, rule checking, constraint mapping |
-| **[L3]** | `artemis` | Boundary & Failure Audit | 20% – 40% | Edge cases, null checks, concurrency hazard scanning |
-| **[L4]** | `athena` | Strategic Architecture | 30% – 50% | Multi-step refactors, API design, trade-off optimization |
-| **[L5]** | `prometheus` | Defensive Engineering | 40% – 60% | Fault tolerance, scaling bottlenecks, lifecycle resilience |
-| **[L6]** | `solstice` | Deep Systemic Derivation | 50% – 70% | Dual-hypothesis synthesis, axiomatic proof, stress-testing |
-| **[L7]** | `hyperion` | First-Principles *(Default)* | 60% – 85% | Comprehensive first-principles decomposition |
-| **[L8]** | `einstein` | Theoretical Mastery | Majority | Theoretical unification, invariant reduction, deep algorithms |
-| **[L9]** | `oracle` | Proof-Grade Authority | Unbounded | Exhaustive formal proofs, zero-placeholder implementations |
-| **Swarm** | `swarm` | 20-Perspective Swarm | Elastic | Breakthrough ideation, divergent multi-perspective synthesis |
-| **Council**| `council` | Adversarial Research Panel | Elastic | Multi-specialist consensus, formal verification, audit |
+| **[L10]** | `proteus` / `dynamic` | **Dynamic Adaptive (Default)** | **Fluid impedance matching** | **Auto-scales across all queries (near-zero on simple, expands on hard)** |
+| **[L0]** | `mortal` / `regular` | Instant Execution | Thinking OFF | Conversational chat, pure classification, basic summarization |
+| **[L1]** | `hermes` | Rapid Instinct | High-velocity check | Instant sanity checks, syntax validation, arithmetic verification |
+| **[L2]** | `apollo` | Disciplined Deduction | Linear implication | Constraint mapping, structured rule following, direct logic |
+| **[L3]** | `artemis` | Boundary & Failure Audit | Edge vulnerability probe | Null checks, type boundaries, concurrency hazard audits |
+| **[L4]** | `athena` | Strategic Architecture | Modular synthesis | Multi-step refactors, API design, trade-off optimization |
+| **[L5]** | `prometheus` | Defensive Engineering | Second-order modeling | Fault tolerance, scaling bottlenecks, lifecycle resilience |
+| **[L6]** | `solstice` | Deep Systemic Derivation | Axiomatic derivation | Dual-hypothesis synthesis, axiomatic proof, stress-testing |
+| **[L7]** | `hyperion` | First-Principles Synthesis | Ground-up decomposition | Comprehensive first-principles decomposition |
+| **[L8]** | `einstein` | Theoretical Mastery | Unified invariant reduction | Theoretical breakthroughs, deep algorithms, continuous derivations |
+| **[L9]** | `oracle` | Proof-Grade Authority | Formal proof verification | Exhaustive formal proofs, zero-placeholder implementations |
+| **Swarm** | `swarm` | 20-Perspective Swarm | Divergent synthesis | Breakthrough ideation, multi-perspective tension resolution |
+| **Council**| `council` | Adversarial Research Panel | Institutional consensus | Multi-specialist consensus, formal verification, audit |
 
 ---
 
@@ -76,6 +78,10 @@ Protium-1.0 provides continuous elastic reasoning calibrated as typical output s
 ### 1. llama.cpp / LM Studio
 Pass any family template via the `--chat-template-file` flag:
 ```bash
+# Default Proteus dynamic reasoning:
+llama-cli -m model.gguf --chat-template-file templates/qwen.jinja -p "Refactor this async worker"
+
+# Explicit level steering:
 llama-cli -m model.gguf --chat-template-file templates/qwen.jinja -p "[L4] Refactor this async worker"
 ```
 
@@ -95,10 +101,17 @@ tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-32B-Instruct")
 with open("templates/qwen.jinja") as f:
     tokenizer.chat_template = f.read()
 
-# Invoke with runtime kwargs:
+# Dynamic default:
 prompt = tokenizer.apply_chat_template(
     [{"role": "user", "content": "Prove that every planar graph is 5-colorable."}],
-    reasoning_effort="oracle",
+    tokenize=False,
+    add_generation_prompt=True
+)
+
+# Explicit level:
+prompt_athena = tokenizer.apply_chat_template(
+    [{"role": "user", "content": "Refactor this code"}],
+    reasoning_effort="athena",
     tokenize=False,
     add_generation_prompt=True
 )
@@ -108,16 +121,15 @@ prompt = tokenizer.apply_chat_template(
 
 ## 🎯 Steering Controls
 
-Protium-1.0 supports instant mid-session steering via inline tags or runtime kwargs:
-
-- **Inline Steering**: Place tags anywhere in your prompt:
-  - `[L2]` or `{REASON:apollo}`: Sets Apollo level.
-  - `[L8]` or `{REASON:einstein}`: Sets Einstein theoretical level.
-  - `[regular]` or `[L0]`: Turns off thinking for pure vanilla response.
-  - `{REASON:i8}`: Silent execution (applies L8 cognitive rigor internally without emitting `<think>`).
+- **Inline Steering**:
+  - `[L10]` or `{REASON:proteus}`: Dynamic unconstrained adaptive reasoning (Default).
+  - `[L4]` or `{REASON:athena}`: Locks to Athena modular balance.
+  - `[L8]` or `{REASON:einstein}`: Locks to Einstein theoretical level.
+  - `[regular]` or `[vanilla]`: Pure clean zero-overhead template without think tags.
+  - `{REASON:i8}`: Silent execution (runs L8 cognitive rigor internally without emitting `<think>`).
 - **Runtime Kwargs**:
-  - `reasoning_effort="hermes"|"apollo"|"artemis"|"athena"|"prometheus"|"solstice"|"hyperion"|"einstein"|"oracle"|"swarm"|"council"`
-  - `thinking_level="10"` through `"100"` (continuous spectrum)
+  - `reasoning_effort="10"|"proteus"|"dynamic"|"auto"`
+  - `reasoning_effort="1".."9"` or named levels (`hermes`, `apollo`, `artemis`, `athena`, `prometheus`, `solstice`, `hyperion`, `einstein`, `oracle`, `swarm`, `council`)
   - `regular=True` or `mode="vanilla"` (zero overhead)
 
 ---