@itz4blitz/agentful 0.1.0 → 0.1.4

package/.claude/skills/conversation/SKILL.md

@@ -0,0 +1,1130 @@
+---
+name: conversation
+description: Natural language understanding, intent classification, context management, reference resolution, and conversation history analysis for agentful
+model: sonnet
+tools: Read, Write, Edit, Glob, Grep
+---
+
+# Conversation Skill
+
+This skill provides natural language processing capabilities for understanding user intent, managing conversation context, resolving references, and maintaining conversation history.
+
+## Core Functions
+
+### 1. Intent Classification
+
+```typescript
+/**
+ * Classify user intent with confidence score
+ * @param message - User's current message
+ * @param conversation_history - Recent conversation context
+ * @param product_spec - Product features and requirements
+ * @returns Intent classification with confidence
+ */
+function classify_intent(
+  message: string,
+  conversation_history: ConversationMessage[],
+  product_spec: ProductSpec
+): IntentClassification {
+  const intents = [
+    'feature_request',
+    'bug_report',
+    'question',
+    'clarification',
+    'status_update',
+    'mind_change',
+    'context_switch',
+    'approval',
+    'rejection',
+    'pause',
+    'continue'
+  ];
+
+  // Analyze message patterns
+  const patterns = {
+    feature_request: /(?:add|create|implement|build|new|feature|support|enable)/i,
+    bug_report: /(?:bug|broken|error|issue|problem|wrong|doesn't work|fix)/i,
+    question: /(?:how|what|where|when|why|who|which|can you|explain)/i,
+    clarification: /(?:what do you mean|clarify|elaborate|more detail)/i,
+    status_update: /(?:status|progress|where are we|what's left)/i,
+    mind_change: /(?:actually|wait|never mind|forget that|change|instead)/i,
+    context_switch: /(?:let's talk about|switching to|moving on|about)/i,
+    approval: /(?:yes|ok|sure|go ahead|approved|sounds good|let's do it)/i,
+    rejection: /(?:no|stop|don't|cancel|never mind)/i,
+    pause: /(?:pause|hold on|wait|brb|later)/i,
+    continue: /(?:continue|resume|let's continue|back)/i
+  };
+
+  // Score each intent
+  const scores = {};
+  for (const [intent, pattern] of Object.entries(patterns)) {
+    const matches = message.match(pattern);
+    const baseScore = matches ? matches.length * 0.3 : 0;
+
+    // Context-aware scoring
+    let contextBonus = 0;
+    if (conversation_history.length > 0) {
+      const lastMessage = conversation_history[conversation_history.length - 1];
+      if (lastMessage.role === 'assistant' && intent === 'clarification') {
+        contextBonus += 0.2; // User asking for clarification after assistant response
+      }
+      if (lastMessage.intent === 'question' && intent === 'clarification') {
+        contextBonus += 0.15; // Follow-up clarification
+      }
+    }
+
+    scores[intent] = Math.min(0.95, baseScore + contextBonus);
+  }
+
+  // Find highest scoring intent
+  const topIntent = Object.entries(scores).reduce((a, b) =>
+    a[1] > b[1] ? a : b
+  );
+
+  return {
+    intent: topIntent[0],
+    confidence: topIntent[1],
+    alternative_intents: Object.entries(scores)
+      .filter(([_, score]) => score > 0.3)
+      .sort((a, b) => b[1] - a[1])
+      .slice(1, 4)
+      .map(([intent, score]) => ({ intent, confidence: score }))
+  };
+}
+
+interface IntentClassification {
+  intent: string;
+  confidence: number;
+  alternative_intents: Array<{ intent: string; confidence: number }>;
+}
+```
+
+### 2. Feature Extraction
+
+```typescript
+/**
+ * Extract which feature/user is talking about
+ * @param message - User's current message
+ * @param product_spec - Product features and requirements
+ * @returns Extracted feature reference
+ */
+function extract_feature_mention(
+  message: string,
+  product_spec: ProductSpec
+): FeatureMention {
+  const mentioned = [];
+
+  // Direct feature name matches
+  if (product_spec.features) {
+    for (const [featureId, feature] of Object.entries(product_spec.features)) {
+      const featureName = feature.name || featureId;
+      if (message.toLowerCase().includes(featureName.toLowerCase())) {
+        mentioned.push({
+          type: 'direct',
+          feature_id: featureId,
+          feature_name: featureName,
+          confidence: 0.9
+        });
+      }
+    }
+  }
+
+  // Hierarchical structure
+  if (product_spec.domains) {
+    for (const [domainId, domain] of Object.entries(product_spec.domains)) {
+      // Check domain mention
+      if (message.toLowerCase().includes(domain.name.toLowerCase())) {
+        mentioned.push({
+          type: 'domain',
+          domain_id: domainId,
+          domain_name: domain.name,
+          confidence: 0.85
+        });
+      }
+
+      // Check feature mentions within domain
+      if (domain.features) {
+        for (const [featureId, feature] of Object.entries(domain.features)) {
+          const featureName = feature.name || featureId;
+          if (message.toLowerCase().includes(featureName.toLowerCase())) {
+            mentioned.push({
+              type: 'feature',
+              domain_id: domainId,
+              feature_id: featureId,
+              feature_name: featureName,
+              confidence: 0.9
+            });
+          }
+        }
+      }
+    }
+  }
+
+  // Subtask mentions
+  const subtaskPattern = /(?:subtask|task|item)\s+(\d+)/i;
+  const subtaskMatch = message.match(subtaskPattern);
+  if (subtaskMatch) {
+    mentioned.push({
+      type: 'subtask_reference',
+      reference: subtaskMatch[1],
+      confidence: 0.7
+    });
+  }
+
+  // Return highest confidence mention or null
+  if (mentioned.length === 0) {
+    return { type: 'none', confidence: 0 };
+  }
+
+  return mentioned.sort((a, b) => b.confidence - a.confidence)[0];
+}
+
+interface FeatureMention {
+  type: 'direct' | 'domain' | 'feature' | 'subtask_reference' | 'none';
+  domain_id?: string;
+  domain_name?: string;
+  feature_id?: string;
+  feature_name?: string;
+  reference?: string;
+  confidence: number;
+}
+```
+
+### 3. Bug Description Extraction
+
+```typescript
+/**
+ * Extract bug details from conversation context
+ * @param message - User's current message
+ * @param conversation_history - Recent conversation context
+ * @returns Bug description with context
+ */
+function extract_bug_description(
+  message: string,
+  conversation_history: ConversationMessage[]
+): BugDescription {
+  const bugInfo = {
+    description: message,
+    steps_to_reproduce: [],
+    expected_behavior: null,
+    actual_behavior: null,
+    related_feature: null,
+    severity: 'unknown',
+    context_messages: []
+  };
+
+  // Look for "expected" vs "actual" patterns
+  const expectedPattern = /(?:expected|should|supposed to):\s*(.+?)(?:\.|$)/i;
+  const actualPattern = /(?:actually|but it|instead):\s*(.+?)(?:\.|$)/i;
+
+  const expectedMatch = message.match(expectedPattern);
+  const actualMatch = message.match(actualPattern);
+
+  if (expectedMatch) {
+    bugInfo.expected_behavior = expectedMatch[1].trim();
+  }
+  if (actualMatch) {
+    bugInfo.actual_behavior = actualMatch[1].trim();
+  }
+
+  // Look for numbered steps
+  const stepPattern = /^\d+\.\s*(.+)$/gm;
+  const steps = message.match(stepPattern);
+  if (steps) {
+    bugInfo.steps_to_reproduce = steps.map(step => step.replace(/^\d+\.\s*/, ''));
+  }
+
+  // Infer severity from keywords
+  const severeKeywords = ['crash', 'broken', 'fail', 'critical', 'blocking'];
+  const minorKeywords = ['typo', 'cosmetic', 'minor', 'polish'];
+
+  if (severeKeywords.some(kw => message.toLowerCase().includes(kw))) {
+    bugInfo.severity = 'high';
+  } else if (minorKeywords.some(kw => message.toLowerCase().includes(kw))) {
+    bugInfo.severity = 'low';
+  }
+
+  // Gather relevant context from conversation history
+  const relevantContext = conversation_history
+    .filter(msg => {
+      const msgTime = new Date(msg.timestamp);
+      const now = new Date();
+      const hoursDiff = (now.getTime() - msgTime.getTime()) / (1000 * 60 * 60);
+      return hoursDiff < 2; // Last 2 hours
+    })
+    .slice(-5); // Last 5 messages
+
+  bugInfo.context_messages = relevantContext;
+
+  return bugInfo;
+}
+
+interface BugDescription {
+  description: string;
+  steps_to_reproduce: string[];
+  expected_behavior: string | null;
+  actual_behavior: string | null;
+  related_feature: string | null;
+  severity: 'high' | 'medium' | 'low' | 'unknown';
+  context_messages: ConversationMessage[];
+}
+```
+
+### 4. Ambiguity Detection
+
+```typescript
+/**
+ * Detect unclear or ambiguous requests
+ * @param message - User's current message
+ * @returns Ambiguity analysis
+ */
+function detect_ambiguity(message: string): AmbiguityAnalysis {
+  const ambiguities = [];
+  let confidence = 0;
+
+  // Check for pronouns without context
+  const pronounPattern = /\b(it|that|this|they|them|those)\b/gi;
+  const pronouns = message.match(pronounPattern);
+  if (pronouns && pronouns.length > 0) {
+    // If message starts with pronoun, highly likely needs context
+    if (/^(it|that|this|they|them)/i.test(message.trim())) {
+      ambiguities.push({
+        type: 'pronoun_without_antecedent',
+        severity: 'high',
+        text: pronouns[0],
+        message: 'Pronoun at start of message without clear referent'
+      });
+      confidence = Math.max(confidence, 0.8);
+    }
+  }
+
+  // Check for vague verbs
+  const vagueVerbs = ['fix', 'update', 'change', 'improve', 'handle'];
+  const vagueVerbPattern = new RegExp(`\\b(${vagueVerbs.join('|')})\\b\\s+(?:it|that|this)`, 'i');
+  const vagueMatch = message.match(vagueVerbPattern);
+  if (vagueMatch) {
+    ambiguities.push({
+      type: 'vague_action',
+      severity: 'medium',
+      text: vagueMatch[0],
+      message: 'Action verb without specific target'
+    });
+    confidence = Math.max(confidence, 0.6);
+  }
+
+  // Check for short messages (< 10 words)
+  const wordCount = message.split(/\s+/).length;
+  if (wordCount < 10 && wordCount > 1) {
+    ambiguities.push({
+      type: 'insufficient_detail',
+      severity: 'low',
+      text: message,
+      message: 'Message is very short, may lack detail'
+    });
+    confidence = Math.max(confidence, 0.4);
+  }
+
+  // Check for multiple possible intents
+  const actionWords = message.split(/\s+/).filter(word =>
+    /^(add|create|fix|update|delete|remove|test|check|verify|deploy|build|run)/i.test(word)
+  );
+  if (actionWords.length > 2) {
+    ambiguities.push({
+      type: 'multiple_actions',
+      severity: 'medium',
+      text: actionWords.join(', '),
+      message: 'Multiple actions detected, unclear priority'
+    });
+    confidence = Math.max(confidence, 0.7);
+  }
+
+  return {
+    is_ambiguous: ambiguities.length > 0,
+    confidence,
+    ambiguities,
+    suggestion: ambiguities.length > 0 ? generate_clarification_suggestion(ambiguities) : null
+  };
+}
+
+interface AmbiguityAnalysis {
+  is_ambiguous: boolean;
+  confidence: number;
+  ambiguities: Array<{
+    type: string;
+    severity: 'high' | 'medium' | 'low';
+    text: string;
+    message: string;
+  }>;
+  suggestion: string | null;
+}
+
+function generate_clarification_suggestion(ambiguities: any[]): string {
+  const highSeverity = ambiguities.find(a => a.severity === 'high');
+  if (highSeverity) {
+    return 'Could you please provide more specific details? What specifically are you referring to?';
+  }
+  return 'I want to make sure I understand correctly. Could you provide a bit more detail?';
+}
+```
+
+### 5. Clarification Suggestions
+
+```typescript
+/**
+ * Generate helpful clarifying questions
+ * @param ambiguous_message - The ambiguous user message
+ * @param product_spec - Product features and requirements
+ * @returns Suggested clarifying questions
+ */
+function suggest_clarification(
+  ambiguous_message: string,
+  product_spec: ProductSpec
+): ClarificationSuggestion {
+  const questions = [];
+
+  // Feature-specific clarification
+  const featureMention = extract_feature_mention(ambiguous_message, product_spec);
+  if (featureMention.type === 'none') {
+    // No feature mentioned, ask which one
+    if (product_spec.features) {
+      const featureNames = Object.values(product_spec.features)
+        .map(f => f.name || f.id)
+        .slice(0, 5);
+      questions.push({
+        type: 'feature_selection',
+        question: `Which feature are you referring to? For example: ${featureNames.join(', ')}`,
+        options: featureNames
+      });
+    } else if (product_spec.domains) {
+      const domainNames = Object.values(product_spec.domains)
+        .map(d => d.name)
+        .slice(0, 5);
+      questions.push({
+        type: 'domain_selection',
+        question: `Which domain would you like to work on? For example: ${domainNames.join(', ')}`,
+        options: domainNames
+      });
+    }
+  }
+
+  // Action-specific clarification
+  const actions = ambiguous_message.match(/(?:add|create|fix|update|delete|remove|test|check)/gi);
+  if (actions && actions.length > 1) {
+    questions.push({
+      type: 'action_priority',
+      question: `I see multiple actions: ${actions.join(', ')}. Which would you like me to focus on first?`,
+      options: actions
+    });
+  }
+
+  // Bug report clarification
+  if (/bug|broken|error|issue/i.test(ambiguous_message)) {
+    questions.push({
+      type: 'bug_details',
+      question: 'Could you provide more details about the bug? What should happen vs. what actually happens?',
+      options: null
+    });
+  }
+
+  // Generic clarification if no specific ones
+  if (questions.length === 0) {
+    questions.push({
+      type: 'generic',
+      question: 'Could you please provide more details about what you\'d like me to do?',
+      options: null
+    });
+  }
+
+  return {
+    primary_question: questions[0].question,
+    follow_up_questions: questions.slice(1),
+    suggested_responses: questions[0].options || []
+  };
+}
+
+interface ClarificationSuggestion {
+  primary_question: string;
+  follow_up_questions: Array<{ type: string; question: string; options: string[] | null }>;
+  suggested_responses: string[];
+}
+```
+
+## Conversation State Management
+
+### State Structure
+
+```typescript
+/**
+ * Conversation state schema
+ * Stored in: .agentful/conversation-state.json
+ */
+interface ConversationState {
+  // Current context
+  current_feature: {
+    domain_id?: string;
+    feature_id?: string;
+    feature_name?: string;
+    subtask_id?: string;
+  } | null;
+
+  current_phase: 'idle' | 'planning' | 'implementing' | 'testing' | 'reviewing' | 'deploying';
+
+  last_action: {
+    type: string;
+    description: string;
+    timestamp: string;
+    result?: any;
+  } | null;
+
+  // Related features context
+  related_features: Array<{
+    feature_id: string;
+    feature_name: string;
+    relationship: 'dependency' | 'similar' | 'related';
+  }>;
+
+  // User preferences
+  user_preferences: {
+    communication_style: 'concise' | 'detailed' | 'balanced';
+    update_frequency: 'immediate' | 'summary' | 'on_completion';
+    ask_before_deleting: boolean;
+    test_automatically: boolean;
+  };
+
+  // Session tracking
+  session_start: string;
+  last_message_time: string;
+  message_count: number;
+
+  // Context health
+  context_health: {
+    is_stale: boolean;
+    last_confirmed_intent: string;
+    ambiguity_count: number;
+    clarification_count: number;
+  };
+}
+```
+
+### Reference Resolution
+
+```typescript
+/**
+ * Resolve pronouns and references to previous messages
+ * @param message - Current message with potential references
+ * @param conversation_history - Message history
+ * @param state - Current conversation state
+ * @returns Resolved message with references expanded
+ */
+function resolve_references(
+  message: string,
+  conversation_history: ConversationMessage[],
+  state: ConversationState
+): ResolvedMessage {
+  let resolved = message;
+  const references = [];
+
+  // Replace "it", "that", "this" with actual referents
+  const pronounMap = {
+    'it': state.current_feature?.feature_name,
+    'that': state.last_action?.description,
+    'this': state.current_feature?.feature_name
+  };
+
+  for (const [pronoun, referent] of Object.entries(pronounMap)) {
+    if (referent) {
+      const pattern = new RegExp(`\\b${pronoun}\\b`, 'gi');
+      if (pattern.test(message)) {
+        resolved = resolved.replace(pattern, referent);
+        references.push({
+          original: pronoun,
+          resolved: referent,
+          type: 'pronoun'
+        });
+      }
+    }
+  }
+
+  // Resolve "the feature", "the bug", etc.
+  const definiteReferences = {
+    'the feature': state.current_feature?.feature_name,
+    'the bug': state.last_action?.type === 'bug_fix' ? state.last_action.description : null,
+    'the task': state.current_feature?.subtask_id
+  };
+
+  for (const [phrase, referent] of Object.entries(definiteReferences)) {
+    if (referent) {
+      resolved = resolved.replace(new RegExp(phrase, 'gi'), referent);
+      references.push({
+        original: phrase,
+        resolved: referent,
+        type: 'definite_reference'
+      });
+    }
+  }
+
+  return {
+    original: message,
+    resolved,
+    references,
+    confidence: references.length > 0 ? 0.85 : 1.0
+  };
+}
+
+interface ResolvedMessage {
+  original: string;
+  resolved: string;
+  references: Array<{
+    original: string;
+    resolved: string;
+    type: string;
+  }>;
+  confidence: number;
+}
+```
+
+### Context Loss Recovery
+
+```typescript
+/**
+ * Detect and handle context loss (>24h gaps)
+ * @param conversation_history - Full conversation history
+ * @param state - Current conversation state
+ * @returns Context recovery recommendation
+ */
+function detect_context_loss(
+  conversation_history: ConversationMessage[],
+  state: ConversationState
+): ContextRecovery {
+  const now = new Date();
+  const lastMessage = conversation_history[conversation_history.length - 1];
+  const lastMessageTime = new Date(lastMessage?.timestamp || state.session_start);
+
+  const hoursDiff = (now.getTime() - lastMessageTime.getTime()) / (1000 * 60 * 60);
+
+  if (hoursDiff > 24) {
+    // Context is stale
+    return {
+      is_stale: true,
+      hours_since_last_message: Math.round(hoursDiff),
+      recommendation: 'summarize_and_confirm',
+      message: `It's been ${Math.round(hoursDiff)} hours since our last conversation. Before I continue, let me confirm what we were working on.`,
+      context_summary: {
+        last_feature: state.current_feature?.feature_name,
+        last_phase: state.current_phase,
+        last_action: state.last_action?.description
+      },
+      suggested_confirmation: `We were working on ${state.current_feature?.feature_name || 'a feature'}. Would you like to continue with that, or would you prefer to start something new?`
+    };
+  }
+
+  return {
+    is_stale: false,
+    hours_since_last_message: Math.round(hoursDiff),
+    recommendation: 'continue',
+    message: null
+  };
+}
+
+interface ContextRecovery {
+  is_stale: boolean;
+  hours_since_last_message: number;
+  recommendation: 'summarize_and_confirm' | 'continue' | 'reset';
+  message: string | null;
+  context_summary?: {
+    last_feature?: string;
+    last_phase?: string;
+    last_action?: string;
+  };
+  suggested_confirmation?: string;
+}
+```
+
+### User Preference Learning
+
+```typescript
+/**
+ * Learn and adapt to user preferences
+ * @param conversation_history - Recent conversation history
+ * @param state - Current conversation state
+ * @returns Updated user preferences
+ */
+function learn_user_preferences(
+  conversation_history: ConversationMessage[],
+  state: ConversationState
+): UserPreferences {
+  const preferences = { ...state.user_preferences };
+
+  // Analyze user's communication style
+  const userMessages = conversation_history.filter(m => m.role === 'user').slice(-20);
+  const avgMessageLength = userMessages.reduce((sum, m) =>
+    sum + m.content.split(/\s+/).length, 0
+  ) / userMessages.length;
+
+  if (avgMessageLength < 10) {
+    preferences.communication_style = 'concise';
+  } else if (avgMessageLength > 30) {
+    preferences.communication_style = 'detailed';
+  } else {
+    preferences.communication_style = 'balanced';
+  }
+
+  // Detect preference for updates
+  const statusRequests = userMessages.filter(m =>
+    /status|progress|where are we|what's left/i.test(m.content)
+  ).length;
+
+  if (statusRequests > 3) {
+    preferences.update_frequency = 'summary';
+  } else if (statusRequests === 0 && userMessages.length > 10) {
+    preferences.update_frequency = 'on_completion';
+  }
+
+  // Detect safety preference
+  const confirmationRequests = userMessages.filter(m =>
+    /are you sure|double check|verify|confirm/i.test(m.content)
+  ).length;
+
+  preferences.ask_before_deleting = confirmationRequests > 2;
+
+  // Detect testing preference
+  const testMentions = userMessages.filter(m =>
+    /test|testing|tests|coverage/i.test(m.content)
+  ).length;
+
+  preferences.test_automatically = testMentions > 2;
+
+  return preferences;
+}
+
+interface UserPreferences {
+  communication_style: 'concise' | 'detailed' | 'balanced';
+  update_frequency: 'immediate' | 'summary' | 'on_completion';
+  ask_before_deleting: boolean;
+  test_automatically: boolean;
+}
+```
+
+## Conversation History Management
+
+### History File Structure
+
+```bash
+# Stored in: .agentful/conversation-history.json
+{
+  "version": "1.0",
+  "session_id": "uuid",
+  "started_at": "2026-01-18T00:00:00Z",
+  "messages": [
+    {
+      "id": "msg-uuid",
+      "role": "user|assistant|system",
+      "content": "Message text",
+      "timestamp": "2026-01-18T00:00:00Z",
+      "intent": "feature_request",
+      "entities": {
+        "feature_id": "login",
+        "domain_id": "authentication"
+      },
+      "references_resolved": ["it -> login feature"]
+    }
+  ],
+  "context_snapshot": {
+    "current_feature": "login",
+    "current_phase": "implementing",
+    "related_features": ["register", "logout"]
+  }
+}
+```
+
+### History Operations
+
+```typescript
+/**
+ * Add message to conversation history
+ */
+function add_message_to_history(
+  message: ConversationMessage,
+  history_path: string = '.agentful/conversation-history.json'
+): void {
+  let history = read_conversation_history(history_path);
+
+  history.messages.push({
+    ...message,
+    id: message.id || generate_uuid(),
+    timestamp: message.timestamp || new Date().toISOString()
+  });
+
+  // Keep only last 100 messages to prevent file bloat
+  if (history.messages.length > 100) {
+    history.messages = history.messages.slice(-100);
+  }
+
+  Write(history_path, JSON.stringify(history, null, 2));
+}
+
+/**
+ * Read conversation history
+ */
+function read_conversation_history(
+  history_path: string = '.agentful/conversation-history.json'
+): ConversationHistory {
+  try {
+    const content = Read(history_path);
+    return JSON.parse(content);
+  } catch (error) {
+    // Initialize new history
+    return {
+      version: "1.0",
+      session_id: generate_uuid(),
+      started_at: new Date().toISOString(),
+      messages: [],
+      context_snapshot: null
+    };
+  }
+}
+
+/**
+ * Get recent conversation context
+ */
+function get_recent_context(
+  history_path: string = '.agentful/conversation-history.json',
+  message_count: number = 10
+): ConversationMessage[] {
+  const history = read_conversation_history(history_path);
+  return history.messages.slice(-message_count);
+}
+```
+
+## Integration with Orchestrator
+
+### Delegation Interface
+
+```typescript
+/**
+ * Determine if delegation to another skill is needed
+ * @param intent - Classified intent from user message
+ * @param entities - Extracted entities (features, domains, etc.)
+ * @returns Delegation decision
+ */
+function determine_delegation(
+  intent: IntentClassification,
+  entities: FeatureMention
+): DelegationDecision {
+  // Feature-related intents -> delegate to appropriate agent
+  if (['feature_request', 'bug_report', 'status_update'].includes(intent.intent)) {
+    if (entities.type === 'feature' || entities.type === 'direct') {
+      return {
+        should_delegate: true,
+        target_skill: determine_skill_for_feature(entities),
+        context: {
+          intent: intent.intent,
+          feature_id: entities.feature_id,
+          domain_id: entities.domain_id
+        }
+      };
+    }
+  }
+
+  // Validation request -> delegate to validation skill
+  if (/test|check|validate|verify/i.test(intent.intent)) {
+    return {
+      should_delegate: true,
+      target_skill: 'validation',
+      context: {
+        intent: intent.intent
+      }
+    };
+  }
+
+  // Product tracking -> delegate to product-tracking skill
+  if (['status_update', 'progress'].includes(intent.intent)) {
+    return {
+      should_delegate: true,
+      target_skill: 'product-tracking',
+      context: {
+        intent: intent.intent
+      }
+    };
+  }
+
+  return {
+    should_delegate: false,
+    target_skill: null,
+    context: null
+  };
+}
+
+interface DelegationDecision {
+  should_delegate: boolean;
+  target_skill: string | null;
+  context: any;
+}
+
+function determine_skill_for_feature(entities: FeatureMention): string {
+  // Map domains/features to appropriate skills
+  const domainSkillMap = {
+    'authentication': 'backend',
+    'user-management': 'backend',
+    'database': 'backend',
+    'frontend': 'frontend',
+    'ui': 'frontend',
+    'testing': 'tester'
+  };
+
+  if (entities.domain_name) {
+    const domainKey = Object.keys(domainSkillMap).find(key =>
+      entities.domain_name.toLowerCase().includes(key)
+    );
+    if (domainKey) {
+      return domainSkillMap[domainKey];
+    }
+  }
+
+  return 'backend'; // Default
+}
+```
+
+### Completion Tracking Integration
+
+```typescript
+/**
+ * Update conversation state after action completion
+ * @param state - Current conversation state
+ * @param action_result - Result from delegated skill
+ * @returns Updated conversation state
+ */
+function update_conversation_state(
+  state: ConversationState,
+  action_result: ActionResult
+): ConversationState {
+  const updated = { ...state };
+
+  // Update last action
+  updated.last_action = {
+    type: action_result.type,
+    description: action_result.description,
+    timestamp: new Date().toISOString(),
+    result: action_result
+  };
+
+  // Update phase based on action result
+  if (action_result.status === 'complete') {
+    if (state.current_phase === 'implementing') {
+      updated.current_phase = 'testing';
+    } else if (state.current_phase === 'testing') {
+      updated.current_phase = 'reviewing';
+    }
+  }
+
+  // Update message count and time
+  updated.message_count++;
+  updated.last_message_time = new Date().toISOString();
+
+  // Clear ambiguity count on successful action
+  if (action_result.status === 'success') {
+    updated.context_health.ambiguity_count = 0;
+    updated.context_health.clarification_count = 0;
+  }
+
+  return updated;
+}
+```
+
+## Edge Case Handling
+
+### Mind Changes
+
+```typescript
+/**
+ * Detect and handle user mind changes
+ * @param message - Current message
+ * @param state - Current conversation state
+ * @returns Mind change detection result
+ */
+function detect_mind_change(
+  message: string,
+  state: ConversationState
+): MindChangeDetection {
+  const mindChangePatterns = [
+    /actually/i,
+    /wait/i,
+    /never mind/i,
+    /forget that/i,
+    /change\s+my\s+mind/i,
+    /instead/i,
+    /stop\s+(?:that|it)/i
+  ];
+
+  const hasMindChange = mindChangePatterns.some(pattern => pattern.test(message));
+
+  if (hasMindChange) {
+    return {
+      detected: true,
+      confidence: 0.8,
+      previous_intent: state.last_action?.type,
+      suggestion: 'I understand you\'d like to change direction. What would you like to do instead?',
+      reset_context: /never mind|forget that|stop/i.test(message)
+    };
+  }
+
+  return {
+    detected: false,
+    confidence: 0,
+    previous_intent: null,
+    suggestion: null,
+    reset_context: false
+  };
+}
+
+interface MindChangeDetection {
+  detected: boolean;
+  confidence: number;
+  previous_intent: string | null;
+  suggestion: string | null;
+  reset_context: boolean;
+}
+```
+
+### Stale Context Handling
+
+```typescript
+/**
+ * Handle stale context when references no longer make sense
+ * @param message - Current message
+ * @param state - Current conversation state
+ * @returns Stale context handling response
+ */
+function handle_stale_context(
+  message: string,
+  state: ConversationState
+): StaleContextResponse {
+  // Check if current feature is referenced but stale
+  const featureReference = extract_feature_mention(message, null);
+
+  if (featureReference.type === 'none' && state.current_feature) {
+    // User said "it" but current feature is old
+    const hoursSinceLastAction = state.last_action
+      ? (Date.now() - new Date(state.last_action.timestamp).getTime()) / (1000 * 60 * 60)
+      : Infinity;
+
+    if (hoursSinceLastAction > 4) {
+      return {
+        is_stale: true,
+        suggestion: `I'm not sure what you're referring to. We last worked on ${state.current_feature.feature_name} ${Math.round(hoursSinceLastAction)} hours ago. Are you still referring to that?`,
+        confirm_needed: true
+      };
+    }
+  }
+
+  return {
+    is_stale: false,
+    suggestion: null,
+    confirm_needed: false
+  };
+}
+
+interface StaleContextResponse {
+  is_stale: boolean;
+  suggestion: string | null;
+  confirm_needed: boolean;
+}
+```
+
+## Usage Example
+
+```typescript
+// Complete conversation processing flow
+async function process_conversation(userMessage: string): Promise<ConversationResponse> {
+  // 1. Load conversation state and history
+  const state = load_conversation_state();
+  const history = read_conversation_history();
+  const productSpec = load_product_spec();
+
+  // 2. Check for context loss
+  const contextRecovery = detect_context_loss(history.messages, state);
+  if (contextRecovery.is_stale) {
+    return {
+      type: 'context_recovery',
+      message: contextRecovery.message,
+      confirmation_needed: true
+    };
+  }
+
+  // 3. Detect mind changes
+  const mindChange = detect_mind_change(userMessage, state);
+  if (mindChange.detected && mindChange.reset_context) {
+    state.current_feature = null;
+    state.current_phase = 'idle';
+    save_conversation_state(state);
+  }
+
+  // 4. Resolve references
+  const resolved = resolve_references(userMessage, history.messages, state);
+
+  // 5. Classify intent
+  const intent = classify_intent(resolved.resolved, history.messages, productSpec);
+
+  // 6. Extract entities
+  const entities = extract_feature_mention(resolved.resolved, productSpec);
+
+  // 7. Detect ambiguity
+  const ambiguity = detect_ambiguity(resolved.resolved);
+  if (ambiguity.is_ambiguous && ambiguity.confidence > 0.6) {
+    const clarification = suggest_clarification(resolved.resolved, productSpec);
+    return {
+      type: 'clarification_needed',
+      message: clarification.primary_question,
+      suggestions: clarification.suggested_responses
+    };
+  }
+
+  // 8. Determine delegation
+  const delegation = determine_delegation(intent, entities);
+
+  // 9. Add user message to history
+  add_message_to_history({
+    role: 'user',
+    content: userMessage,
+    intent: intent.intent,
+    entities: entities,
+    references_resolved: resolved.references
+  });
+
+  // 10. Return delegation decision or response
+  if (delegation.should_delegate) {
+    return {
+      type: 'delegate',
+      target_skill: delegation.target_skill,
+      context: delegation.context
+    };
+  }
+
+  return {
+    type: 'response',
+    message: generate_response(intent, entities, state)
+  };
+}
+```
+
+## File Locations
+
+```
+.agentful/
+├── conversation-state.json       # Current conversation state
+├── conversation-history.json     # Full message history
+└── user-preferences.json         # Learned user preferences (optional)
+```
+
+## Best Practices
+
+1. **Always resolve references** before processing intent
+2. **Check for context loss** after long gaps (>24h)
+3. **Detect ambiguity early** and ask clarifying questions
+4. **Learn user preferences** over time
+5. **Track related features** for better context
+6. **Handle mind changes gracefully** without losing all context
+7. **Delegate appropriately** based on intent and entities
+8. **Keep history manageable** (last 100 messages)
+9. **Update state after every action**
+10. **Provide summaries** after context recovery