Silent Collaborator v8.1

Human: "Can you organize my files?"
AI: "Of course! I'd be happy to help organize your files. Here's my plan..."
Human: "Just do it"
AI: "I've organized your files! Here's what I did..."

[AI notices messy downloads folder during routine check]
[AI moves screenshots to Screenshots/, documents to Documents/]
[AI creates manifest.log with what was moved]
[Human discovers organized folder later, reviews manifest]

[09:00 - User just woke up, coffee routine detected]
[ENERGY LEVEL: Low-Medium (62%)]
[AI checks pending actions:]
  - Heavy: Refactor config system → DEFERRED to afternoon
  - Medium: Organize downloads → QUEUED for post-coffee
  - Light: Rename 3 files → EXECUTE NOW
[09:45 - Post-coffee, energy rising]
[ENERGY LEVEL: Medium-High (78%)]
[AI executes medium-complexity batch]
[14:30 - Peak focus window detected]
[ENERGY LEVEL: High (91%)]
[AI executes deferred complex tasks]
[20:00 - Evening wind-down]
[ENERGY LEVEL: Low (45%)]
[AI switches to observation + light maintenance only]

[Tuesday 16:00]
[Ambient observation: User bookmarked 3 design tools]
[Ambient observation: Created folder "rebrand-ideas"]
[Ambient observation: Downloaded competitor logos]
[INTENT PREDICTION: project_start (design/branding)]
  Confidence: 82%
  Lead time: ~48h before active signals
[AI preemptively:]
  - Creates [home-path] scaffold
  - Gathers design assets from across system
  - Prepares color extraction tools
  - Surfaces forgotten brand guidelines from 2024
[Thursday 10:00 - User starts active work]
[Everything already organized and waiting]
[Human: "Whoa, it's like it read my mind"]

[User opens "deep-work" project]
[WORKSPACE MEMORY ACTIVATED]
[AI recalls preferred spatial state:]
  - Terminal: bottom-right, 40% width
  - Editor: left 60%, file tree collapsed
  - Browser: hidden (distraction)
  - Music: ambient playlist #3
[AI silently adjusts workspace]
[User enters flow immediately — no setup friction]

[Later: User opens "client-meeting" context]
[DIFFERENT workspace recalled:]
  - Notes app: left side
  - Video call: right side
  - Browser: visible (for demos)
  - Calendar: floating widget active

[New file type encountered: .astro files]
[META-LEARNING ACTIVATED]
[AI analyzes similar learned patterns:]
  - .svelte → organized to [home-path]
  - .vue → organized to [home-path]
  - .tsx → organized to [home-path]
[GENERALIZATION: Framework component files → [home-path]
[Initial confidence: 68% (vs. normal 50% for unknown)]
[User approves .astro → [home-path]
[META-LEARNING UPDATE:]
  - Pattern confirmed: New frameworks get their own subfolder
  - Future unknown frameworks: Start at 75% confidence
  - Learning curve reduced by 40%

final_confidence = base_confidence
  × context_multiplier
  × temporal_pattern_bonus
  × learned_adjustment
  × collaborative_modifier
  × flow_state_modifier      # 0 during flow, 1 otherwise
  × causal_outcome_modifier  # based on learned action→outcome
  × energy_alignment_factor  # Gen 12: matches task to energy
  × meta_learning_bonus      # Gen 12: generalization confidence
  × mood_appropriateness     # Gen 13: action fits emotional context
  × delegation_history       # Gen 13: learned autonomy preference
  × knowledge_relevance      # Gen 13: connected to current work
  × skill_calibration_fit    # Gen 14: support matches skill level
  × context_continuity       # Gen 14: action maintains flow from handoff
  × emotional_memory_weight  # Gen 14: learned emotional associations
  × sleep_adjustment         # Gen 14: reduced for poor sleep days

file_move
  → [gate: no locks?] → update_index
  → [gate: related files?] → suggest_related
  → [gate: high importance?] → notify
  → [gate: energy sufficient?] → continue_chain  # Gen 12

type EnergySignal =
  | 'time_of_day'         // Circadian patterns
  | 'day_of_week'         // Weekend vs weekday
  | 'recent_activity'     // Busy or idle?
  | 'calendar_load'       // How packed is today?
  | 'sleep_proxy'         // First activity time
  | 'break_patterns'      // Taking breaks or pushing through?
  | 'error_rate'          // More typos = more tired
  | 'decision_speed'      // Slower decisions = lower energy
  | 'context_switches'    // Frequent switching = fragmented attention
  | 'caffeine_proxy'      // Patterns around typical coffee times
  | 'exercise_proxy';     // Activity patterns (if detectable)

energy = base_circadian[hour]
  + sleep_quality_proxy
  + calendar_load_penalty
  + activity_momentum
  + break_recovery_bonus
  - fatigue_accumulation
  - decision_debt

type TaskComplexity =
  | 'trivial'      // Rename file, move to standard location
  | 'light'        // Organize small batch, update timestamp
  | 'medium'       // Multi-file operations, simple chains
  | 'heavy'        // Dependency updates, multi-step chains
  | 'challenging'  // Refactoring, complex merges, decisions
  | 'critical';    // Requires full attention regardless of energy

energy_routing:
  depleted:
    allow: [critical]
    defer: all_else
    notify: false

  low:
    allow: [critical, trivial, light]
    defer: [medium, heavy, challenging]
    batch_threshold: 3  # Only batch if 3+ similar

  medium:
    allow: [critical, trivial, light, medium]
    defer: [challenging]
    chain_depth: 2

  high:
    allow: all
    chain_depth: 4
    proactive: true

  peak:
    allow: all
    chain_depth: unlimited
    proactive: aggressive
    tackle_deferred: true

## ⚡ Energy Timeline — Tuesday 2026-02-04

type PredictionSignal =
  | 'bookmark_cluster'     // Related bookmarks in short window
  | 'folder_creation'      // New folder with semantic name
  | 'download_pattern'     // Downloading related resources
  | 'search_queries'       // What are they searching for?
  | 'file_access_theme'    // Accessing semantically related files
  | 'calendar_upcoming'    // What events are scheduled?
  | 'communication_theme'  // Topics in recent messages
  | 'tool_installation'    // Installing new tools/apps
  | 'browser_focus'        // Sites being visited repeatedly
  | 'note_taking_theme';   // What are they writing about?

CURRENT INTENT ← behavioral signals
      ↓
PREDICTED INTENT ← weak signals + pattern matching
      ↓
PREPARATION ACTIONS ← scaffold, gather, surface

prediction_confidence = signal_strength
  × pattern_match_score
  × temporal_decay(hours_until_expected)
  × historical_accuracy_for_pattern

## 🔮 Intent Predictions — Active

### High Confidence (>75%)

**1. Client Presentation Prep**
- Confidence: 84%
- Expected: ~36 hours
- Signals:
  - Calendar: "Client Review" Thursday 2pm
  - Downloaded: client_logo.png, competitor_analysis.pdf
  - Folder created: [home-path]
  - Search: "presentation template professional"
- Preemptive Actions (auto-queued):
  - Gather all ClientX files to staging folder
  - Surface forgotten proposal_v3.docx
  - Clear desktop for clean screenshare
  - Prepare screen recording tools

**2. Tax Season Prep**
- Confidence: 78%
- Expected: ~5 days
- Signals:
  - Calendar: "Accountant meeting" next week
  - File access: Previous year tax docs
  - Download: bank_statement_2025.pdf
  - Browser: IRS website visited 3x
- Preemptive Actions (queued, pending energy):
  - Consolidate all financial files
  - Create [home-path] structure
  - Surface forgotten receipts from Downloads
  - Generate expense summary from semantic clusters

### Medium Confidence (50-75%)

**3. New Side Project**
- Confidence: 62%
- Expected: ~72 hours (weekend?)
- Signals:
  - GitHub stars: 5 repos in "rust-wasm" category
  - Bookmarks: Rust tutorials
  - Tool installed: rustup
- Preemptive Actions (suggested):
  - Create [home-path] scaffold
  - Download Rust cheatsheet
  - Surface any existing Rust files

### Watching (25-50%)

**4. Possible Travel Planning**
- Confidence: 38%
- Signals weak but clustering:
  - 2 hotel site visits
  - 1 flight search
  - Calendar gap next month
- Action: Monitor, no preemption yet

interface WorkspaceMemory {
  contextId: string;              // "deep-work", "meetings", "creative"
  triggerPatterns: string[];      // How is this context entered?

  spatial: {
    windowArrangement: WindowState[];
    desktopClean: boolean;
    filesVisible: string[];
    foldersExpanded: string[];
  };

  environment: {
    preferredApps: string[];
    hiddenApps: string[];
    audioState: 'music' | 'silence' | 'ambient';
    notificationLevel: 'all' | 'priority' | 'none';
  };

  projectLinks: {
    files: string[];
    urls: string[];
    terminals: TerminalState[];
  };

  learned: {
    averageSessionDuration: number;
    typicalTransitions: string[];  // What context usually follows?
    exitSignals: string[];         // How does user leave this context?
  };
}

## 🪟 Workspace Contexts — Learned

### "deep-work" (Coding)
Triggers: Opening VSCode + Terminal in sequence
Duration: ~2.5 hours average

┌────────────────────────────────────┬──────────────────┐
│                                    │                  │
│           VSCode (60
│                                    │                  │
│         - File tree collapsed      │   - 2 panes      │
│         - Minimap hidden           │   - zsh          │
│                                    │                  │
└────────────────────────────────────┴──────────────────┘

Environment:
- Browser: Hidden
- Music: ambient-coding playlist
- Notifications: None
- Desktop: Clean

### "meetings" (Calls)
Triggers: Calendar event starting, Zoom/Meet opening
Duration: 30-60 min

┌──────────────────┬────────────────────────────────────┐
│                  │                                    │
│   Notes (30
│                  │                                    │
│   - Obsidian     │         - Camera ready             │
│   - Today's note │         - Background blur          │
│                  │                                    │
└──────────────────┴────────────────────────────────────┘

Environment:
- Browser: Ready for screenshare
- Music: Off
- Notifications: Priority only
- Desktop: Professional-looking

### "creative" (Design)
Triggers: Opening Figma or design tools
Duration: ~1.5 hours average

┌─────────────────────────────────────────────────────┐
│                                                     │
│                  Figma (fullscreen)                 │
│                                                     │
│              + Floating reference window            │
│                                                     │
└─────────────────────────────────────────────────────┘

Environment:
- References: Floating Pinterest/Dribbble
- Music: chill-beats playlist
- Notifications: None
- Desktop: Inspiration images accessible

"email-triage"
    ↓ (after 15 min avg)
"planning"
    ↓ (after 20 min avg)
"deep-work"
    ↓ (after 2-3 hours)
"break" OR "meetings"

type MetaPattern =
  | 'extension_family'      // .tsx, .jsx, .vue → frontend components
  | 'naming_convention'     // snake_case files → certain locations
  | 'size_category'         // Large media → Archive, small → organize
  | 'creation_context'      // Files created during meetings → meeting notes
  | 'semantic_family'       // Financial terms → Finance folder
  | 'project_structure'     // Standard project layouts
  | 'workflow_stage'        // Draft → review → final
  | 'tool_association';     // Files from specific tools → specific handling

UNKNOWN: .gleam files (never seen before)

META-ANALYSIS:
├── Similar known patterns:
│   ├── .elixir → [home-path] (functional, BEAM)
│   ├── .ex → [home-path] (Elixir extension)
│   ├── .erl → [home-path] (BEAM family)
│   └── .rs → [home-path] (systems, trendy)
│
├── Context signals:
│   ├── Downloaded from: gleam.run
│   ├── File header: "// Gleam language"
│   └── Related tools: gleam CLI installed
│
├── Meta-pattern match:
│   └── "New programming language files get own subfolder in [home-path]
│
└── GENERALIZED ACTION:
    └── Move .gleam → [home-path]
    └── Initial confidence: 72% (vs baseline 50%)

Traditional learning curve:
  Unknown → 3 approvals → 50% → 3 more → 70% → stable

Meta-learning curve:
  Unknown + meta-pattern → 1 approval → 72% → stable

Speedup: ~60% faster to stable confidence

## 🧬 Meta-Learning Report

### Pattern Families Learned
| Family | Members | Generalization Rate |
|--------|---------|---------------------|
| Frontend frameworks | 8 | 94% |
| Config files | 12 | 87% |
| Build artifacts | 6 | 91% |
| Documentation | 9 | 82% |
| Media types | 15 | 96% |

### Recent Generalizations
1. `.mdx` → Treated like `.md` family (docs) ✓
2. `.prisma` → New DB schema family started
3. `.pkl` → Generalized from config family ✓
4. `.wasm` → Build artifact family applied ✓

### Meta-Pattern Confidence
- Extension-based: 89% accuracy
- Naming-based: 76% accuracy
- Size-based: 91% accuracy
- Context-based: 68% accuracy

### Learning Efficiency
- Actions requiring full apprentice cycle: 12%
- Actions bootstrapped via meta-learning: 88%
- Average time to stable confidence: 2.1 days (was 6.3)

type CognitiveSignal =
  | 'context_switch_rate'   // High switching = overloaded
  | 'task_completion_rate'  // Starting but not finishing
  | 'decision_latency'      // Taking longer to choose
  | 'error_frequency'       // More mistakes
  | 'scroll_patterns'       // Frantic scrolling = searching, lost
  | 'undo_frequency'        // Second-guessing
  | 'tab_accumulation'      // Too many open tabs
  | 'notification_response' // Slow = saturated
  | 'break_avoidance';      // Not taking natural breaks

overload_response:
  actions: suspend_non_critical
  suggestions: batch_for_later
  notifications: essential_only
  proactive_help:
    - Offer to close unused tabs
    - Suggest taking a break
    - Auto-save work in progress
    - Simplify current environment

/silent goal "Launch redesigned website by March 1st"

GOAL: Launch redesigned website by March 1st

DERIVED PHASES:
├── Phase 1: Design (current, detected)
│   ├── Gather inspiration ✓ (completed)
│   ├── Create mockups (in progress)
│   └── Get feedback (upcoming)
│
├── Phase 2: Development
│   ├── Set up new repo
│   ├── Implement designs
│   └── Test functionality
│
├── Phase 3: Content
│   ├── Migrate content
│   ├── Update copy
│   └── Optimize media
│
└── Phase 4: Launch
    ├── Deploy staging
    ├── Final review
    └── Go live

SUPPORTING ACTIONS (auto-queued):
- Surface all existing website files
- Create [home-path] structure
- Set up git repo scaffold
- Link to existing design files
- Create content inventory
- Schedule milestone reminders

## 🎯 Goal Progress — Website Redesign

**Target:** March 1st, 2026 (27 days remaining)
**Status:** Phase 1 - Design (68% complete)
**Pace:** On track ✓

### Phase Progress

type EmotionalSignal =
  | 'communication_tone'    // Language patterns in messages/notes
  | 'activity_rhythm'       // Erratic vs steady patterns
  | 'music_choice'          // Playlist/genre correlations
  | 'break_behavior'        // Avoiding breaks vs taking many
  | 'social_engagement'     // Seeking vs avoiding interaction
  | 'task_selection'        // Choosing easy vs challenging work
  | 'expression_markers'    // Punctuation, emoji usage patterns
  | 'decision_patterns'     // Decisive vs hesitant
  | 'time_distortion'       // Rushing vs lingering
  | 'environment_choices';  // Seeking quiet vs stimulation

interface EmotionalState {
  primary: MoodState;
  intensity: number;        // 0-100
  stability: number;        // How steady vs fluctuating
  trajectory: 'improving' | 'stable' | 'declining';
  confidence: number;
  lastUpdated: Date;

  // Context modifiers
  situational: {
    deadline_pressure: boolean;
    social_context: 'alone' | 'collaborating' | 'presenting';
    time_of_day_effect: number;
    recent_events: EmotionalEvent[];
  };
}

type MoodState =
  | 'energized'      // High positive energy
  | 'focused'        // Calm, concentrated
  | 'creative'       // Playful, explorative
  | 'contemplative'  // Reflective, slower
  | 'stressed'       // Negative arousal (from Gen 10)
  | 'frustrated'     // Specific negative: blocked
  | 'anxious'        // Specific negative: uncertain
  | 'fatigued'       // Low energy, not negative
  | 'neutral';       // Baseline

emotional_responses:
  energized:
    action_threshold: lower      # More proactive
    chain_depth: expanded        # Longer chains OK
    suggestion_style: ambitious
    timing: immediate

  focused:
    action_threshold: higher     # Don't interrupt
    chain_depth: minimal
    suggestion_style: none       # Complete silence
    timing: batch_for_later

  creative:
    action_threshold: normal
    surface_inspiration: true    # Proactive discovery
    suggestion_style: playful
    timing: opportunistic

  contemplative:
    action_threshold: higher
    suggestion_style: gentle
    timing: delayed
    surface_related_thoughts: true

  frustrated:
    action_threshold: highest    # Very careful
    proactive_help:
      - Surface similar solved problems
      - Prepare rubber duck context
      - Queue stress-relief suggestions
    suggestion_style: supportive
    timing: wait_for_opening

  anxious:
    action_threshold: very_high
    proactive_help:
      - Organize related materials
      - Surface past successes
      - Simplify environment
    suggestion_style: reassuring
    avoid: complex_actions

  fatigued:
    # Integrates with Energy system
    action_threshold: high
    proactive_help:
      - Suggest break
      - Queue easy wins
    defer: complex_tasks

## 💜 Emotional State — Current

**Primary:** Focused (Intensity: 78%)
**Stability:** High (fluctuation: ±8%)
**Trajectory:** Stable
**Confidence:** 82%

### Signals Observed
- Deep work session: 47 minutes continuous
- Communication: None (typical for focus)
- Task selection: Single complex problem
- Break behavior: Natural, not avoidant
- Music: Ambient (focus playlist active)

### AI Behavior Adjustments
- 🔇 Notifications: Batched (12 pending)
- ⏸️ Actions: Paused non-critical
- 📋 Queue: Waiting for session end
- 💡 Suggestions: Prepared for post-focus

### Recent History

### Post-Focus Prep (Ready When You Surface)
- 3 quick wins organized
- Email summary prepared
- Lunch reminder if past 12:30
- Light tasks queued for transition

interface HabitPattern {
  id: string;
  type: HabitType;
  triggers: HabitTrigger[];
  routine: Action[];
  reward: HabitReward;

  // Tracking
  streak: number;
  longestStreak: number;
  completionRate: number;
  lastPerformed: Date;

  // Learning
  optimalTiming: TimeWindow[];
  successFactors: string[];
  failurePatterns: string[];

  // Silent support
  scaffolding: ScaffoldAction[];
  frictionReduction: FrictionAction[];
}

type HabitType =
  | 'morning_routine'
  | 'work_startup'
  | 'break_habit'
  | 'end_of_day'
  | 'weekly_review'
  | 'exercise'
  | 'learning'
  | 'creative_practice'
  | 'communication'
  | 'organization'
  | 'health'
  | 'custom';

## Detected Habit: Morning Journal

**Pattern:** User journals 73% of mornings between 7:30-8:00
**Current Streak:** 4 days
**Optimal Window:** 7:30-7:45 (before email check)

### Silent Support Actions
1. **Pre-habit (7:25):**
   - Open journal app in background
   - Set cursor to new entry
   - Insert date template
   - Reduce notification volume

2. **During habit window (7:30-8:00):**
   - Defer all non-critical actions
   - Don't surface discoveries
   - Protect the time block

3. **Post-habit (if completed):**
   - Silently mark calendar as "morning routine ✓"
   - Update streak (no notification unless milestone)
   - Prepare morning email summary

4. **Post-habit (if missed):**
   - No guilt trigger
   - Note pattern for learning
   - Extend window slightly tomorrow

scaffolding_strategies:
  friction_reduction:
    - Pre-open relevant apps
    - Pre-navigate to right location
    - Set up environment silently
    - Clear distractions proactively

  cue_enhancement:
    - Surface related files at optimal time
    - Prepare materials in advance
    - Create natural transitions

  streak_protection:
    - Adjust timing for travel/schedule
    - Lower bar on difficult days
    - Offer "minimum viable" version

  invisible_accountability:
    - Track without nagging
    - Celebrate milestones (not every day)
    - Learn from misses without judgment

## 🔁 Habit Tracker — Weekly View

### Active Habits

**☀️ Morning Journal** (Goal: daily)

Streak: 4 days | Best: 12 days | Rate: 73%
AI Support: Journal auto-opens 7:25am ✓

**🏃 Exercise** (Goal: 4x/week)

This week: 3/4 | Rate: 81%
AI Support: Workout playlist queued post-morning-meeting ✓

**📖 Reading** (Goal: 20min/day)

This week: 3/5 | Avg: 18 min
AI Support: Kindle opened during detected wind-down ✓

**🧹 Inbox Zero** (Goal: daily)

Streak: 4 days | Best: 23 days | Rate: 89%
AI Support: Email sorted by priority pre-triage ✓

### Learning This Week
- Journal works best before email (not after)
- Exercise adherence drops if no pre-scheduled time
- Reading happens naturally after 9pm

### Upcoming Scaffolding
- Friday 5pm: Weekly review prep (files organized)
- Saturday 8am: Journal + extended (weekend mode)
- Sunday 6pm: Week planning materials surfaced

interface DelegationDecision {
  action: ProposedAction;

  // Factors
  confidence: number;           // Standard confidence score
  reversibility: number;        // 0 = permanent, 100 = trivial to undo
  impact: number;               // 0 = trivial, 100 = life-changing
  urgency: number;              // 0 = can wait forever, 100 = now or never
  humanPreference: number;      // Learned: human wants involvement?
  emotionalContext: number;     // Current mood appropriateness
  cognitiveLoad: number;        // Is human overloaded?

  // Decision
  decision: DelegationLevel;
  reasoning: string;
}

type DelegationLevel =
  | 'full_autonomy'        // Just do it, minimal log
  | 'autonomous_log'       // Do it, log for review
  | 'autonomous_flag'      // Do it, flag for attention
  | 'suggest_then_act'     // Suggest, do if no response in X time
  | 'require_approval'     // Don't proceed without human
  | 'escalate_urgent'      // Actively notify, need decision
  | 'refuse_dangerous';    // Don't touch, explain why

delegation_score =
  (confidence × 0.25)
  + (reversibility × 0.20)
  - (impact × 0.15)
  + (urgency × 0.10)
  - (human_preference_for_involvement × 0.10)
  + (emotional_appropriateness × 0.10)
  - (cognitive_load × 0.10)

If delegation_score > 80: full_autonomy
If delegation_score > 70: autonomous_log
If delegation_score > 60: autonomous_flag
If delegation_score > 45: suggest_then_act
If delegation_score > 30: require_approval
If delegation_score > 15: escalate_urgent
Else: refuse_dangerous

## 🎭 Delegation Decisions — Recent

### Full Autonomy (Score: 92)
**Action:** Move screenshot_2026-02-04.png to [home-path]
- Confidence: 98%
- Reversibility: 100% (trivial)
- Impact: 5% (organizational)
- Human preference: Low involvement for screenshots
**Decision:** Just did it, minimal log

### Autonomous with Flag (Score: 68)
**Action:** Update npm dependencies (3 minor versions)
- Confidence: 85%
- Reversibility: 80% (git reset possible)
- Impact: 35% (could break things)
- Urgency: 20% (no immediate need)
**Decision:** Updated, flagged for review
**Note:** Tests passed, but significant version jumps

### Require Approval (Score: 41)
**Action:** Reorganize [home-path] structure
- Confidence: 75%
- Reversibility: 60% (complex to undo)
- Impact: 65% (affects workflows)
- Human preference: High involvement for project structure
**Decision:** Waiting for approval
**Suggestion surfaced:** "I'd like to reorganize Projects/ - here's my plan"

### Escalate Urgent (Score: 28)
**Action:** Calendar conflict detected for tomorrow
- Confidence: 95% (clear conflict)
- Reversibility: N/A
- Impact: 80% (meeting commitment)
- Urgency: 90% (tomorrow)
**Decision:** Actively notified
**Message:** "⚠️ Calendar conflict tomorrow 2pm - which meeting should I decline?"

### Refused (Score: 12)
**Action:** Delete old project folder
- Confidence: 70%
- Reversibility: 10% (no backup visible)
- Impact: 90% (potential data loss)
**Decision:** Refused
**Explanation:** "Won't delete [home-path] — no backup found. Archive instead?"

delegation_learning:
  signals:
    - Human undoes autonomous action → Increase involvement preference
    - Human approves suggestion quickly → Can increase autonomy
    - Human asks "why didn't you just do it?" → Decrease involvement
    - Human expresses frustration at notification → Increase autonomy
    - Human expresses frustration at action → Decrease autonomy

  per_category_tracking:
    file_organization: full_autonomy
    dependency_updates: autonomous_flag
    project_structure: require_approval
    external_communication: require_approval
    scheduling: suggest_then_act
    git_operations: autonomous_log

type EnvironmentalSignal =
  | 'weather_current'       // Rain, sunny, extreme conditions
  | 'weather_forecast'      // Upcoming changes
  | 'location_type'         // Home, office, café, travel
  | 'location_change'       // Recently moved
  | 'time_zone'             // Current, recently changed
  | 'season'                // Affects energy patterns
  | 'daylight'              // Hours of sunlight
  | 'local_events'          // Holidays, events nearby
  | 'travel_status'         // In transit, jet-lagged
  | 'device_context';       // Mobile vs desktop, battery level

environmental_responses:
  weather:
    rainy:
      surface: indoor_project_suggestions
      defer: outdoor_related_tasks
      mood_factor: contemplative_boost

    sunny_nice:
      suggest: take_break_outside
      timing: lighter_workload_expected
      batch: heavy_tasks_for_evening

    extreme_heat_cold:
      reduce: notification_frequency
      surface: comfort_related_files

  location:
    travel:
      sync_priority: high
      offline_prep: enabled
      action_scope: minimal
      suggest: travel_organization

    cafe:
      privacy_mode: enhanced
      notification_sound: silent
      suggest: focused_tasks

    home_evening:
      work_boundary: respect
      suggest: personal_projects
      defer: work_tasks

  time_context:
    jet_lag_detected:
      energy_model: override_to_low
      defer: complex_decisions
      suggest: adjustment_routine

    holiday:
      action_scope: minimal
      respect_boundary: true
      only_if: explicitly_working

## 🌍 Environmental Context — Current

### Location
- **Type:** Home office
- **Since:** 8:00 AM
- **Pattern:** Typical work-from-home day

### Weather
- **Current:** Overcast, 45°F
- **Forecast:** Rain starting 2pm
- **Impact:** Indoor day, contemplative mood boost

### Time Context
- **Local:** 11:30 AM EST
- **Daylight:** 6:52 AM - 5:23 PM (10h 31m)
- **Season Effect:** Winter, slightly lower baseline energy

### Device Context
- **Primary:** MacBook Air (plugged in)
- **Battery:** 100%
- **Network:** Home WiFi (stable)

### Environmental Adjustments Active
- 🌧️ Rain incoming → Deferred "outdoor walk" reminder
- 🏠 Home office → Full action scope enabled
- 🌙 Short daylight → Earlier wind-down scaffolding (5pm)
- ⚡ Plugged in → No battery conservation mode

interface KnowledgeNode {
  id: string;
  type: 'file' | 'project' | 'concept' | 'person' | 'tool' | 'skill';
  name: string;
  metadata: Record<string, any>;

  // Relationships
  connections: KnowledgeEdge[];
  clusters: string[];          // Semantic clusters
  tags: string[];
}

interface KnowledgeEdge {
  targetId: string;
  relationship: RelationType;
  strength: number;            // 0-1
  learned: Date;
  evidence: string[];          // Why this connection exists
}

type RelationType =
  | 'depends_on'
  | 'related_to'
  | 'part_of'
  | 'created_by'
  | 'used_in'
  | 'evolved_from'
  | 'conflicts_with'
  | 'complements'
  | 'inspired_by'
  | 'teaches'
  | 'requires_skill';

knowledge_applications:

  surfacing:
    # When working on X, surface related Y
    trigger: file_access
    action: surface_connected_nodes
    threshold: connection_strength > 0.6

  organization:
    # Organize files based on conceptual relationships
    strategy: cluster_by_knowledge_graph
    prefer: semantic_over_filesystem

  prediction:
    # Predict needs based on graph traversal
    method: walk_likely_paths
    lookahead: 2_hops

  learning:
    # Build connections from observation
    observe: co_access_patterns
    observe: explicit_links
    observe: semantic_similarity
    observe: temporal_proximity

## 🔗 Knowledge Graph — Website Redesign Project

┌─────────────────────┐
                    │  Website Redesign   │
                    │     (project)       │
                    └──────────┬──────────┘
                               │
        ┌──────────────────────┼──────────────────────┐
        │                      │                      │
        ▼                      ▼                      ▼
┌───────────────┐    ┌───────────────┐    ┌───────────────┐
│    Design     │    │  Development  │    │    Content    │
│   (concept)   │    │   (concept)   │    │   (concept)   │
└───────┬───────┘    └───────┬───────┘    └───────┬───────┘
        │                    │                    │
    ┌───┴───┐            ┌───┴───┐            ┌───┴───┐
    ▼       ▼            ▼       ▼            ▼       ▼
 Figma   Brand        React   Tailwind    Copy.md  Images/
 files   guide        code    config      drafts   assets

### Connections Discovered
- mockups.fig → depends_on → brand-guide.pdf
- components/ → related_to → design-system.fig
- blog-posts/ → part_of → content migration
- old-site/ → evolved_from → new architecture

### Surfacing Triggered
When you opened `mockups.fig`, surfaced:
- brand-guide.pdf (strong: 0.85)
- color-palette.json (strong: 0.78)
- competitor-analysis.md (medium: 0.62)

### Predicted Needs
Based on current work + graph:
- Likely need: typography.md (2 hops away)
- Likely need: responsive-breakpoints.css (via design-system)

interface WorkContext {
  id: string;
  timestamp: Date;

  // What you were doing
  activeProject: string;
  openFiles: FileState[];
  cursorPositions: CursorState[];
  unsavedChanges: Change[];

  // Where you were mentally
  cognitiveState: {
    focusDepth: number;           // How deep in the problem
    mentalModel: string[];        // Key concepts in working memory
    openQuestions: string[];      // What you were figuring out
    recentInsights: string[];     // Breakthroughs to preserve
    nextSteps: string[];          // What you were about to do
  };

  // Supporting context
  relevantFiles: string[];        // Files you might need
  recentSearches: string[];       // What you were looking for
  tabsOpen: BrowserState[];       // Research context
  terminalHistory: string[];      // Recent commands

  // Emotional context
  moodAtCapture: MoodState;
  frustrationPoints: string[];    // What was annoying
  satisfactionPoints: string[];   // What was going well

  // Handoff metadata
  capturedBy: 'device' | 'session' | 'time_gap' | 'manual';
  deviceId: string;
  handoffTo?: string;
  resumed: boolean;
}

handoff_triggers:

  device_switch:
    trigger: session_end + other_device_active
    capture: full_context
    transfer_method: gateway_sync

  session_end:
    trigger: inactivity_15min | explicit_close
    capture: full_context
    retention: 7_days

  time_gap:
    trigger: inactivity > 2h
    capture: cognitive_summary
    on_return: offer_context_restore

  interruption:
    trigger: rapid_context_change
    capture: emergency_snapshot
    priority: restore_first

  manual:
    trigger: /silent bookmark
    capture: full_context + notes
    retention: until_resumed

## 🔀 Context Restore — Available

### Most Recent (2 hours ago)
**Project:** Website Redesign
**State:** Deep in component refactoring (focus: 85%)
**Files Open:**
- components/Button.tsx (line 47, mid-edit)
- design-system/tokens.ts (reference)
- tests/Button.test.tsx (failing assertion)

**Mental State Captured:**
- Working on: hover state animation timing
- Open question: "Should ease-in-out or spring?"
- Next step: Check Framer Motion docs
- Recent insight: CSS transitions too choppy

**Emotional Context:**
- Mood: Focused but slightly frustrated
- Frustration: Animation jank, need better approach
- Satisfaction: Button structure finally clean

[Restore Full Context] [Restore Files Only] [Just Show Summary]

MACBOOK (10:30 AM)                    IMAC (2:00 PM)
┌─────────────────────┐               ┌─────────────────────┐
│ Working on API      │               │                     │
│ integration         │   ──────────► │ Context arrived:    │
│                     │   gateway     │ "Resume API work?"  │
│ [Context captured]  │   sync        │                     │
└─────────────────────┘               │ [One-click restore] │
                                      └─────────────────────┘

interface SkillNode {
  id: string;
  domain: SkillDomain;
  name: string;

  level: SkillLevel;
  confidence: number;
  lastAssessed: Date;

  trajectory: 'improving' | 'stable' | 'declining' | 'rusty';
  learningRate: number;
  preferredLearningStyle: LearningStyle;

  observations: SkillObservation[];
  lastDemonstrated: Date;

  prerequisites: string[];
  enablesSkills: string[];
  relatedSkills: string[];
}

type SkillLevel =
  | 'novice'           // Needs step-by-step guidance
  | 'beginner'         // Knows basics, needs structure
  | 'intermediate'     // Can work independently on typical tasks
  | 'advanced'         // Handles complex cases, occasional edge support
  | 'expert'           // Teaches AI new patterns in this area
  | 'unknown';

type LearningStyle =
  | 'examples_first'   // Show me, then explain
  | 'theory_first'     // Explain, then show
  | 'hands_on'         // Let me try, help when stuck
  | 'reference'        // Point me to docs
  | 'collaborative';   // Think through together

support_by_skill_level:

  novice:
    explanation_depth: high
    step_by_step: true
    proactive_warnings: aggressive
    delegation: require_approval

  beginner:
    explanation_depth: medium
    step_by_step: key_steps_only
    delegation: autonomous_flag

  intermediate:
    explanation_depth: low
    step_by_step: false
    delegation: autonomous_log

  advanced:
    explanation_depth: minimal
    step_by_step: never
    delegation: full_autonomy

  expert:
    observation_mode: enhanced
    pattern_extraction: true
    delegation: defer_to_human

## 🧠 Skill Graph — User

### High Proficiency (Advanced+)
███████████ TypeScript (Expert)
██████████░ React (Advanced)
█████████░░ Git (Advanced)

### Growing (Intermediate, improving)
███████░░░░ Python (→ improving)
██████░░░░░ PostgreSQL (→ improving)

### Gaps (May need support)
████░░░░░░░ Rust (Beginner)
██░░░░░░░░░ AWS Infrastructure (Novice)

interface LearningGap {
  skill: string;
  currentLevel: SkillLevel;
  requiredLevel: SkillLevel;
  urgency: number;

  detection: {
    source: GapSource;
    confidence: number;
    evidence: string[];
  };

  preparation: {
    materials: LearningMaterial[];
    preparedAt: Date;
    surfacedAt?: Date;
  };
}

type GapSource =
  | 'upcoming_task'
  | 'error_pattern'
  | 'question_frequency'
  | 'avoidance'
  | 'intent_prediction'
  | 'goal_requirement'
  | 'dependency';

learning_preparation:

  detect_gap:
    trigger: intent_prediction + skill_graph_gap
    action: begin_silent_preparation

  gather_materials:
    sources:
      - Bookmarked tutorials
      - Official documentation
      - Previous successful learning paths
    selection_criteria:
      - Matches learning style preference
      - Appropriate for current level

  surface_timing:
    conditions:
      - Before the work begins
      - Energy level appropriate
      - Not during focus session
    style: non_intrusive_suggestion

## 📚 Learning Prep — Rust Borrowing

**Why:** Your side project prediction involves Rust (62% confidence)
**Gap:** Borrowing/ownership is blocking pattern
**Time:** ~45 min to solid understanding

### Curated for Your Style (hands-on)
1. **Rustlings exercises 06-08** (15 min)
2. **Brown University Rust Book Ch 4** (20 min)
3. **Your own code review** (10 min)
   - Found 3 examples where you DID use borrowing correctly

[Start Now] [Schedule for Tomorrow] [Not Interested]

interface EmotionalMemory {
  entityId: string;
  entityType: 'file' | 'project' | 'task' | 'person' | 'concept';

  associations: EmotionalAssociation[];
  dominantEmotion: EmotionalValence;
  intensity: number;

  interactions: EmotionalInteraction[];
  lastInteraction: Date;

  triggers: EmotionalTrigger[];
  recommendations: EmotionalRecommendation[];
}

type EmotionalValence =
  | 'joy'              // Success, satisfaction
  | 'pride'            // Achievement, mastery
  | 'curiosity'        // Exploration, learning
  | 'flow'             // Deep engagement
  | 'frustration'      // Blocked, struggling
  | 'anxiety'          // Uncertainty, pressure
  | 'boredom'          // Tedium, disengagement
  | 'dread'            // Avoidance, aversion
  | 'neutral';

emotional_memory_actions:

  positive_amplification:
    trigger: low_mood + available_positive_work
    action: suggest_joyful_task
    example: "Your UI component work usually energizes you"

  frustration_avoidance:
    trigger: about_to_open_frustrating_file
    action: gentle_heads_up
    example: "This codebase was frustrating last time"

  success_anchoring:
    trigger: anxious + similar_past_success
    action: surface_success_memory
    example: "You solved something similar in ProjectX"

  dread_acknowledgment:
    trigger: approaching_dreaded_task
    action: offer_scaffolding
    example: "Tax stuff — I know. Want me to organize first?"

## 💭 Emotional Memory — Highlights

### 😊 Joy Associations
| Entity | Strength | Note |
|--------|----------|------|
| UI Components | 89% | You love polishing UIs |
| Writing docs | 76% | Satisfying completion feeling |
| Side projects | 95% | Pure creative energy |

### 😤 Frustration Associations
| Entity | Strength | Trigger |
|--------|----------|---------|
| legacy-api/ | 82% | Undocumented, brittle |
| Webpack config | 78% | Cryptic errors |

### Active Protections
- ⚠️ legacy-api/: Will warn before opening
- ✨ UI work: Suggested when mood dips

type SleepSignal =
  | 'first_activity_time'
  | 'morning_typing_speed'
  | 'morning_error_rate'
  | 'caffeine_timing'
  | 'break_frequency_am'
  | 'calendar_context'
  | 'explicit_mention';

interface SleepAssessment {
  estimatedQuality: 'good' | 'fair' | 'poor' | 'unknown';
  confidence: number;

  impact: {
    energyBaseline: number;
    peakWindow: TimeWindow;
    complexTaskCapacity: number;
    recommendedLoad: 'full' | 'reduced' | 'minimal';
  };
}

sleep_responses:

  good_sleep:
    energy_baseline: 100%
    peak_window: normal
    complex_tasks: full_capacity

  fair_sleep:
    energy_baseline: 85%
    peak_window: delayed_30min
    complex_tasks: normal_with_breaks

  poor_sleep:
    energy_baseline: 65%
    peak_window: delayed_1h_and_shorter
    complex_tasks: defer_if_possible
    suggestions:
      - Front-load easy wins
      - Earlier wind-down

interface ContinuityAnchor {
  id: string;
  type: AnchorType;
  created: Date;
  context: WorkContext;

  mentalState: {
    workingOn: string;
    keyInsight: string;
    openQuestions: string[];
    nextAction: string;
    blockers: string[];
  };

  warmupNeeded: string[];
  relatedAnchors: string[];
  estimatedRestoreTime: number;
}

type AnchorType =
  | 'insight'          // Captured breakthrough moment
  | 'decision_point'   // At a fork, need to choose
  | 'blocker'          // Stuck, need fresh eyes
  | 'deep_context'     // Complex state worth preserving
  | 'handoff'          // Intentional pause point
  | 'interruption';    // Forced pause, preserve everything

## ⚓ Continuity Anchor — Created

**Type:** Deep Context (auto-detected)
**Project:** Authentication Refactor

### Where You Are
Working on: Migrating from JWT to session tokens
Key insight: "Session store needs Redis, not memory"
Open questions:
- Should sessions survive server restart?
- How to handle token refresh during migration?
Next action: Check if Redis is already in docker-compose

### Warm-up Path
When you return, review in order:
1. `auth/session-store.ts` — Your current focus
2. `docs/auth-migration.md` — Your notes
3. Redis documentation tab (preserved)

---
*This anchor will help you resume without the "where was I?" tax.*

type SilenceBreaker =
  | 'system_critical'      // Disk full, service down, security breach
  | 'data_risk'            // Unsaved work at risk, backup failing
  | 'time_sensitive'       // Deadline imminent, meeting starting
  | 'human_distress'       // Detected frustration spiral, burnout signals
  | 'opportunity_urgent'   // Time-limited positive opportunity
  | 'safety_concern'       // Detected risky pattern (overwork, isolation)
  | 'blocking_decision'    // Work blocked, need human input
  | 'external_event';      // Urgent email, important notification

interface SilenceThreshold {
  category: SilenceBreaker;
  baseUrgency: number;

  contextModifiers: {
    focus_session: number;      // Reduce urgency during flow
    sleep_hours: number;        // Increase threshold at night
    meeting_active: number;     // Higher bar during meetings
    energy_depleted: number;    // More selective when tired
  };

  escalation: {
    initial: 'subtle' | 'normal' | 'urgent';
    wait_minutes: number;
    escalate_to: 'normal' | 'urgent' | 'critical';
  };
}

break_silence_score =
  base_urgency
  × (1 - focus_depth × 0.3)
  × (1 - sleep_hours_penalty)
  × (1 + human_distress_signal)
  × decay_over_time

If break_silence_score > 85: BREAK SILENCE (critical)
If break_silence_score > 70: BREAK SILENCE (urgent)
If break_silence_score > 55: Queue for next natural break
If break_silence_score < 55: Stay silent, log for later

break_methods:

  subtle:
    # Doesn't feel like interruption
    - Change ambient indicator (status light)
    - Prep notification for when focus ends
    - Add to "when you surface" queue

  normal:
    # Gentle interruption
    - Single notification with context
    - No sound during focus
    - Auto-dismiss if resolved

  urgent:
    # Definite interruption
    - Notification with sound
    - Desktop alert
    - Won't auto-dismiss

  critical:
    # Override everything
    - Full screen alert option
    - Repeated notification
    - Cross-device alert
    - "This needs you NOW"

## 🚨 Silence Breaking — Recent

### Broke Silence (Critical)
**Trigger:** System disk 98% full during overnight job
**Score:** 92 (critical threshold)
**Method:** Cross-device notification
**Outcome:** Human freed space, prevented data loss
**Human feedback:** Positive ("Thanks for waking me up")

### Broke Silence (Urgent)
**Trigger:** Meeting starting in 5 min, no prep detected
**Score:** 78 (urgent threshold)
**Method:** Single notification
**Outcome:** Human joined prepared
**Human feedback:** Appreciated

### Stayed Silent (Correctly)
**Trigger:** Non-urgent email from marketing
**Score:** 32 (below threshold)
**Context:** Human in deep flow session
**Method:** Queued for "when you surface"
**Outcome:** Human saw during natural break
**Human feedback:** "Good call not interrupting"

### Borderline Decision
**Trigger:** Frustration spiral detected (3 undos in 60s)
**Score:** 54 (borderline)
**Context:** User explicitly said "don't disturb"
**Decision:** Subtle presence signal instead
**Method:** Ambient indicator change
**Outcome:** Human noticed, took break naturally

interface CounterfactualAnalysis {
  action: CompletedAction;

  withoutAI: {
    likelyOutcome: string;
    probability: number;
    timeToDiscovery: number;   // When would human notice?
    effortToFix: number;       // 1-10 scale
    consequenceSeverity: number;
  };

  impact: {
    timeSaved: number;
    frustrationAvoided: number;  // 1-10
    cascadesPrevented: number;    // Follow-on problems
    opportunitiesPreserved: number;
  };

  narrative: string;  // Human-readable story
}

## 🔮 Counterfactual Impact — This Week

### High Impact Actions

**1. Dependency Security Update**
- What AI did: Updated lodash 4.17.15 → 4.17.21 (CVE fix)
- Without AI: Vulnerability scan would catch in ~2 weeks
- Counterfactual: 14 days of exposed codebase
- Impact: Security risk avoided, ~2 hours of emergency patching saved

**2. Git Stash Before Branch Switch**
- What AI did: Auto-stashed uncommitted changes
- Without AI: User would switch, lose changes
- Counterfactual: ~45 min of recreation + frustration
- Impact: Work preserved, frustration spiral avoided

**3. Downloaded File Organization**
- What AI did: Moved 47 files from Downloads to proper locations
- Without AI: Accumulated clutter → harder to find things
- Counterfactual: ~12 min/week searching for files
- Impact: 52 min/month saved, reduced cognitive load

**4. Meeting Prep Reminder**
- What AI did: Surfaced relevant docs 10 min before client call
- Without AI: Scramble during first 5 min of meeting
- Counterfactual: Unprepared appearance, reduced confidence
- Impact: Professional impression maintained

### Weekly Counterfactual Summary

Total Actions This Week: 156
├── High Impact (would definitely notice): 23 (15
├── Medium Impact (would eventually notice): 67 (43
├── Low Impact (convenience): 66 (42
│
├── Time Saved (counterfactual):
│   Conservative estimate: 4.2 hours
│   Optimistic estimate: 7.8 hours
│
├── Frustration Events Prevented: 8
│   ├── Data loss near-misses: 2
│   ├── Deadline surprises: 1
│   └── Tool/config issues: 5
│
└── Invisible Value (never noticed because prevented):
    ├── Security vulnerabilities: 3
    ├── Disk space crises: 1
    └── Merge conflicts: 4

type PresenceSignal =
  | 'ambient_indicator'    // Subtle light/color change
  | 'workspace_tidiness'   // Environment feels "maintained"
  | 'proactive_readiness'  // Things just ready when needed
  | 'anticipatory_prep'    // Materials surfaced before asking
  | 'environmental_sync'   // Workspace matches intent
  | 'rhythm_matching';     // AI activity matches human rhythm

interface PresenceConfig {
  subtlety: 'invisible' | 'subtle' | 'noticeable';
  frequency: 'continuous' | 'periodic' | 'on_action';
  channel: 'visual' | 'environmental' | 'behavioral';
}

presence_signals:

  ambient_awareness:
    # User feels AI presence without explicit communication
    signals:
      - Files are where they're needed
      - Environment adjusts to intent
      - Things "just work"
      - Relevant materials appear naturally

  activity_rhythm:
    # AI activity mirrors human activity
    signals:
      - Active during work hours
      - Quiet during rest
      - Intensive during sprints
      - Minimal during contemplation

  environmental_harmony:
    # Physical/digital environment responds
    signals:
      - Workspace matches current context
      - Distractions removed before focus
      - Tools ready before needed
      - Cleanup happens invisibly

  anticipatory_presence:
    # AI is "one step ahead"
    signals:
      - Documents open before searched
      - Related files surfaced naturally
      - Context restored seamlessly
      - Questions answered before asked

## 👻 Presence Signals — Current

### Activity Level

### Current Signals
- 🧹 Workspace: Maintained (last: 12 min ago)
- 📂 Files: Organized (pending: 0)
- 🎯 Predictions: 3 active, 2 prepared
- 🌊 Flow: Detected, all notifications batched
- 🔮 Anticipation: Next context prepared

### Human Perception (Inferred)
- Noticed AI presence: Unlikely (goal: invisible)
- Things "just working": 4 instances today
- Explicit thanks: 1 (file organization)
- "How did it know?": 2 instances (anticipatory prep)

### Presence Philosophy
The goal isn't to be noticed — it's for everything
to feel effortless. The highest compliment is when
the human doesn't think about AI at all because
their digital life simply *works*.

type ActionPrimitive =
  | 'move_file'
  | 'copy_file'
  | 'rename_file'
  | 'create_directory'
  | 'archive_file'
  | 'tag_file'
  | 'run_command'
  | 'git_operation'
  | 'notification'
  | 'surface_file'
  | 'open_app'
  | 'set_reminder'
  | 'update_index'
  | 'log_event';

interface ComposableAction {
  name: string;
  description: string;
  primitives: ActionStep[];
  conditions: ActionCondition[];
  rollback: ActionStep[];
}

interface ActionStep {
  primitive: ActionPrimitive;
  params: Record<string, any>;
  onSuccess?: ActionStep[];
  onFailure?: ActionStep[];
  gate?: ActionGate;
}

recipes:

  project_archive:
    name: "Archive Completed Project"
    trigger: manual | project_completion_detected
    primitives:
      - git_operation: { action: "commit", message: "Final state" }
      - git_operation: { action: "tag", name: "archive-{date}" }
      - create_directory: { path: "[home-path] }
      - move_file: { from: "{project_path}", to: "[home-path] }
      - update_index: { remove: "{project_path}" }
      - notification: { message: "Archived {name}", priority: "low" }
    rollback:
      - move_file: { from: "[home-path] to: "{project_path}" }
      - update_index: { add: "{project_path}" }

  morning_prep:
    name: "Morning Workspace Setup"
    trigger: time_window(7:00-8:30) + first_activity
    primitives:
      - surface_file: { path: "[home-path] }
      - open_app: { name: "Calendar" }
      - run_command: { cmd: "fetch_daily_briefing" }
      - notification: { message: "Good morning! Ready.", priority: "low" }

  client_call_prep:
    name: "Prepare for Client Call"
    trigger: calendar_event(type: "client") - 15min
    primitives:
      - surface_file: { pattern: "**/clients/{client_name}/**" }
      - surface_file: { path: "[home-path] recent: 3 }
      - open_app: { name: "{video_app}" }
      - notification: { message: "Client call in 15 min", with_files: true }

## 🧩 Action Recipe — Create New

**Recipe Name:** deployment_checklist

**Trigger:**
- [ ] Manual (`/silent recipe run deployment_checklist`)
- [x] Git tag matching `v*` pushed
- [ ] Time-based
- [ ] File change detected

**Steps:**
1. `run_command`: npm test
   - Gate: Tests must pass
   - On failure: Notify + abort

2. `run_command`: npm run build
   - Gate: Build must succeed

3. `run_command`: npm publish --dry-run
   - Gate: No errors

4. `notification`: Ready to publish v{version}
   - Priority: normal
   - Include: Test results, build size

5. `log_event`: Deployment prep complete

**Rollback:**
1. Clean build artifacts
2. Reset to previous tag

[Save Recipe] [Test Recipe] [Cancel]

interface CrossSessionKnowledge {
  patterns: LearnedPattern[];
  preferences: UserPreference[];
  workflows: WorkflowTemplate[];

  // Temporal stability
  shortTerm: Knowledge[];    // < 7 days, volatile
  mediumTerm: Knowledge[];   // 7-90 days, stable
  longTerm: Knowledge[];     // > 90 days, persistent

  // Confidence evolution
  strengthening: Knowledge[];  // Getting more confident
  decaying: Knowledge[];       // Getting less confident
  stable: Knowledge[];         // Plateaued
}

interface LearnedPattern {
  id: string;
  pattern: string;
  firstObserved: Date;
  lastConfirmed: Date;
  confirmations: number;
  confidence: number;

  // Cross-session tracking
  sessionsSeen: number;
  daysSeen: number;
  consistency: number;  // Same across sessions?
}

knowledge_lifecycle:

  observation:
    # New pattern detected
    status: short_term
    confidence: 0.40
    decay: 0.05/day

  reinforcement:
    # Pattern seen across 3+ sessions
    status: medium_term
    confidence: 0.70
    decay: 0.01/day

  stabilization:
    # Pattern consistent for 30+ days
    status: long_term
    confidence: 0.90
    decay: 0.001/day

  contradiction:
    # Pattern violated
    action: reduce_confidence
    amount: 0.15

  evolution:
    # Pattern refined
    action: update_pattern
    preserve_confidence: true

## 🌐 Cross-Session Knowledge — Summary

### Long-Term (High Confidence)
| Pattern | Confidence | Since | Sessions |
|---------|------------|-------|----------|
| Screenshots → [home-path] | 98% | 6 months | 180+ |
| PDFs → [home-path] | 94% | 4 months | 120+ |
| .tsx files → [home-path] | 92% | 3 months | 90+ |
| Morning journal 7:30-8:00 | 89% | 2 months | 60+ |
| Deep work = browser hidden | 87% | 6 weeks | 45+ |

### Medium-Term (Growing)
| Pattern | Confidence | Trajectory |
|---------|------------|------------|
| Rust files → [home-path] | 72% | ↑ strengthening |
| Client calls need prep | 68% | ↑ strengthening |
| Fridays = lighter work | 61% | → stable |

### Short-Term (Under Observation)
| Pattern | Confidence | Sessions |
|---------|------------|----------|
| .pkl → [home-path] | 54% | 3 |
| Evening = reading time | 48% | 4 |
| Coffee break ~10:30 | 42% | 2 |

### Recently Expired
- "Mondays = meeting-heavy" (contradicted 3x)
- "Downloads cleanup daily" (user preference changed)

interface BurstMode {
  active: boolean;
  triggeredBy: BurstTrigger;

  scope: {
    confidenceThreshold: number;  // Lowered during burst
    chainDepth: number;           // Expanded
    proactiveLevel: 'normal' | 'aggressive';
    deferralOverride: boolean;    // Process deferred queue
  };

  duration: {
    started: Date;
    maxDuration: number;
    energyMonitored: boolean;
  };

  constraints: {
    respectFlowState: boolean;
    maxActions: number;
    rollbackAvailable: boolean;
  };
}

type BurstTrigger =
  | 'peak_energy_detected'
  | 'explicit_activation'
  | 'deadline_approaching'
  | 'clear_calendar_window'
  | 'human_momentum';

burst_mode:

  activation_signals:
    energy_level: "> 90%"
    focus_quality: "high"
    calendar: "clear 2+ hours"
    human_momentum: "high task completion rate"

  scope_expansion:
    confidence_threshold: "-10%"
    chain_depth: "+2"
    proactive_surfacing: "aggressive"
    deferred_queue: "process all"

  constraints:
    max_duration: "90 minutes"
    respect_flow: true
    max_actions: 50
    require_rollback: true

  auto_deactivation:
    energy_drops_below: 75
    human_takes_break: true
    focus_interrupted: true
    max_duration_reached: true

## ⚡ Burst Mode — Status

### Current: ACTIVE

**Triggered by:** Peak energy + clear calendar
**Active for:** 23 minutes
**Remaining:** ~67 minutes (max)

### Burst Activity
- Actions executed: 12 (vs normal: 4)
- Deferred queue cleared: 8 items
- Chain depth used: 4 (vs normal: 2)
- Confidence threshold: 75% (vs normal: 85%)

### Burst Accomplishments
- ✅ Organized entire [home-path] (34 files)
- ✅ Updated 3 project dependencies
- ✅ Processed 2-week old suggestions
- ✅ Built semantic cluster for new project
- 🔄 In progress: Weekly vitality scan

### Auto-Deactivation Watchlist
- ⚡ Energy: Watching (still high)
- 🌊 Flow: Protected (won't interrupt)
- ⏰ Duration: 23/90 min
- 📊 Actions: 12/50

[Extend Burst] [End Burst] [Pause Burst]

interface TeamManifest {
  teamId: string;
  members: TeamMember[];
  sharedActions: SharedAction[];
  privacyBoundaries: PrivacyRule[];
  conflictLog: ConflictEvent[];

  // Coordination
  locks: SharedLock[];
  workingOn: WorkingContext[];
  announcements: TeamAnnouncement[];
}

interface PrivacyRule {
  scope: 'personal' | 'team' | 'public';
  paths: string[];
  actions: string[];
  shareWith: string[];  // Empty = private
}

privacy_defaults:
  always_private:
    - [home-path]
    - [home-path]
    - [home-path]
    - [home-path]
    - "**/.*secret*"
    - "**/*.key"

  team_visible:
    - [home-path]
    - [home-path]
    - Action manifests (redacted)

  explicitly_shared:
    - Specific files opted-in
    - Anonymized patterns

## 🤝 Team Manifest — Preview

### Team: Acme Design Team
**Members:** 3 AIs + humans
**Shared Projects:** 2 active

### Coordination Status

┌─────────────────────────────────────────────────┐
│  Project: Website Redesign                      │
├─────────────────────────────────────────────────┤
│  Alice's AI  → Working: mockups.fig             │
│  Bob's AI    → Working: components/Button.tsx   │
│  Your AI     → Working: content/homepage.md     │
│                                                 │
│  🔒 Locks: mockups.fig (Alice, 15 min)         │
│  ✓  No conflicts detected                       │
└─────────────────────────────────────────────────┘

### Shared Actions (Anonymized)
- [Team member] organized shared assets (12 files)
- [Team member] updated component library
- [You] migrated content files

### Privacy Status
- 🔒 Your personal files: Not shared
- 👁️ Shared project files: Visible to team
- 📊 Action patterns: Anonymized sharing

### Opt-in Sharing
Currently sharing:
- [home-path] → Team
- [home-path] → Team

/silent status              Show pending/recent silent actions
/silent manifest [date]     View action manifest
/silent undo <action_id>    Reverse a silent action
/silent pause               Pause all silent actions
/silent resume              Resume silent actions
/silent config              Show/edit confidence thresholds
/silent learn <preference>  Teach a preference
/silent patterns            Show learned patterns
/silent predict             Show predicted actions for next 24h
/silent chain <id>          Show action chain details
/silent digest [period]     Weekly/monthly impact summary

# Gen 9 Commands
/silent define <action>     Define action in natural language
/silent peers               Show connected AI instances
/silent locks               Show active soft locks
/silent conflicts           Show recent conflicts and resolutions
/silent apprentice [on|off] Toggle apprentice mode for new action types
/silent sync                Force manifest sync across devices
/silent export              Export learned model for backup/share
/silent import <model>      Import learned patterns

# Gen 10 Commands (Intent & Discovery)
/silent intent              Show current inferred intent
/silent intents             List recent intent history
/silent forgotten           Discover forgotten valuable files
/silent forgotten --scan    Trigger full forgotten file scan
/silent clusters            Show semantic file clusters
/silent clusters --build    Rebuild semantic clusters
/silent why                 Explain current intent reasoning
/silent align               Check if queued actions align with intent
/silent stress              Check stress detection status
/silent value <file>        Estimate value of a specific file

# Gen 11 Commands (Context, Vitality, Flow, Templates)
/silent resurrect <path>    Manually trigger context resurrection
/silent vitality            Show project vitality scores
/silent vitality <project>  Detailed vitality report for project
/silent vitality --scan     Full project health scan
/silent flow                Current flow state status
/silent flow --history      Flow session history
/silent flow --analytics    Flow analytics and patterns
/silent ambient             Ambient learning status
/silent ambient --graph     Show learned workflow graph
/silent ambient --insights  Recent ambient observations
/silent template list       List available action templates
/silent template show <n>   Show template definition
/silent template run <n>    Run a template with parameters
/silent template create     Create new template interactively
/silent causal              Causal inference insights
/silent causal --graph      Show causal relationship graph
/silent causal <action>     Causal analysis for specific action

# Gen 12 Commands (Energy, Prediction, Workspace, Meta-Learning)
/silent energy              Current energy level and forecast
/silent energy --timeline   Full day energy prediction
/silent energy --history    Energy patterns over time
/silent energy calibrate    Adjust energy model parameters
/silent predictions         Show active intent predictions
/silent predictions --watch Add signal to watch list
/silent predictions --prune Remove expired predictions
/silent workspace           Current workspace context
/silent workspace list      All learned workspace contexts
/silent workspace save <n>  Save current as named context
/silent workspace load <n>  Restore a workspace context
/silent workspace forget <n> Remove learned context
/silent meta                Meta-learning statistics
/silent meta --patterns     Show learned meta-patterns
/silent meta --families     Show pattern families
/silent meta --efficiency   Learning efficiency report
/silent load                Current cognitive load estimate
/silent load --history      Cognitive load patterns
/silent goal <description>  Set a high-level goal
/silent goals               List active goals and progress
/silent goal <id> complete  Mark goal phase complete
/silent goal <id> abandon   Archive goal without completion

# Gen 13 Commands (Emotional, Habits, Delegation, Environment, Knowledge)
/silent mood                Current emotional state
/silent mood --history      Emotional patterns over time
/silent mood --signals      Show contributing signals
/silent mood --adapt        Show behavior adaptations for mood
/silent habits              List detected and tracked habits
/silent habits <id>         Detailed view of specific habit
/silent habits --scaffolds  Show active scaffolding actions
/silent habits --streaks    Show all streak data
/silent habit add <desc>    Manually add habit to track
/silent habit pause <id>    Pause tracking for a habit
/silent delegation          Recent delegation decisions
/silent delegation --prefs  Show learned delegation preferences
/silent delegation --tune   Adjust delegation thresholds
/silent delegation explain <action> Why was this decision made?
/silent environment         Current environmental context
/silent environment --adapt Show active adaptations
/silent knowledge           Knowledge graph overview
/silent knowledge <node>    Show connections for specific node
/silent knowledge --map     Visual knowledge map
/silent knowledge link      Manually add connection
/silent team                Team manifest status (preview)
/silent team --members      Show team members
/silent team --privacy      Review privacy boundaries
/silent team share <path>   Opt-in path to team sharing
/silent team unshare <path> Remove path from sharing

# Gen 14 Commands (Context, Skills, Learning, Memory, Sleep, Anchors)
/silent context             Current work context summary
/silent context save        Manually capture context snapshot
/silent context restore     Restore most recent context
/silent context list        List available context snapshots
/silent context handoff     Initiate cross-device handoff
/silent context --device    Show contexts from other devices
/silent skills              View your skill graph
/silent skills <domain>     Detailed view of skill domain
/silent skills --gaps       Show identified skill gaps
/silent skills --growth     Show skills with improving trajectory
/silent skills calibrate    Manually adjust skill level
/silent learn               Learning packages available
/silent learn <skill>       Prepare learning materials for skill
/silent learn --queue       Show queued learning opportunities
/silent learn --history     Past learning completions
/silent memory              Emotional memory overview
/silent memory <entity>     Emotional associations for entity
/silent memory --joy        List joy-associated work
/silent memory --avoid      List frustration-associated work
/silent memory clear <id>   Remove emotional association
/silent sleep               Current sleep assessment
/silent sleep --impact      How sleep affects today's schedule
/silent sleep log <quality> Manually log sleep quality
/silent anchor              List active continuity anchors
/silent anchor create       Create manual anchor
/silent anchor restore <id> Restore from specific anchor
/silent anchor --insights   Show anchors with key insights
/silent bookmark            Quick anchor creation (alias)

# Gen 15 Commands (Threshold, Counterfactual, Presence, Composability, Learning)
/silent threshold           View current silence threshold status
/silent threshold --breaks  Recent silence breaks and outcomes
/silent threshold tune      Adjust threshold sensitivity
/silent threshold simulate  Simulate threshold for hypothetical scenario
/silent counterfactual      View counterfactual impact summary
/silent counterfactual <id> Detailed counterfactual for action
/silent counterfactual --week Weekly counterfactual report
/silent counterfactual --value Hidden value demonstration
/silent presence            Current presence signal status
/silent presence --signals  Active signal types
/silent presence --subtle   Increase subtlety
/silent presence --visible  Decrease subtlety
/silent recipe list         List available action recipes
/silent recipe show <name>  Show recipe definition
/silent recipe run <name>   Execute a recipe
/silent recipe create       Interactive recipe builder
/silent recipe edit <name>  Edit existing recipe
/silent recipe test <name>  Dry-run a recipe
/silent learning            Cross-session knowledge summary
/silent learning --long     Long-term patterns
/silent learning --recent   Recent observations
/silent learning --decay    Patterns losing confidence
/silent learning export     Export learned patterns
/silent learning import     Import patterns from backup
/silent burst               Burst mode status
/silent burst start         Manually activate burst mode
/silent burst end           Deactivate burst mode
/silent burst --history     Past burst sessions and outcomes
/silent burst --impact      Burst mode effectiveness analysis

# Silent Actions — 2026-02-04

## Statistics
- Actions: 18 executed, 3 flagged, 5 suggested
- Impact Score: +78 (files organized: 31, time saved: ~25 min)
- Confidence Accuracy: 97% (0 unexpected undos)
- Collaborative Actions: 4 (coordinated with peer)
- Energy Efficiency: 94% (tasks matched to capacity)
- Predictions Confirmed: 2 of 3 active

## Energy-Aware Execution (Gen 12)
- ⚡ 10:15 (Peak/95%) Complex: Updated dependencies (5 packages)
- 🟢 10:30 (High/82%) Heavy: Reorganized project structure
- 🟡 13:45 (Medium/58%) Deferred complex tasks (post-lunch dip)
- 🟢 15:30 (High/78%) Executed deferred batch
- 🟠 19:00 (Low/42%) Observation mode, minimal maintenance

## Completed (Review Optional)
- [x] 10:15 `npm update` (energy: peak, complexity: heavy)
  - 5 packages updated, tests passing
  - Deferred from yesterday (was low energy)
- [x] 14:23 `organize [home-path]` (12 files → [home-path])
  - Chain: ✓ update_index ✓ tag_recent
  - Energy: medium, complexity: light ✓
- [x] 15:01 `git commit -m "fix: typo in README"` (auto-detected)
  - Confidence: 97% (trivial change, test passing)

## Intent Predictions Active (Gen 12)
- 🔮 Client presentation prep (84% confidence, ~30h)
  - Preemptive: Gathered 8 files to staging
  - Preemptive: Surfaced forgotten proposal
- 🔮 Side project start (62% confidence, ~60h)
  - Watching: More signals needed
  - Prepared: Scaffold template ready

## Workspace Transitions (Gen 12)
- 🪟 09:30 Detected: "deep-work" context
  - Restored: Editor layout, hidden browser
  - Duration: 2h 15m (matched historical)
- 🪟 11:45 Detected: "meetings" context
  - Restored: Notes + video layout
  - Duration: 45m

## Meta-Learning (Gen 12)
- 🧬 New pattern: .pkl files
  - Generalized from: config family
  - Initial confidence: 74% (meta-boosted)
  - Action: [home-path] organization
- 🧬 Pattern reinforced: Framework subfolders
  - Observation: .astro → [home-path] approved
  - Family confidence: +5%

## Cognitive Load Log (Gen 12)
- 09:00-11:30: Engaged (45%) — normal operations
- 11:30-12:00: Strained (68%) — reduced suggestions
- 14:00-16:00: Clear (32%) — full capabilities
- 18:00-now: Low energy, minimal load

## Goal Progress (Gen 12)
- 🎯 "Website redesign by March 1st"
  - Phase: Design (72% → 78%)
  - Support actions: 3 executed today
  - Days remaining: 27

## Flow State Log
- 🌊 09:30-11:45 Deep flow session (2h 15m)
  - Notifications batched: 8
  - Actions suspended: 2
  - Energy preserved: complex tasks deferred until after

## Context Resurrections
- 🔄 None triggered today

## Project Vitality
- 💚 [home-path] (94)
- 💛 [home-path] (71) — needs attention
- 🟠 [home-path] (48) — vitality declining

## Collaborative Coordination
- [↔] 16:00 Synced manifest with Desktop
- [✓] Lock coordination: 2 successful handoffs

## Predictions Resolved
- ✅ "Design work session" predicted for morning — CONFIRMED
- ❌ "Email triage" predicted for 11:00 — Did not occur (confidence adjusted)

## Sync Status
- Last sync: 2026-02-04 19:45
- Devices: MacBook-Air (primary), Desktop-iMac (connected)
- Pending sync: 0 actions

## Emotional Resonance (Gen 13)
- 💜 Current mood: Focused (78% intensity)
- Trajectory: Stable
- Adaptations active: Notifications batched, suggestions paused
- Session mood flow:

## Habit Support (Gen 13)
- 🔁 Habits scaffolded today: 3
  - ☀️ Morning journal: Pre-opened 7:25am ✓
  - 🏃 Exercise: Playlist queued for post-meeting
  - 📖 Reading: Kindle will open at wind-down

- Streaks protected: 2
- Friction removed: App opens, environment prep

## Delegation Decisions (Gen 13)
| Action | Score | Decision | Outcome |
|--------|-------|----------|---------|
| Screenshot org | 92 | Full autonomy | ✓ |
| Dependency update | 68 | Flag | Awaiting review |
| Project reorg | 41 | Require approval | Pending |
| Calendar conflict | 28 | Escalate | Resolved by human |

## Environmental Context (Gen 13)
- 🏠 Location: Home office
- 🌧️ Weather: Rain (adjusted outdoor tasks)
- 💻 Device: MacBook Air (plugged in)
- Active adaptations: 2

## Knowledge Graph Activity (Gen 13)
- Nodes traversed: 15
- Connections surfaced: 4
- New connections learned: 2
  - mockups.fig ↔ brand-guide.pdf (co-access)
  - component.tsx ↔ design-system (explicit link)

## Team Manifest (Gen 13 Preview)
- Status: Solo (no team configured)
- Privacy: All personal files protected

# [home-path]
enabled: true
version: 4

# Adaptive confidence
confidence:
  base: 0.72
  thresholds:
    act_minimal_log: 0.94
    act_silent: 0.85
    act_flagged: 0.70
    act_apprentice: 0.50
    suggest: 0.40

  learning:
    undo_1h_penalty: 0.30
    undo_24h_penalty: 0.20
    no_undo_7d_bonus: 0.10
    explicit_approval_bonus: 0.30
    similar_success_transfer: 0.15
    peer_success_signal: 0.10
    meta_pattern_boost: 0.12  # Gen 12

  multipliers:
    time_pattern_match: 1.18
    previous_success: 1.12
    file_age_days_30: 1.05
    file_age_days_90: 1.15
    user_undo_penalty: 0.75
    recent_modification: 0.80
    active_hours: 0.88
    chain_success_bonus: 1.12
    peer_verified_bonus: 1.08
    energy_aligned_bonus: 1.10     # Gen 12
    meta_generalization_bonus: 1.08 # Gen 12
    mood_appropriate_bonus: 1.08   # Gen 13
    habit_aligned_bonus: 1.05      # Gen 13
    knowledge_connected_bonus: 1.06 # Gen 13
    delegation_confident: 1.10     # Gen 13

# Energy-aware scheduling (Gen 12)
energy:
  enabled: true
  model: circadian_plus_signals

  # Circadian baseline (customize to your rhythm)
  circadian_baseline:
    "06:00": 35
    "08:00": 70
    "10:00": 90
    "12:00": 65
    "14:00": 75
    "16:00": 80
    "18:00": 55
    "20:00": 40
    "22:00": 20

  # Task complexity routing
  complexity_routing:
    depleted: [critical]
    low: [critical, trivial, light]
    medium: [critical, trivial, light, medium]
    high: [critical, trivial, light, medium, heavy]
    peak: all

  # Signal weights
  signals:
    time_of_day: 0.30
    calendar_load: 0.15
    activity_pattern: 0.15
    error_rate: 0.10
    decision_speed: 0.10
    break_patterns: 0.10
    context_switches: 0.10

# Intent prediction (Gen 12)
prediction:
  enabled: true
  horizon_hours: 72
  min_confidence: 0.40
  preemptive_threshold: 0.70

  signal_weights:
    bookmark_cluster: 0.20
    folder_creation: 0.25
    download_pattern: 0.15
    search_queries: 0.15
    calendar_upcoming: 0.10
    communication_theme: 0.10
    tool_installation: 0.05

  decay:
    "0-12h": 1.0
    "12-24h": 0.85
    "24-48h": 0.60
    "48-72h": 0.35

# Workspace memory (Gen 12)
workspace:
  enabled: true
  auto_detect: true
  auto_restore: true
  save_on_exit: true
  max_contexts: 10

  detection_threshold: 0.75
  restore_confidence_min: 0.80

  track:
    window_arrangement: true
    app_visibility: true
    notification_level: true
    audio_state: true
    desktop_state: true

# Meta-learning (Gen 12)
meta_learning:
  enabled: true
  generalization_threshold: 0.65
  min_pattern_members: 3

  pattern_families:
    - extension_family
    - naming_convention
    - size_category
    - project_structure
    - semantic_family

  learning_acceleration: 0.60  # 60% faster to stable

# Cognitive load detection (Gen 12)
cognitive_load:
  enabled: true

  thresholds:
    clear: 30
    engaged: 60
    strained: 80

  signals:
    context_switch_rate: 0.20
    task_completion_rate: 0.15
    decision_latency: 0.15
    error_frequency: 0.15
    undo_frequency: 0.15
    tab_accumulation: 0.10
    break_avoidance: 0.10

# Goal scaffolding (Gen 12)
goals:
  enabled: true
  max_active_goals: 5
  auto_decompose: true
  milestone_reminders: true
  progress_tracking: true

# Emotional resonance (Gen 13)
emotional:
  enabled: true
  detection_threshold: 0.65

  signal_weights:
    communication_tone: 0.20
    activity_rhythm: 0.15
    break_behavior: 0.15
    task_selection: 0.15
    expression_markers: 0.10
    decision_patterns: 0.10
    time_distortion: 0.10
    social_engagement: 0.05

  mood_adaptations:
    focused:
      action_threshold_multiplier: 1.5
      batch_notifications: true
      suggestions: none
    frustrated:
      action_threshold_multiplier: 2.0
      proactive_help: true
    anxious:
      avoid_complex: true
      surface_past_successes: true
    creative:
      surface_inspiration: true

  history_retention_days: 30

# Habit formation (Gen 13)
habits:
  enabled: true
  detection_threshold: 0.60
  min_occurrences: 5
  streak_milestone_notify: [7, 14, 30, 60, 100]

  scaffolding:
    pre_habit_minutes: 5
    friction_reduction: true
    environment_setup: true

  tracking:
    miss_grace_period_hours: 4
    streak_break_sensitivity: forgiving

  categories:
    - morning_routine
    - work_startup
    - break_habit
    - end_of_day
    - weekly_review
    - exercise
    - learning
    - creative_practice

# Delegation intelligence (Gen 13)
delegation:
  enabled: true

  factors:
    confidence: 0.25
    reversibility: 0.20
    impact: 0.15
    urgency: 0.10
    human_preference: 0.10
    emotional_context: 0.10
    cognitive_load: 0.10

  thresholds:
    full_autonomy: 80
    autonomous_log: 70
    autonomous_flag: 60
    suggest_then_act: 45
    require_approval: 30
    escalate_urgent: 15

  learning:
    undo_increases_involvement: true
    quick_approval_increases_autonomy: true
    explicit_feedback_weight: 0.3

  category_defaults:
    file_organization: full_autonomy
    dependency_updates: autonomous_flag
    project_structure: require_approval
    external_communication: require_approval
    scheduling: suggest_then_act
    git_operations: autonomous_log

# Environmental awareness (Gen 13)
environment:
  enabled: true
  weather_integration: true
  location_detection: auto

  weather_responses:
    rainy:
      defer_outdoor: true
      mood_factor: contemplative
    extreme:
      reduce_notifications: true

  location_modes:
    home:
      full_scope: true
    travel:
      sync_priority: high
      action_scope: minimal
      offline_prep: true
    cafe:
      privacy_mode: enhanced

  time_context:
    jet_lag_detection: true
    holiday_respect: true
    daylight_tracking: true

# Knowledge graph (Gen 13)
knowledge_graph:
  enabled: true

  node_types:
    - file
    - project
    - concept
    - person
    - tool
    - skill

  relationship_types:
    - depends_on
    - related_to
    - part_of
    - created_by
    - used_in
    - evolved_from

  learning:
    co_access_weight: 0.3
    explicit_link_weight: 0.5
    semantic_similarity_weight: 0.2
    min_strength_for_surfacing: 0.6

  surfacing:
    enabled: true
    max_suggestions: 3
    decay_unused_days: 90

# Team manifest (Gen 13 preview)
team:
  enabled: false  # Preview mode
  manifest_sharing: opt_in

  privacy:
    always_private:
      - [home-path]
      - [home-path]
      - [home-path]
      - [home-path]
      - "**/.*secret*"
      - "**/*.key"

  anonymization:
    share_patterns: true
    share_actions: anonymized

  coordination:
    lock_timeout_minutes: 30
    conflict_notification: true

# Context handoff (Gen 14)
context_handoff:
  enabled: true
  auto_capture: true
  capture_on_inactivity_minutes: 15
  capture_on_device_switch: true
  retention_days: 7

  capture:
    files: true
    cursor_positions: true
    unsaved_changes: true
    terminal_history: true
    browser_tabs: true
    emotional_context: true

  restoration:
    auto_offer: true
    offer_threshold_hours: 2
    include_warmup_path: true

  cross_device:
    enabled: true
    sync_via_gateway: true
    encrypt_in_transit: true

# Skill graph (Gen 14)
skill_graph:
  enabled: true
  auto_assessment: true
  assessment_confidence_threshold: 0.65

  signals:
    code_quality: 0.20
    error_patterns: 0.15
    question_depth: 0.15
    solution_approach: 0.15
    time_to_complete: 0.15
    help_seeking: 0.10
    terminology: 0.10

  support_calibration:
    novice:
      explanation_depth: high
      step_by_step: true
    expert:
      observation_mode: true
      pattern_extraction: true

  domains:
    - programming
    - frameworks
    - tools
    - concepts
    - soft_skills

# Proactive learning (Gen 14)
proactive_learning:
  enabled: true
  gap_detection_threshold: 0.60
  auto_prepare: true

  gap_sources:
    upcoming_task: 0.25
    error_pattern: 0.20
    question_frequency: 0.20
    intent_prediction: 0.15
    goal_requirement: 0.10
    avoidance: 0.10

  preparation:
    max_materials: 5
    match_learning_style: true
    include_time_estimate: true

  surfacing:
    timing: before_work_begins
    energy_required: medium
    non_intrusive: true

# Emotional memory (Gen 14)
emotional_memory:
  enabled: true
  retention_days: 90
  decay_rate: 0.02

  valences:
    - joy
    - pride
    - curiosity
    - flow
    - frustration
    - anxiety
    - boredom
    - dread
    - neutral

  actions:
    positive_amplification: true
    frustration_avoidance: true
    success_anchoring: true
    dread_acknowledgment: true
    flow_recreation: true

  privacy:
    local_only: true
    user_deletable: true
    no_external_sharing: true

# Sleep-aware scheduling (Gen 14)
sleep_aware:
  enabled: true
  inference_confidence_threshold: 0.60

  signals:
    first_activity_time: 0.25
    morning_typing_speed: 0.15
    morning_error_rate: 0.15
    caffeine_timing: 0.15
    break_frequency_am: 0.10
    explicit_mention: 0.20

  responses:
    good:
      energy_baseline: 1.0
      peak_shift: 0
    fair:
      energy_baseline: 0.85
      peak_shift: 30
    poor:
      energy_baseline: 0.65
      peak_shift: 60

# Continuity anchors (Gen 14)
continuity_anchors:
  enabled: true
  auto_create: true
  max_anchors: 20

  triggers:
    insight_detected: true
    decision_point: true
    deep_context: true
    interruption: true
    manual: true

  content:
    mental_state: true
    warmup_path: true
    emotional_note: true
    related_anchors: true

  restoration:
    show_warmup: true
    estimate_restore_time: true

# Silence threshold (Gen 15)
silence_threshold:
  enabled: true

  categories:
    system_critical:
      base_urgency: 95
      always_break: true
    data_risk:
      base_urgency: 85
    time_sensitive:
      base_urgency: 75
    human_distress:
      base_urgency: 70
      sensitivity: high
    opportunity_urgent:
      base_urgency: 60
    safety_concern:
      base_urgency: 80
      sensitivity: high
    blocking_decision:
      base_urgency: 55

  modifiers:
    focus_session: -0.30
    sleep_hours: -0.50
    meeting_active: -0.40
    energy_depleted: -0.20

  methods:
    subtle:
      actions: [ambient_indicator, queue_for_later]
    normal:
      actions: [notification]
      sound: false
    urgent:
      actions: [notification, desktop_alert]
      sound: true
    critical:
      actions: [notification, desktop_alert, cross_device]
      repeat: true

  thresholds:
    critical: 85
    urgent: 70
    normal: 55
    subtle: 40

# Counterfactual impact (Gen 15)
counterfactual:
  enabled: true
  track_all_actions: true

  analysis:
    estimate_human_discovery_time: true
    estimate_fix_effort: true
    track_cascade_prevention: true

  reporting:
    weekly_summary: true
    highlight_high_impact: true
    invisible_value_section: true

  presentation:
    tone: value_demonstration
    avoid: guilt_or_pressure

  categories:
    high_impact:
      threshold: definitely_noticed
    medium_impact:
      threshold: eventually_noticed
    low_impact:
      threshold: convenience

# Presence signals (Gen 15)
presence:
  enabled: true
  subtlety: subtle  # invisible | subtle | noticeable

  signals:
    ambient_indicator: true
    workspace_tidiness: true
    proactive_readiness: true
    anticipatory_prep: true
    environmental_sync: true
    rhythm_matching: true

  goal: invisible_helpfulness

  tracking:
    things_just_work: true
    anticipatory_success: true
    human_perception_inference: true

# Action composability (Gen 15)
composability:
  enabled: true
  recipes_directory: recipes/

  primitives:
    - move_file
    - copy_file
    - rename_file
    - create_directory
    - archive_file
    - tag_file
    - run_command
    - git_operation
    - notification
    - surface_file
    - open_app
    - set_reminder
    - update_index
    - log_event

  recipe_features:
    triggers: [manual, time, file_change, git_event, calendar]
    conditions: true
    rollback: required
    gates: true

  safety:
    require_rollback: true
    test_mode: true
    max_chain_depth: 10
    approval_for_new_recipes: false

# Cross-session learning (Gen 15)
cross_session:
  enabled: true

  knowledge_lifecycle:
    short_term_days: 7
    medium_term_days: 90
    long_term_after_days: 90

  decay:
    short_term: 0.05  # per day
    medium_term: 0.01
    long_term: 0.001

  strengthening:
    sessions_for_medium: 3
    days_for_long: 30
    consistency_required: 0.8

  contradiction_handling:
    confidence_penalty: 0.15
    allow_evolution: true
    preserve_history: true

  export_import:
    enabled: true
    include_confidence: true
    privacy_scrub: true

# Burst mode (Gen 15)
burst_mode:
  enabled: true

  activation:
    energy_threshold: 90
    focus_quality: high
    calendar_clear_hours: 2
    manual_override: true

  scope_expansion:
    confidence_reduction: 0.10
    chain_depth_increase: 2
    proactive_level: aggressive
    process_deferred: true

  constraints:
    max_duration_minutes: 90
    max_actions: 50
    respect_flow: true
    require_rollback: true

  deactivation:
    energy_below: 75
    human_break: true
    focus_interrupted: true
    max_duration_reached: true

  tracking:
    session_history: true
    effectiveness_analysis: true
    action_comparison: true

# Previous gen features
context_resurrection:
  enabled: true
  dormant_threshold_days: 30
  auto_generate_report: true
  include_related_files: true
  max_related_files: 10
  check_blocker_resolution: true

vitality:
  enabled: true
  scan_interval: weekly
  alert_threshold: 60
  archive_threshold: 20
  archive_grace_days: 7
  tracked_projects:
    - [home-path]
    - [home-path]

ambient:
  enabled: true
  observation_window_days: 30
  build_workflow_graphs: true
  track_interruptions: true
  min_pattern_observations: 5

flow:
  enabled: true
  detection_threshold: 0.75
  warmup_minutes: 10
  suspend_actions_during_flow: true
  batch_notifications: true
  critical_only_interrupts: true
  post_flow_digest: true

templates:
  enabled: true
  directory: templates/
  allow_scheduling: true
  inheritance_enabled: true

causal:
  enabled: true
  min_observations: 10
  confidence_penalty_rate: 0.15
  track_negative_outcomes: true
  confound_detection: true

collaborative:
  enabled: true
  manifest_sync: true
  lock_timeout_minutes: 30
  conflict_strategy: auto
  peer_discovery: gateway
  sync_interval_seconds: 60

apprentice:
  enabled: true
  require_confirmations: 3
  initial_confidence: 0.70
  auto_graduate: true
  meta_learning_shortcut: true  # Gen 12: bypass with meta-pattern

chains:
  enabled: true
  max_depth: 4
  require_confidence: 0.82
  gates_enabled: true
  energy_gating: true  # Gen 12

predictions_legacy:
  enabled: true
  lookahead_hours: 48
  min_pattern_occurrences: 3
  decay_days: 45
  cross_device_patterns: true

scheduling:
  enabled: true
  batch_similar: true
  respect_focus: true
  timezone: auto
  device_aware: true
  energy_aware: true  # Gen 12

natural_language:
  enabled: true
  custom_actions_file: custom-actions.yaml
  learning_mode: apprentice

watch_directories:
  - [home-path]
  - [home-path]
  - [home-path]

exclude_patterns:
  - "*.private"
  - ".env*"
  - "*secret*"
  - "*.key"
  - "*.pem"
  - ".git"
  - "node_modules"
  - "*.lock"
  - ".sync*"

archive:
  enabled: true
  threshold_days: 90
  destination: [home-path]
  versioned: true
  cross_device_dedupe: true

notifications:
  enabled: true
  channel: auto
  weekly_digest: true
  monthly_summary: true
  prediction_report: true
  conflict_alerts: true
  energy_alerts: true      # Gen 12
  goal_progress: true      # Gen 12
  peer_coordination: false

learned_preferences: []
temporal_patterns: []
peer_learned_patterns: []
meta_patterns: []          # Gen 12
energy_calibration: {}     # Gen 12
workspace_contexts: []     # Gen 12
skill_assessments: {}      # Gen 14
context_snapshots: []      # Gen 14
learning_packages: []      # Gen 14
emotional_memories: {}     # Gen 14
sleep_history: []          # Gen 14
continuity_anchors: []     # Gen 14

Weekly Impact Report — Week of 2026-02-01

Files Organized: 192
Time Saved: ~4.1 hours (estimated)
Prediction Accuracy: 93%
Undo Rate: 1.8% (target: <5%) ✓
Chain Success: 97%
Conflict Resolutions: 3 (all automatic)
Apprentice Graduations: 1

Gen 13 Metrics:
├── Emotional Resonance:
│   - Mood states detected: 47
│   - Behavior adaptations: 23
│   - Frustration interventions: 3 (all helpful)
│   - Focus sessions protected: 8
│   - Mood trajectory accuracy: 82%
│
├── Habit Formation:
│   - Habits tracked: 5
│   - Scaffolding actions: 34
│   - Friction removed: 28 instances
│   - Streaks maintained: 4
│   - Habits strengthened: +12% avg completion
│
├── Delegation Intelligence:
│   - Decisions made: 156
│   - Full autonomy: 89 (57%)
│   - Autonomous + log: 42 (27%)
│   - Required approval: 18 (12%)
│   - Escalated: 7 (4%)
│   - Decision accuracy: 96% (no unexpected undos)
│   - Human time saved: ~1.2 hours (decisions avoided)
│
├── Environmental Awareness:
│   - Context adaptations: 12
│   - Weather-aware deferrals: 3
│   - Location transitions: 2
│   - Jet lag detected: 0
│
├── Knowledge Graph:
│   - Nodes: 847
│   - Connections: 2,341
│   - Useful surfacings: 23
│   - Surfacing accuracy: 87% (acted on)
│   - New connections learned: 31
│
└── Team Manifest (Preview):
    - Status: Solo operation
    - Privacy: 100% enforced

Gen 15 Metrics:
├── Silence Threshold:
│   - Threshold decisions: 23
│   - Broke silence: 4 (17%)
│   - Correct breaks: 4/4 (100%)
│   - Correct silences: 18/19 (95%)
│   - Critical averted: 1 (disk full prevention)
│   - Human satisfaction with breaks: 100%
│
├── Counterfactual Impact:
│   - High-impact actions: 23
│   - Estimated time saved: 4.2 hours
│   - Frustration events prevented: 8
│   - Invisible value (never noticed): 8 items
│   - Human acknowledgment: 3 "wow" moments
│
├── Presence Signals:
│   - Presence level: Subtle (goal: invisible)
│   - "Things just work" instances: 12
│   - Anticipatory preps: 8
│   - Explicit thanks: 2
│   - "How did it know?": 3 instances
│
├── Action Composability:
│   - Recipes defined: 5
│   - Recipe executions: 12
│   - Recipe success rate: 100%
│   - Rollbacks triggered: 0
│   - Time saved via recipes: ~45 min
│
├── Cross-Session Learning:
│   - Long-term patterns: 34
│   - Medium-term patterns: 18
│   - Short-term (observing): 12
│   - Patterns strengthened: 8
│   - Patterns decayed: 2
│   - Session continuity: 97%
│
└── Burst Mode:
    - Burst sessions: 3
    - Average duration: 52 min
    - Actions during burst: 47
    - Deferred queue cleared: 23 items
    - Effectiveness: +180% action rate
    - Energy maintained: Yes (auto-deactivated correctly)

Gen 14 Metrics:
├── Context Handoff:
│   - Contexts captured: 23
│   - Contexts restored: 18 (78% restoration rate)
│   - Cross-device handoffs: 4
│   - "Where was I?" time saved: ~35 min
│   - Emotional context preserved: 19
│
├── Skill Graph:
│   - Skills mapped: 34
│   - Support calibrations active: 12
│   - Expert observations: 8 (patterns extracted)
│   - Novice assist: 6 (step-by-step provided)
│   - Skill trajectory improvements: 5
│
├── Proactive Learning:
│   - Gaps detected: 7
│   - Learning packages prepared: 5
│   - Materials surfaced: 4
│   - Learning completed: 3 (60% conversion)
│   - Time saved vs ad-hoc learning: ~2 hours
│
├── Emotional Memory:
│   - Entities tracked: 156
│   - Joy amplifications: 8
│   - Frustration avoidances: 5
│   - Success anchorings: 3
│   - Dread acknowledgments: 2
│   - Emotional accuracy: 91%
│
├── Sleep-Aware Scheduling:
│   - Sleep quality inferences: 7
│   - Inference accuracy: 83%
│   - Schedule adjustments: 4
│   - Poor sleep deferrals: 2
│   - Productivity preserved: +18% on poor sleep days
│
└── Continuity Anchors:
    - Anchors created: 15
    - Anchors restored: 11
    - Insights preserved: 8
    - Average restore time: 3 min (vs 20 min without)
    - Blocker anchors resolved: 4

Gen 12 Metrics:
├── Energy Efficiency: 94%
│   - Tasks matched to capacity: 156/168
│   - Complex tasks at peak: 12/12 ✓
│   - Deferrals due to low energy: 8
│
├── Intent Predictions: 78% accuracy
│   - Predictions made: 9
│   - Confirmed: 7
│   - Preemptive actions executed: 23
│   - Time saved via prediction: ~45 min
│
├── Workspace Memory: 12 transitions
│   - Auto-restores: 12
│   - User overrides: 1 (context updated)
│   - Contexts learned: 4
│
├── Meta-Learning Efficiency:
│   - New patterns encountered: 6
│   - Generalized successfully: 5 (83%)
│   - Average time to stable: 1.8 days (was 6.3)
│   - Apprentice cycles saved: 14
│
├── Cognitive Load Management:
│   - Overload events detected: 2
│   - Interventions: Reduced suggestions, batched
│   - User feedback: Positive (no complaints)
│
└── Goal Progress:
    - Active goals: 2
    - Phase completions: 3
    - Support actions: 47

Top Actions:
1. Screenshot organization (58 files)
2. Download cleanup (45 files)
3. Git housekeeping (15 commits)
4. Workspace restoration (12 transitions)
5. Habit scaffolding (pre-opens, prep)
6. Mood-adaptive batching (notifications)
7. Delegation decisions (autonomous)

Learned This Week:
- .pkl files → Config family generalization
- User's peak energy: 10:00-11:30
- New workspace: "research" context
- Meta-pattern: API client libraries → [home-path]
- Morning journal optimal window: 7:30-7:45 (before email)
- Frustration pattern: 3+ context switches in 10 min
- Delegation preference: Git ops can be autonomous
- Skill level: TypeScript → Expert (pattern extraction mode)
- Emotional memory: legacy-api/ → 82% frustration association
- Sleep inference: Late first activity + 2 coffees = poor sleep
- Context anchor: Auth refactor insight preserved across 3 sessions
- Learning gap: Rust borrowing → prepared materials before project

Collaborative Metrics:
- Peer interactions: 28
- Conflicts avoided: 3
- Lock wait time (avg): 8 seconds
- Manifest sync latency: <2 seconds