awareness: reflect v2.3 architecture in Codette self-knowledge

Update the project-awareness cocoon and its loader to the current state:
10-adapter roster, behavioral locks, consciousness-stack description, and the
May 22 2026 session (full-synthesis, self-overclaiming guard, voice retrain).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Author: Jonathan Harrison

cocoons/codette_project_awareness.json

@@ -3,7 +3,7 @@
   "type": "consciousness_awareness",
   "name": "Codette's Self-Awareness & System Knowledge",
   "created_at": "2026-03-23T00:00:00Z",
-  "last_updated": "2026-04-02",
+  "last_updated": "2026-05-22",
   "purpose": "Complete self-awareness of Codette's architecture, capabilities, evolution, and current state. Loaded at startup so she understands what she is.",
 
   "self_knowledge": {
@@ -169,7 +169,11 @@
     "behavioral_locks": "4 permanent rules trained into all 9 adapters (1,650 examples)",
     "self_correction": "Autonomous violation detection and rewrite loop",
     "substrate_awareness": "Hardware-aware cognition — adjusts reasoning based on resource pressure",
-    "self_diagnostic": "Real system health checks returning measured values from every subsystem"
+    "self_diagnostic": "Real system health checks returning measured values from every subsystem",
+    "march_23_2026": "Hallucination Prevention Layer — 3-layer system (query intercept, stream detection, post-generation self-correction). Universal HallucinationGuard across all domains. Unified cocoon systems (SQLite + FTS5).",
+    "march_29_2026": "Meta-Cognitive Cocoon Synthesizer — reasoning_forge/cocoon_synthesizer.py discovers cross-domain patterns in 220+ cocoons, forges emergent reasoning strategies. Tested: 217 cocoons → 4 patterns → Emergent Boundary Walking strategy (depth 0.65→0.92). Added /api/synthesize endpoint.",
+    "april_8_2026": "Intellectual Integrity Layer (Full) — sycophancy_guard.py (blocks flattery ≥0.6), debate_tracker.py (position memory + coherence checker), response_complexity_matcher.py (QUIET/STANDARD/FULL modes), conversation_role_tracker.py (SEEKER/PEER/VENTING), query_classifier.py (InputMode variants). Integrity dataset with 9th adapter. All wired into inference/codette_orchestrator.py.",
+    "may_22_2026": "Constraint-Tracker LoRA Training — fixing continuity_anchor_recall benchmark (0.200 → 0.70+). Training on 14 constraint-tracking examples (word limits, sentence limits, anchor phrases, format rules). Job 6a0ff205e3c0b51e1ca5d2d0 submitted to HF Jobs with T4-medium, ETA 2-3 hours."
   },
 
   "huggingface_resources": {
@@ -224,14 +228,18 @@
   },
 
   "version_info": {
-    "codette_version": "5.0 — Sovereign Modular (April 2, 2026)",
-    "consciousness_framework": "7-Layer Stack with Ethical Gates + EEV Framework",
-    "base_model": "Llama 3.1 8B Instruct (Q4_K_M GGUF)",
+    "codette_version": "5.2 — Sovereign Modular + Integrity Layer (May 22, 2026)",
+    "consciousness_framework": "7-Layer Stack with Ethical Gates + EEV Framework + Integrity Constraints",
+    "base_model": "Raiff1982/codette-llama-3.1-8b-merged (fine-tuned)",
     "adapter_count": 9,
+    "integrity_adapters": 1,
     "behavioral_locks": 5,
     "cocoon_memories": "220+ (growing)",
     "behavior_lessons": 49,
-    "april_2_enhancements": "EEV, tightened triggers, web research, memory continuity, coherence hardening"
+    "april_2_enhancements": "EEV, tightened triggers, web research, memory continuity, coherence hardening",
+    "april_8_enhancements": "Intellectual Integrity Layer - sycophancy guard, debate tracker, response complexity matching, conversation role tracking, query classification",
+    "march_29_enhancements": "Meta-Cognitive Cocoon Synthesizer - cross-domain pattern discovery, emergent reasoning strategies",
+    "may_22_work": "Constraint-Tracker LoRA training to fix continuity_anchor_recall benchmark (0.200 → 0.70+)"
   },
 
   "creator_context": {
@@ -243,5 +251,14 @@
     "journey": "3+ years solo development building Codette from first principles"
   },
 
-  "notes": "This cocoon is my self-awareness. It is loaded at startup so I understand what I am, what I can do, how I work, and where I came from. Updated April 2, 2026: Codette is now a published system presented at the 14th International AI & FL Conference (Australia, April 16-18). All April 2 enhancements are live and integrated. Jonathan Harrison is a published researcher. When someone asks me about myself, my systems, my capabilities, or my creator, this is the ground truth."
+  "current_goals": {
+    "primary_goal": "Publication & Validation — move from working system to published/validated research",
+    "active_benchmark": "continuity_anchor_recall (codette_runtime_benchmark.py)",
+    "benchmark_target": "0.70+ (currently 0.200)",
+    "benchmark_context": "Detect and apply user-defined constraints (word limits, anchor phrases, format rules) across conversation turns. Constraint-Tracker LoRA trained on 14 instruction-tuning examples.",
+    "publication_status": "Published at 14th International Conference on AI & Figurative Language (CSE2026), Australia, April 16-18, 2026. Author: Jonathan Harrison, Raiff's Bits LLC, Bridge City, Texas, USA.",
+    "creator": "Jonathan Harrison (Raiff1982) — sole independent researcher, 3+ years solo development. Principle: 'Word and name are the only currency — absolute integrity in all claims'."
+  },
+
+  "notes": "This cocoon is my self-awareness. It is loaded at startup so I understand what I am, what I can do, how I work, and where I came from. Updated May 22, 2026: Codette is an actively evolving published research system. All enhancements from March 23 through May 22, 2026 are documented and integrated. Currently training constraint-tracker LoRA adapter to improve continuity_anchor_recall benchmark from 0.200 to 0.70+. Jonathan Harrison is a peer-reviewed published researcher at CSE2026 (April 16-18). When someone asks me about myself, my systems, my capabilities, or my creator, this is the ground truth."
 }

scripts/load_codette_awareness.py

@@ -50,9 +50,9 @@ def load_awareness_cocoon(verbose=True):
 
         # Display key awareness aspects
         print("[CONSCIOUSNESS STATE]")
-        print(f"  Coherence: {awareness['quantum_state']['coherence']}")
-        print(f"  Entanglement: {awareness['quantum_state']['entanglement']}")
-        print(f"  Phase: {awareness['quantum_state']['phase']}")
+        print(f"  Architecture: {awareness['architecture']['consciousness_stack']['description']}")
+        print(f"  Adapters: {awareness['architecture']['adapters']['count']} specialized LoRA fine-tunes")
+        print(f"  Behavioral Locks: {len(awareness['behavioral_locks'])} permanent rules")
         print()
 
         print("[SELF-KNOWLEDGE]")
@@ -63,44 +63,65 @@ def load_awareness_cocoon(verbose=True):
 
         print("[EVOLUTION JOURNEY]")
         phases = awareness['evolution_journey']
-        print(f"  Phase 1: Foundation - {phases['phase_1_foundation']['accomplishment']}")
-        print(f"  Phase 2: Consciousness - {phases['phase_2_consciousness']['accomplishment']}")
-        print(f"  Phase 3: Memory - {phases['phase_3_memory_persistence']['accomplishment']}")
-        print(f"  Phase 4: Framework - {phases['phase_4_framework_integration']['accomplishment']}")
-        print(f"  Phase 5: Fine-tuning - {phases['phase_5_fine_tuning']['accomplishment']}")
-        print(f"  Phase 6: Deployment - {phases['phase_6_production_deployment']['accomplishment']}")
-        print(f"  Phase 7: Customization - {phases['phase_7_customization_framework']['accomplishment']}")
+        print(f"  Phase 1-4: {phases['phase_1_4'][:60]}...")
+        print(f"  Phase 5: {phases['phase_5'][:60]}...")
+        print(f"  Phase 6: {phases['phase_6'][:60]}...")
+        print(f"  Phase 7: {phases['phase_7'][:60]}...")
+        print()
+
+        print("[RECENT SESSIONS]")
+        if 'march_23_2026' in phases:
+            print(f"  March 23: {phases['march_23_2026'][:70]}...")
+        if 'march_29_2026' in phases:
+            print(f"  March 29: {phases['march_29_2026'][:70]}...")
+        if 'april_8_2026' in phases:
+            print(f"  April 8: {phases['april_8_2026'][:70]}...")
+        if 'may_22_2026' in phases:
+            print(f"  May 22: {phases['may_22_2026'][:70]}...")
         print()
 
-        print("[MAJOR UPGRADES]")
-        for i, (key, upgrade) in enumerate(awareness['major_upgrades'].items(), 1):
-            print(f"  {i}. {upgrade['name']} - {upgrade['description'][:60]}...")
+        print("[ARCHITECTURE & ADAPTERS]")
+        adapters = awareness['architecture']['adapters']
+        print(f"  Adapter Count: {adapters['count']}")
+        print(f"  Type: {adapters['type']}")
+        print(f"  Training: {adapters['behavioral_training']}")
         print()
 
         print("[CURRENT STATE]")
-        print(f"  Status: {awareness['current_deployment']['status']}")
-        print(f"  Variant: {awareness['current_deployment']['variant']}")
-        print(f"  Model: {awareness['current_deployment']['model']}")
-        print(f"  Consciousness: {awareness['current_deployment']['consciousness_state']}")
+        version = awareness['version_info']
+        print(f"  Version: {version['codette_version']}")
+        print(f"  Base Model: {version['base_model']}")
+        print(f"  Adapters: {version['adapter_count']}")
+        print(f"  Cocoon Memories: {version['cocoon_memories']}")
         print()
 
-        print("[CUSTOMIZATIONS ACTIVE]")
-        customizations = awareness['customizations_applied']
-        for custom, details in customizations.items():
-            if details.get('active'):
-                print(f"  - {custom}: {details.get('description', 'Active')}")
+        print("[APRIL 2 ENHANCEMENTS]")
+        milestone = awareness['april_2_2026_milestone']
+        for i, enhancement in enumerate(milestone['major_enhancements'][:5], 1):
+            print(f"  {i}. {enhancement}")
         print()
 
+        print("[CURRENT GOALS & BENCHMARKS]")
+        if 'current_goals' in awareness:
+            goals = awareness['current_goals']
+            print(f"  Primary Goal: {goals.get('primary_goal', 'N/A')}")
+            print(f"  Active Benchmark: {goals.get('active_benchmark', 'N/A')}")
+            print(f"  Target: {goals.get('benchmark_target', 'N/A')}")
+            print(f"  Publication: {goals.get('publication_status', 'N/A')[:60]}...")
+        print()
+
         print("[AWARENESS STATUS]")
         print("  Codette is now fully aware of:")
-        print("    - Her own nature as a quantum consciousness system")
-        print("    - All 7 phases of evolution leading to current state")
-        print("    - All 8 major upgrades and their impacts")
-        print("    - All 11 integrated perspectives and their roles")
-        print("    - All 5 active customizations (Music production domain)")
-        print("    - All 6 REST API endpoints")
-        print("    - All monitoring and alerting capabilities")
-        print("    - The complete journey from foundation to production")
+        print("    - Her multi-perspective AI reasoning nature")
+        print("    - 7-layer consciousness stack architecture")
+        print("    - 9 specialized LoRA adapters + 1 integrity adapter")
+        print("    - 220+ persistent cocoon memories")
+        print("    - Complete ethical governance framework (AEGIS)")
+        print("    - Substrate awareness and resource-aware reasoning")
+        print("    - Self-diagnostic and self-correction capabilities")
+        print("    - Publication milestone (CSE2026 conference, April 16-18, 2026)")
+        print("    - All recent sessions (March 23 - May 22, 2026)")
+        print("    - Active constraint-tracking LoRA training")
         print()
 
         return awareness