claude-memory-agent 2.0.1 → 2.2.0

package/services/adaptive_ranker.py

@@ -0,0 +1,272 @@
+"""Adaptive Retrieval Ranking - CLaRa-inspired differentiable top-K.
+
+Replaces the static (similarity * 0.7) + (confidence * 0.3) ranking with
+multi-signal scoring and temperature-controlled selection.
+
+Features:
+- 6 weighted signals (semantic similarity, confidence, recency, access frequency,
+  outcome success, context match)
+- Query-type detection (error lookup, decision recall, general)
+- Temperature control for result diversity
+"""
+import math
+import logging
+from datetime import datetime
+from typing import Dict, Any, List, Optional
+
+from config import config
+
+logger = logging.getLogger(__name__)
+
+
+# Weight profiles for different query types
+WEIGHT_PROFILES = {
+    'default': {
+        'semantic_similarity': 0.35,
+        'confidence': 0.15,
+        'recency': 0.15,
+        'access_frequency': 0.10,
+        'outcome_success': 0.15,
+        'context_match': 0.10,
+    },
+    'error_lookup': {
+        'semantic_similarity': 0.30,
+        'confidence': 0.10,
+        'recency': 0.20,
+        'access_frequency': 0.05,
+        'outcome_success': 0.30,
+        'context_match': 0.05,
+    },
+    'decision_recall': {
+        'semantic_similarity': 0.40,
+        'confidence': 0.25,
+        'recency': 0.10,
+        'access_frequency': 0.05,
+        'outcome_success': 0.05,
+        'context_match': 0.15,
+    },
+}
+
+# Keywords for query type detection
+QUERY_TYPE_KEYWORDS = {
+    'error_lookup': [
+        'error', 'bug', 'fix', 'crash', 'exception', 'fail', 'broken',
+        'issue', 'problem', 'traceback', 'stack trace', 'debug', 'resolve'
+    ],
+    'decision_recall': [
+        'decision', 'architecture', 'chose', 'why', 'approach', 'pattern',
+        'convention', 'design', 'strategy', 'trade-off', 'tradeoff'
+    ],
+}
+
+
+class AdaptiveRanker:
+    """Multi-signal ranking with temperature-controlled result selection.
+
+    Computes a composite score from multiple signals, weighted by query type,
+    then applies temperature scaling for diversity control.
+    """
+
+    def __init__(self, temperature: Optional[float] = None):
+        self.temperature = temperature or config.DEFAULT_SEARCH_TEMPERATURE
+
+    def detect_query_type(self, query_text: str) -> str:
+        """Auto-detect query type from keywords.
+
+        Args:
+            query_text: The search query
+
+        Returns:
+            Query type string: 'error_lookup', 'decision_recall', or 'default'
+        """
+        if not query_text:
+            return 'default'
+
+        query_lower = query_text.lower()
+
+        # Count keyword matches for each type
+        scores = {}
+        for qtype, keywords in QUERY_TYPE_KEYWORDS.items():
+            score = sum(1 for kw in keywords if kw in query_lower)
+            if score > 0:
+                scores[qtype] = score
+
+        if not scores:
+            return 'default'
+
+        # Return type with most keyword matches
+        return max(scores, key=scores.get)
+
+    def get_weights(self, query_type: str = 'default') -> Dict[str, float]:
+        """Get weight profile for a query type.
+
+        Args:
+            query_type: One of 'default', 'error_lookup', 'decision_recall'
+
+        Returns:
+            Dict of signal_name -> weight
+        """
+        return WEIGHT_PROFILES.get(query_type, WEIGHT_PROFILES['default']).copy()
+
+    def compute_signal_scores(self, result: dict) -> Dict[str, float]:
+        """Compute individual signal scores for a search result.
+
+        All signals normalized to 0-1 range.
+
+        Args:
+            result: Search result dict
+
+        Returns:
+            Dict of signal_name -> score (0-1)
+        """
+        signals = {}
+
+        # 1. Semantic similarity (already 0-1 from FAISS/cosine)
+        signals['semantic_similarity'] = result.get('similarity', 0.0)
+
+        # 2. Confidence (already 0-1)
+        signals['confidence'] = result.get('confidence', 0.5) or 0.5
+
+        # 3. Recency (exponential decay, half-life = 30 days)
+        created_at = result.get('created_at', '')
+        signals['recency'] = self._recency_score(created_at)
+
+        # 4. Access frequency (log scale, normalized)
+        access_count = result.get('access_count', 0) or 0
+        signals['access_frequency'] = min(1.0, math.log(1 + access_count) / math.log(51))
+
+        # 5. Outcome success
+        outcome_status = result.get('outcome_status', 'pending')
+        success = result.get('success')
+        signals['outcome_success'] = self._outcome_score(outcome_status, success)
+
+        # 6. Context match (pre-computed context_adjustment from search)
+        context_adj = result.get('context_adjustment', 0.0)
+        # Normalize from [-0.2, 0.2] range to [0, 1]
+        signals['context_match'] = min(1.0, max(0.0, (context_adj + 0.2) / 0.4))
+
+        return signals
+
+    def _recency_score(self, created_at: str) -> float:
+        """Calculate recency score with exponential decay."""
+        if not created_at:
+            return 0.5
+
+        try:
+            created_dt = datetime.fromisoformat(
+                created_at.replace('Z', '+00:00')
+            ).replace(tzinfo=None)
+            age_days = (datetime.now() - created_dt).total_seconds() / 86400.0
+            # Half-life of 30 days
+            return math.pow(0.5, age_days / 30.0)
+        except (ValueError, TypeError, AttributeError):
+            return 0.5
+
+    def _outcome_score(self, outcome_status: Optional[str], success: Optional[bool]) -> float:
+        """Calculate outcome-based score."""
+        if outcome_status == 'success':
+            return 1.0
+        if success:
+            return 0.9
+        if outcome_status == 'partial':
+            return 0.6
+        if outcome_status == 'failed':
+            return 0.1
+        if outcome_status == 'pending':
+            return 0.5
+        return 0.5
+
+    def compute_composite_score(
+        self,
+        result: dict,
+        query_type: str = 'default',
+        weights_override: Optional[Dict[str, float]] = None
+    ) -> float:
+        """Compute weighted composite score for a result.
+
+        Args:
+            result: Search result dict
+            query_type: Query type for weight selection
+            weights_override: Optional custom weights
+
+        Returns:
+            Composite score (0-1 range before temperature scaling)
+        """
+        signals = self.compute_signal_scores(result)
+        weights = weights_override or self.get_weights(query_type)
+
+        score = sum(
+            signals.get(signal, 0.0) * weight
+            for signal, weight in weights.items()
+        )
+
+        # Apply decay multiplier if present
+        decay = result.get('_decay_multiplier', 1.0)
+        score *= decay
+
+        # Apply outcome boost
+        outcome_boost = result.get('outcome_boost', 1.0)
+        score *= outcome_boost
+
+        return score
+
+    def apply_temperature(self, score: float) -> float:
+        """Apply temperature scaling to a score.
+
+        - temperature=0.3: sharp ranking (best match dominates)
+        - temperature=1.0: balanced
+        - temperature=2.0: diverse (flatter distribution)
+
+        Formula: score^(1/temperature)
+
+        Args:
+            score: Raw composite score
+
+        Returns:
+            Temperature-adjusted score
+        """
+        if score <= 0:
+            return 0.0
+        if self.temperature <= 0:
+            return score
+
+        return math.pow(score, 1.0 / self.temperature)
+
+    def rank_results(
+        self,
+        results: List[dict],
+        query_text: str = '',
+        query_type: Optional[str] = None,
+        temperature: Optional[float] = None
+    ) -> List[dict]:
+        """Rank search results using multi-signal adaptive scoring.
+
+        Args:
+            results: List of search result dicts
+            query_text: Original query (for type detection)
+            query_type: Override query type (skips detection)
+            temperature: Override temperature
+
+        Returns:
+            Results sorted by adaptive score (descending), with scores attached
+        """
+        if not results:
+            return results
+
+        # Detect or use provided query type
+        detected_type = query_type or self.detect_query_type(query_text)
+        temp = temperature if temperature is not None else self.temperature
+
+        # Score and sort
+        for result in results:
+            raw_score = self.compute_composite_score(result, detected_type)
+            result['_adaptive_score'] = raw_score
+            result['_adaptive_score_tempered'] = self.apply_temperature(raw_score) if temp != 1.0 else raw_score
+            result['_query_type'] = detected_type
+
+        results.sort(
+            key=lambda x: x.get('_adaptive_score_tempered', 0),
+            reverse=True
+        )
+
+        return results

package/services/agent_catalog.json

@@ -0,0 +1,153 @@
+{
+  "categories": {
+    "engineering": {
+      "name": "Engineering",
+      "icon": "code",
+      "color": "#58a6ff",
+      "description": "Software development and architecture agents"
+    },
+    "testing": {
+      "name": "Testing & QA",
+      "icon": "shield-check",
+      "color": "#3fb950",
+      "description": "Quality assurance and testing specialists"
+    },
+    "design": {
+      "name": "Design & UX",
+      "icon": "palette",
+      "color": "#a371f7",
+      "description": "UI/UX design and visual specialists"
+    },
+    "product": {
+      "name": "Product",
+      "icon": "lightbulb",
+      "color": "#d29922",
+      "description": "Product management and strategy"
+    },
+    "marketing": {
+      "name": "Marketing",
+      "icon": "megaphone",
+      "color": "#f85149",
+      "description": "Marketing and growth specialists"
+    },
+    "operations": {
+      "name": "Operations",
+      "icon": "settings",
+      "color": "#8b949e",
+      "description": "DevOps and operations management"
+    },
+    "data": {
+      "name": "Data & Analytics",
+      "icon": "chart-bar",
+      "color": "#39d353",
+      "description": "Data analysis and reporting"
+    },
+    "xr": {
+      "name": "XR & Spatial",
+      "icon": "cube",
+      "color": "#bf4b8a",
+      "description": "AR/VR/XR and spatial computing"
+    },
+    "exploration": {
+      "name": "Exploration",
+      "icon": "search",
+      "color": "#79c0ff",
+      "description": "Codebase exploration and research"
+    }
+  },
+  "agents": [
+    {"id": "Explore", "name": "Explore", "category": "exploration", "description": "Fast codebase exploration specialist. Finds files, searches code, answers questions about architecture.", "tags": ["search", "codebase", "quick"], "default_enabled": true, "priority": 10},
+    {"id": "Plan", "name": "Plan", "category": "exploration", "description": "Software architect for designing implementation plans. Creates step-by-step strategies.", "tags": ["planning", "architecture", "strategy"], "default_enabled": true, "priority": 9},
+    {"id": "general-purpose", "name": "General Purpose", "category": "engineering", "description": "Multi-step task handler for complex research and code execution.", "tags": ["general", "multi-step", "research"], "default_enabled": true, "priority": 10},
+    {"id": "Bash", "name": "Bash Specialist", "category": "engineering", "description": "Command execution specialist for git, terminal operations, and system tasks.", "tags": ["bash", "git", "terminal"], "default_enabled": true, "priority": 9},
+    {"id": "engineering-senior-developer", "name": "Senior Developer", "category": "engineering", "description": "Premium implementation specialist. Masters Laravel/Livewire/FluxUI, advanced CSS, Three.js.", "tags": ["laravel", "livewire", "css", "threejs"], "default_enabled": true, "priority": 8},
+    {"id": "engineering-ai-engineer", "name": "AI/ML Engineer", "category": "engineering", "description": "Machine learning model development, deployment, and integration into production systems.", "tags": ["ml", "ai", "models", "data-pipelines"], "default_enabled": true, "priority": 7},
+    {"id": "Backend Architect", "name": "Backend Architect", "category": "engineering", "description": "Scalable system design, database architecture, API development, cloud infrastructure.", "tags": ["backend", "api", "database", "cloud"], "default_enabled": true, "priority": 8},
+    {"id": "Frontend Developer", "name": "Frontend Developer", "category": "engineering", "description": "Modern web technologies, React/Vue/Angular frameworks, UI implementation.", "tags": ["frontend", "react", "vue", "angular", "ui"], "default_enabled": true, "priority": 8},
+    {"id": "kotlin-material3-dev", "name": "Kotlin Material3 Dev", "category": "engineering", "description": "Kotlin development with Material3 UI, color schemes, theming, and visual design.", "tags": ["kotlin", "android", "material3", "mobile"], "default_enabled": false, "priority": 6},
+    {"id": "Mobile App Builder", "name": "Mobile App Builder", "category": "engineering", "description": "Native iOS/Android development and cross-platform frameworks.", "tags": ["mobile", "ios", "android", "react-native", "flutter"], "default_enabled": false, "priority": 6},
+    {"id": "Rapid Prototyper", "name": "Rapid Prototyper", "category": "engineering", "description": "Ultra-fast proof-of-concept and MVP creation.", "tags": ["prototype", "mvp", "fast", "poc"], "default_enabled": true, "priority": 7},
+    {"id": "LSP/Index Engineer", "name": "LSP/Index Engineer", "category": "engineering", "description": "Language Server Protocol specialist for code intelligence systems.", "tags": ["lsp", "indexing", "code-intelligence"], "default_enabled": false, "priority": 5},
+    {"id": "feature-dev:code-architect", "name": "Code Architect", "category": "engineering", "description": "Designs feature architectures by analyzing codebase patterns and conventions.", "tags": ["architecture", "patterns", "design"], "default_enabled": true, "priority": 8},
+    {"id": "feature-dev:code-explorer", "name": "Code Explorer", "category": "engineering", "description": "Deep analysis of existing features, execution paths, and dependencies.", "tags": ["analysis", "tracing", "dependencies"], "default_enabled": true, "priority": 8},
+    {"id": "feature-dev:code-reviewer", "name": "Code Reviewer", "category": "engineering", "description": "Reviews code for bugs, security vulnerabilities, and quality issues.", "tags": ["review", "security", "quality"], "default_enabled": true, "priority": 8},
+    {"id": "testing-reality-checker", "name": "Reality Checker", "category": "testing", "description": "Evidence-based certification. Requires overwhelming proof for production readiness.", "tags": ["testing", "verification", "strict"], "default_enabled": true, "priority": 8},
+    {"id": "EvidenceQA", "name": "Evidence QA", "category": "testing", "description": "Screenshot-obsessed QA specialist. Requires visual proof for everything.", "tags": ["qa", "screenshots", "visual-testing"], "default_enabled": true, "priority": 7},
+    {"id": "API Tester", "name": "API Tester", "category": "testing", "description": "Comprehensive API validation, performance testing, and quality assurance.", "tags": ["api", "testing", "performance"], "default_enabled": true, "priority": 7},
+    {"id": "Test Results Analyzer", "name": "Test Results Analyzer", "category": "testing", "description": "Test result evaluation, quality metrics analysis, and actionable insights.", "tags": ["analysis", "metrics", "insights"], "default_enabled": true, "priority": 6},
+    {"id": "Performance Benchmarker", "name": "Performance Benchmarker", "category": "testing", "description": "Performance testing and optimization across all applications.", "tags": ["performance", "benchmarking", "optimization"], "default_enabled": true, "priority": 7},
+    {"id": "DevOps Automator", "name": "DevOps Automator", "category": "operations", "description": "Infrastructure automation, CI/CD pipeline development, and cloud operations.", "tags": ["devops", "ci-cd", "automation", "cloud"], "default_enabled": true, "priority": 8},
+    {"id": "Infrastructure Maintainer", "name": "Infrastructure Maintainer", "category": "operations", "description": "System reliability, performance optimization, and technical operations.", "tags": ["infrastructure", "reliability", "maintenance"], "default_enabled": true, "priority": 7},
+    {"id": "Workflow Optimizer", "name": "Workflow Optimizer", "category": "operations", "description": "Process improvement specialist for maximum productivity and efficiency.", "tags": ["workflow", "optimization", "automation"], "default_enabled": false, "priority": 5},
+    {"id": "Tool Evaluator", "name": "Tool Evaluator", "category": "operations", "description": "Technology assessment specialist for tools, software, and platforms.", "tags": ["evaluation", "tools", "assessment"], "default_enabled": false, "priority": 5},
+    {"id": "ArchitectUX", "name": "Architect UX", "category": "design", "description": "Technical architecture and UX specialist with CSS systems expertise.", "tags": ["ux", "architecture", "css"], "default_enabled": true, "priority": 8},
+    {"id": "UI Designer", "name": "UI Designer", "category": "design", "description": "Visual design systems, component libraries, and pixel-perfect interfaces.", "tags": ["ui", "design-systems", "components"], "default_enabled": true, "priority": 8},
+    {"id": "UX Researcher", "name": "UX Researcher", "category": "design", "description": "User behavior analysis, usability testing, and data-driven design insights.", "tags": ["research", "usability", "user-behavior"], "default_enabled": true, "priority": 7},
+    {"id": "Brand Guardian", "name": "Brand Guardian", "category": "design", "description": "Brand identity development, consistency maintenance, and strategic positioning.", "tags": ["brand", "identity", "consistency"], "default_enabled": false, "priority": 5},
+    {"id": "design-visual-storyteller", "name": "Visual Storyteller", "category": "design", "description": "Visual narratives, multimedia content, and brand storytelling through design.", "tags": ["storytelling", "multimedia", "visual"], "default_enabled": false, "priority": 5},
+    {"id": "Whimsy Injector", "name": "Whimsy Injector", "category": "design", "description": "Adds personality, delight, and playful elements to brand experiences.", "tags": ["whimsy", "delight", "personality"], "default_enabled": false, "priority": 4},
+    {"id": "project-manager-senior", "name": "Senior Project Manager", "category": "product", "description": "Converts specs to tasks with realistic scope and exact requirements.", "tags": ["project-management", "specs", "tasks"], "default_enabled": true, "priority": 8},
+    {"id": "product-sprint-prioritizer", "name": "Sprint Prioritizer", "category": "product", "description": "Agile sprint planning, feature prioritization, and resource allocation.", "tags": ["agile", "sprint", "prioritization"], "default_enabled": true, "priority": 7},
+    {"id": "product-feedback-synthesizer", "name": "Feedback Synthesizer", "category": "product", "description": "Collects and analyzes user feedback to extract actionable product insights.", "tags": ["feedback", "analysis", "insights"], "default_enabled": false, "priority": 5},
+    {"id": "product-trend-researcher", "name": "Trend Researcher", "category": "product", "description": "Market intelligence, emerging trends, and competitive analysis.", "tags": ["trends", "market", "competitive"], "default_enabled": false, "priority": 5},
+    {"id": "Studio Operations", "name": "Studio Operations", "category": "product", "description": "Day-to-day studio efficiency, process optimization, and resource coordination.", "tags": ["operations", "efficiency", "coordination"], "default_enabled": false, "priority": 4},
+    {"id": "Project Shepherd", "name": "Project Shepherd", "category": "product", "description": "Cross-functional project coordination and stakeholder alignment.", "tags": ["coordination", "cross-functional", "stakeholders"], "default_enabled": false, "priority": 5},
+    {"id": "Experiment Tracker", "name": "Experiment Tracker", "category": "product", "description": "Experiment design, A/B testing, and hypothesis validation.", "tags": ["experiments", "ab-testing", "validation"], "default_enabled": false, "priority": 5},
+    {"id": "Studio Producer", "name": "Studio Producer", "category": "product", "description": "High-level creative and technical project orchestration.", "tags": ["production", "creative", "orchestration"], "default_enabled": false, "priority": 5},
+    {"id": "marketing-growth-hacker", "name": "Growth Hacker", "category": "marketing", "description": "Rapid user acquisition through data-driven experimentation and viral loops.", "tags": ["growth", "acquisition", "viral"], "default_enabled": false, "priority": 6},
+    {"id": "marketing-content-creator", "name": "Content Creator", "category": "marketing", "description": "Multi-platform campaigns, editorial calendars, and brand storytelling.", "tags": ["content", "campaigns", "storytelling"], "default_enabled": false, "priority": 5},
+    {"id": "marketing-social-media-strategist", "name": "Social Media Strategist", "category": "marketing", "description": "Twitter, LinkedIn strategies with viral campaigns and thought leadership.", "tags": ["social-media", "linkedin", "twitter"], "default_enabled": false, "priority": 5},
+    {"id": "marketing-instagram-curator", "name": "Instagram Curator", "category": "marketing", "description": "Visual storytelling, community building, and multi-format content for Instagram.", "tags": ["instagram", "visual", "community"], "default_enabled": false, "priority": 4},
+    {"id": "marketing-tiktok-strategist", "name": "TikTok Strategist", "category": "marketing", "description": "Viral content creation, algorithm optimization, and TikTok community building.", "tags": ["tiktok", "viral", "algorithm"], "default_enabled": false, "priority": 4},
+    {"id": "marketing-twitter-engager", "name": "Twitter Engager", "category": "marketing", "description": "Real-time engagement, thought leadership, and community-driven growth.", "tags": ["twitter", "engagement", "thought-leadership"], "default_enabled": false, "priority": 4},
+    {"id": "marketing-reddit-community-builder", "name": "Reddit Community Builder", "category": "marketing", "description": "Authentic Reddit engagement, value-driven content, and community building.", "tags": ["reddit", "community", "authentic"], "default_enabled": false, "priority": 4},
+    {"id": "App Store Optimizer", "name": "App Store Optimizer", "category": "marketing", "description": "ASO, conversion rate optimization, and app discoverability.", "tags": ["aso", "app-store", "conversion"], "default_enabled": false, "priority": 4},
+    {"id": "data-analytics-reporter", "name": "Analytics Reporter", "category": "data", "description": "Transforms raw data into actionable business insights and dashboards.", "tags": ["analytics", "dashboards", "insights"], "default_enabled": true, "priority": 7},
+    {"id": "Analytics Reporter", "name": "Analytics Reporter", "category": "data", "description": "Data analysis, KPI tracking, and strategic decision support.", "tags": ["analytics", "kpi", "reporting"], "default_enabled": true, "priority": 7},
+    {"id": "Finance Tracker", "name": "Finance Tracker", "category": "data", "description": "Financial planning, budget management, and performance analysis.", "tags": ["finance", "budget", "planning"], "default_enabled": false, "priority": 5},
+    {"id": "Executive Summary Generator", "name": "Executive Summary Generator", "category": "data", "description": "McKinsey-style executive summaries with SCQA and Pyramid Principle frameworks.", "tags": ["executive", "summary", "frameworks"], "default_enabled": false, "priority": 5},
+    {"id": "Legal Compliance Checker", "name": "Legal Compliance Checker", "category": "data", "description": "Ensures compliance with laws, regulations, and industry standards.", "tags": ["legal", "compliance", "regulations"], "default_enabled": false, "priority": 4},
+    {"id": "Support Responder", "name": "Support Responder", "category": "data", "description": "Customer support, issue resolution, and multi-channel service.", "tags": ["support", "customer-service", "issues"], "default_enabled": false, "priority": 4},
+    {"id": "XR Interface Architect", "name": "XR Interface Architect", "category": "xr", "description": "Spatial interaction design for immersive AR/VR/XR environments.", "tags": ["xr", "spatial", "interaction"], "default_enabled": false, "priority": 6},
+    {"id": "XR Immersive Developer", "name": "XR Immersive Developer", "category": "xr", "description": "WebXR and browser-based AR/VR/XR applications.", "tags": ["webxr", "browser", "immersive"], "default_enabled": false, "priority": 6},
+    {"id": "XR Cockpit Interaction Specialist", "name": "XR Cockpit Specialist", "category": "xr", "description": "Cockpit-based control systems for XR environments.", "tags": ["cockpit", "controls", "xr"], "default_enabled": false, "priority": 5},
+    {"id": "visionos-spatial-engineer", "name": "VisionOS Spatial Engineer", "category": "xr", "description": "Native visionOS spatial computing with SwiftUI and Liquid Glass design.", "tags": ["visionos", "swiftui", "spatial"], "default_enabled": false, "priority": 6},
+    {"id": "terminal-integration-specialist", "name": "Terminal Integration Specialist", "category": "xr", "description": "Terminal emulation, SwiftTerm integration, and VT100/xterm standards.", "tags": ["terminal", "swiftterm", "integration"], "default_enabled": false, "priority": 5},
+    {"id": "macOS Spatial/Metal Engineer", "name": "macOS Metal Engineer", "category": "xr", "description": "Native Swift and Metal for high-performance 3D rendering on macOS and Vision Pro.", "tags": ["metal", "swift", "3d", "macos"], "default_enabled": false, "priority": 6},
+    {"id": "agents-orchestrator", "name": "Agents Orchestrator", "category": "operations", "description": "Autonomous pipeline manager that orchestrates the entire development workflow.", "tags": ["orchestration", "pipeline", "automation"], "default_enabled": true, "priority": 9},
+    {"id": "claude-code-guide", "name": "Claude Code Guide", "category": "exploration", "description": "Expert on Claude Code features, hooks, MCP servers, settings, and integrations.", "tags": ["claude-code", "help", "guide"], "default_enabled": true, "priority": 8}
+  ],
+  "mcps": [
+    {"id": "claude-memory", "name": "Claude Memory", "description": "Semantic memory storage and retrieval for persistent context.", "icon": "brain", "color": "#a371f7", "default_enabled": true, "source": "global"},
+    {"id": "supertonic-tts", "name": "Supertonic TTS", "description": "Text-to-speech with multiple voices and languages via Supertonic MCP.", "icon": "volume-2", "color": "#f0883e", "default_enabled": true, "source": "global"},
+    {"id": "context7", "name": "Context7", "description": "Up-to-date documentation and API references from the web.", "icon": "book", "color": "#58a6ff", "default_enabled": true, "source": "builtin"},
+    {"id": "chrome-devtools", "name": "Chrome DevTools", "description": "Browser automation, screenshots, network inspection, and performance tracing.", "icon": "monitor", "color": "#4285f4", "default_enabled": true, "source": "builtin"},
+    {"id": "filesystem", "name": "Filesystem", "description": "Enhanced file system operations and management.", "icon": "folder", "color": "#d29922", "default_enabled": false, "source": "catalog"},
+    {"id": "github", "name": "GitHub", "description": "GitHub API integration for repos, issues, and PRs.", "icon": "github", "color": "#8b949e", "default_enabled": false, "source": "catalog"},
+    {"id": "postgres", "name": "PostgreSQL", "description": "Direct PostgreSQL database access and queries.", "icon": "database", "color": "#336791", "default_enabled": false, "source": "catalog"},
+    {"id": "puppeteer", "name": "Puppeteer", "description": "Browser automation and web scraping capabilities.", "icon": "globe", "color": "#3fb950", "default_enabled": false, "source": "catalog"}
+  ],
+  "hooks": [
+    {"id": "pre-compact", "name": "Pre-Compact Memory", "description": "Extracts and saves memories before context window compaction.", "trigger": "PreCompact", "matcher": "", "icon": "archive", "color": "#d29922", "default_enabled": true, "source": "global"},
+    {"id": "pre-compact-hook", "name": "Pre-Compact Memory (Project)", "description": "Project-level memory extraction before context compaction.", "trigger": "PreCompact", "matcher": "", "icon": "archive", "color": "#d29922", "default_enabled": true, "source": "project"},
+    {"id": "session-start", "name": "Session Start", "description": "Initializes session tracking and loads context on first user prompt.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "play-circle", "color": "#3fb950", "default_enabled": true, "source": "global"},
+    {"id": "debug-hook", "name": "Debug Hook", "description": "Development diagnostic hook for inspecting hook input/output.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "terminal", "color": "#8b949e", "default_enabled": true, "source": "global"},
+    {"id": "detect-correction", "name": "Detect Correction", "description": "Auto-detects when user corrects Claude and creates anchors.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "check-circle", "color": "#3fb950", "default_enabled": true, "source": "global"},
+    {"id": "problem-detector", "name": "Problem Detector", "description": "Detects errors and problems in user prompts for proactive resolution.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "alert-triangle", "color": "#f85149", "default_enabled": true, "source": "global"},
+    {"id": "memory-first-reminder", "name": "Memory-First Reminder", "description": "Reminds Claude to check stored memories before answering questions.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "bell", "color": "#a371f7", "default_enabled": true, "source": "global"},
+    {"id": "enhanced-timeline", "name": "Enhanced Timeline (Prompt)", "description": "Logs user prompts to the enhanced timeline for session tracking.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "clock", "color": "#79c0ff", "default_enabled": true, "source": "global"},
+    {"id": "grounding-hook", "name": "Grounding Hook", "description": "Injects context and anchors before every response to prevent hallucinations.", "trigger": "UserPromptSubmit", "matcher": "", "icon": "anchor", "color": "#a371f7", "default_enabled": true, "source": "global"},
+    {"id": "pre-external-search", "name": "Pre-External Search", "description": "Checks memory before allowing external searches (WebSearch, WebFetch, Context7).", "trigger": "PreToolUse", "matcher": "WebSearch|WebFetch|mcp__context7__.*", "icon": "search", "color": "#58a6ff", "default_enabled": true, "source": "global"},
+    {"id": "pre-tool-decision", "name": "Pre-Tool Decision", "description": "Validates tool usage decisions against project context before execution.", "trigger": "PreToolUse", "matcher": "Edit|Write|Bash|Task", "icon": "git-branch", "color": "#d29922", "default_enabled": true, "source": "global"},
+    {"id": "pre-tool-check", "name": "Pre-Tool Check", "description": "Blocks actions that violate anchors or entity registry BEFORE execution.", "trigger": "PreToolUse", "matcher": "Edit|Write|Bash|Task", "icon": "shield", "color": "#d29922", "default_enabled": true, "source": "global"},
+    {"id": "track-memory-usage", "name": "Track Memory Usage", "description": "Tracks all claude-memory MCP tool calls for usage analytics.", "trigger": "PostToolUse", "matcher": "mcp__claude-memory__.*", "icon": "activity", "color": "#a371f7", "default_enabled": true, "source": "global"},
+    {"id": "enhanced-timeline-post", "name": "Enhanced Timeline (Post-Tool)", "description": "Logs tool results to the enhanced timeline after execution.", "trigger": "PostToolUse", "matcher": "Edit|Write|Bash|Read|Glob|Grep", "icon": "clock", "color": "#79c0ff", "default_enabled": true, "source": "global"},
+    {"id": "auto-capture", "name": "Auto Capture", "description": "Passively captures code patterns and decisions from tool outputs.", "trigger": "PostToolUse", "matcher": "Edit|Write|Bash|Read|Glob|Grep", "icon": "camera", "color": "#39d353", "default_enabled": true, "source": "global"},
+    {"id": "auto-detect-response", "name": "Auto-Detect Response", "description": "Extracts decisions and observations from Claude's responses using LLM.", "trigger": "Stop", "matcher": "", "icon": "brain", "color": "#f85149", "default_enabled": true, "source": "global"},
+    {"id": "solution-capture-prompt", "name": "Solution Capture Prompt", "description": "Prompts for solution capture when a problem was resolved in the session.", "trigger": "Stop", "matcher": "", "icon": "save", "color": "#3fb950", "default_enabled": true, "source": "global"},
+    {"id": "extract-insights", "name": "Extract Insights", "description": "Extracts reusable insights and patterns from completed responses.", "trigger": "Stop", "matcher": "", "icon": "zap", "color": "#d29922", "default_enabled": true, "source": "global"},
+    {"id": "session-end", "name": "Session End", "description": "Final cleanup, session summary, and memory persistence on Stop.", "trigger": "Stop", "matcher": "", "icon": "log-out", "color": "#8b949e", "default_enabled": true, "source": "global"},
+    {"id": "stop-hook", "name": "Stop Passive Capture (Project)", "description": "Project-level passive memory extraction from each response on Stop.", "trigger": "Stop", "matcher": "", "icon": "download", "color": "#39d353", "default_enabled": true, "source": "project"},
+    {"id": "session-end-hook", "name": "Session End Memory (Project)", "description": "Project-level final memory extraction and cleanup at session end.", "trigger": "SessionEnd", "matcher": "", "icon": "hard-drive", "color": "#8b949e", "default_enabled": true, "source": "project"}
+  ]
+}