chore(deploy): bump image to 7840ef9e [skip ci]

fix(multi): resolve 26 bugs found in comprehensive codebase review
Core fixes (issues #12, #13, #22): - #12: undo now preserves score delta instead of restoring snapshot score - #13: take_from_foundation defaults to false (non-standard house rule) - #22: check_win validates full suit sequence, not just card count Engine fixes: - #8: replay keyboard input guard against non-replay state - #9: help modal scrims.is_empty() guard added - #10: settings modal scrims.is_empty() guard added - #11: sync_plugin builds payload at poll time (not task-spawn time) - #14: server replay mode case-sensitivity fix ("Classic") - #15: play_by_seed_plugin confirmed flag set to true on launch - #16: replay back-step debounce via Local<bool> + StateChangedEvent; register StateChangedEvent in ReplayOverlayPlugin (fixes 52 tests) - #17: time-attack timer ignores win-summary overlay - #18: HUD dropdown glyphs U+25BE → U+2193 (FiraMono-safe arrow) - #19: theme plugin applies immediate visual update on A→B→A switch - #20: SyncAuthError / SyncBusyOverlay split into separate entities so auth errors are visible after busy overlay is hidden - #21: handle_forfeit ordered before update_stats_on_new_game - #23: server merge uses correct avg_time_seconds and games_lost math - #24: win_summary migrated to ModalScrim pattern - #25: card_animation apply_deferred between animation systems - #26: cursor_plugin HashMap access uses .get() with fallback - #27: auto_complete mid-sequence deactivation guard - #28: feedback_anim SettleAnim ordered before FoundationFlourish - #29: achievement_plugin iterates all win events; adds scrims guard - #30: leaderboard modal scrims.is_empty() guard added - #31: server auth tmp file cleanup on rename failure - #32: sync_setup modal scrims.is_empty() guard added - #33: font_plugin uses match fallback; TokioRuntimeResource graceful current-thread fallback on runtime init failure Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-19 20:19:02 +00:00 · 2026-05-19 13:14:47 -07:00 · 2026-05-19 11:32:07 -07:00 · 2026-05-19 11:32:07 -07:00 · 2026-05-19 11:32:07 -07:00 · 2026-05-19 11:32:07 -07:00
39 changed files with 1081 additions and 199 deletions
@@ -0,0 +1,7 @@
 # Claude Flow runtime files
 data/
 logs/
 sessions/
 neural/
 *.log
 *.tmp
@@ -0,0 +1,403 @@
 # RuFlo V3 - Complete Capabilities Reference
 > Generated: 2026-05-19T00:18:20.864Z
 > Full documentation: https://github.com/ruvnet/claude-flow
 ## 📋 Table of Contents
 1. [Overview](#overview)
 2. [Swarm Orchestration](#swarm-orchestration)
 3. [Available Agents (60+)](#available-agents)
 4. [CLI Commands (26 Commands, 140+ Subcommands)](#cli-commands)
 5. [Hooks System (27 Hooks + 12 Workers)](#hooks-system)
 6. [Memory & Intelligence (RuVector)](#memory--intelligence)
 7. [Hive-Mind Consensus](#hive-mind-consensus)
 8. [Performance Targets](#performance-targets)
 9. [Integration Ecosystem](#integration-ecosystem)
 ---
 ## Overview
 RuFlo V3 is a domain-driven design architecture for multi-agent AI coordination with:
 - **15-Agent Swarm Coordination** with hierarchical and mesh topologies
 - **HNSW Vector Search** - 150x-12,500x faster pattern retrieval
 - **SONA Neural Learning** - Self-optimizing with <0.05ms adaptation
 - **Byzantine Fault Tolerance** - Queen-led consensus mechanisms
 - **MCP Server Integration** - Model Context Protocol support
 ### Current Configuration
 | Setting | Value |
 |---------|-------|
 | Topology | hierarchical-mesh |
 | Max Agents | 15 |
 | Memory Backend | hybrid |
 | HNSW Indexing | Enabled |
 | Neural Learning | Enabled |
 | LearningBridge | Enabled (SONA + ReasoningBank) |
 | Knowledge Graph | Enabled (PageRank + Communities) |
 | Agent Scopes | Enabled (project/local/user) |
 ---
 ## Swarm Orchestration
 ### Topologies
 | Topology | Description | Best For |
 |----------|-------------|----------|
 | `hierarchical` | Queen controls workers directly | Anti-drift, tight control |
 | `mesh` | Fully connected peer network | Distributed tasks |
 | `hierarchical-mesh` | V3 hybrid (recommended) | 10+ agents |
 | `ring` | Circular communication | Sequential workflows |
 | `star` | Central coordinator | Simple coordination |
 | `adaptive` | Dynamic based on load | Variable workloads |
 ### Strategies
 - `balanced` - Even distribution across agents
 - `specialized` - Clear roles, no overlap (anti-drift)
 - `adaptive` - Dynamic task routing
 ### Quick Commands
 ```bash
 # Initialize swarm
 npx @claude-flow/cli@latest swarm init --topology hierarchical --max-agents 8 --strategy specialized
 # Check status
 npx @claude-flow/cli@latest swarm status
 # Monitor activity
 npx @claude-flow/cli@latest swarm monitor
 ```
 ---
 ## Available Agents
 ### Core Development (5)
 `coder`, `reviewer`, `tester`, `planner`, `researcher`
 ### V3 Specialized (4)
 `security-architect`, `security-auditor`, `memory-specialist`, `performance-engineer`
 ### Swarm Coordination (5)
 `hierarchical-coordinator`, `mesh-coordinator`, `adaptive-coordinator`, `collective-intelligence-coordinator`, `swarm-memory-manager`
 ### Consensus & Distributed (7)
 `byzantine-coordinator`, `raft-manager`, `gossip-coordinator`, `consensus-builder`, `crdt-synchronizer`, `quorum-manager`, `security-manager`
 ### Performance & Optimization (5)
 `perf-analyzer`, `performance-benchmarker`, `task-orchestrator`, `memory-coordinator`, `smart-agent`
 ### GitHub & Repository (9)
 `github-modes`, `pr-manager`, `code-review-swarm`, `issue-tracker`, `release-manager`, `workflow-automation`, `project-board-sync`, `repo-architect`, `multi-repo-swarm`
 ### SPARC Methodology (6)
 `sparc-coord`, `sparc-coder`, `specification`, `pseudocode`, `architecture`, `refinement`
 ### Specialized Development (8)
 `backend-dev`, `mobile-dev`, `ml-developer`, `cicd-engineer`, `api-docs`, `system-architect`, `code-analyzer`, `base-template-generator`
 ### Testing & Validation (2)
 `tdd-london-swarm`, `production-validator`
 ### Agent Routing by Task
 | Task Type | Recommended Agents | Topology |
 |-----------|-------------------|----------|
 | Bug Fix | researcher, coder, tester | mesh |
 | New Feature | coordinator, architect, coder, tester, reviewer | hierarchical |
 | Refactoring | architect, coder, reviewer | mesh |
 | Performance | researcher, perf-engineer, coder | hierarchical |
 | Security | security-architect, auditor, reviewer | hierarchical |
 | Docs | researcher, api-docs | mesh |
 ---
 ## CLI Commands
 ### Core Commands (12)
 | Command | Subcommands | Description |
 |---------|-------------|-------------|
 | `init` | 4 | Project initialization |
 | `agent` | 8 | Agent lifecycle management |
 | `swarm` | 6 | Multi-agent coordination |
 | `memory` | 11 | AgentDB with HNSW search |
 | `mcp` | 9 | MCP server management |
 | `task` | 6 | Task assignment |
 | `session` | 7 | Session persistence |
 | `config` | 7 | Configuration |
 | `status` | 3 | System monitoring |
 | `workflow` | 6 | Workflow templates |
 | `hooks` | 17 | Self-learning hooks |
 | `hive-mind` | 6 | Consensus coordination |
 ### Advanced Commands (14)
 | Command | Subcommands | Description |
 |---------|-------------|-------------|
 | `daemon` | 5 | Background workers |
 | `neural` | 5 | Pattern training |
 | `security` | 6 | Security scanning |
 | `performance` | 5 | Profiling & benchmarks |
 | `providers` | 5 | AI provider config |
 | `plugins` | 5 | Plugin management |
 | `deployment` | 5 | Deploy management |
 | `embeddings` | 4 | Vector embeddings |
 | `claims` | 4 | Authorization |
 | `migrate` | 5 | V2→V3 migration |
 | `process` | 4 | Process management |
 | `doctor` | 1 | Health diagnostics |
 | `completions` | 4 | Shell completions |
 ### Example Commands
 ```bash
 # Initialize
 npx @claude-flow/cli@latest init --wizard
 # Spawn agent
 npx @claude-flow/cli@latest agent spawn -t coder --name my-coder
 # Memory operations
 npx @claude-flow/cli@latest memory store --key "pattern" --value "data" --namespace patterns
 npx @claude-flow/cli@latest memory search --query "authentication"
 # Diagnostics
 npx @claude-flow/cli@latest doctor --fix
 ```
 ---
 ## Hooks System
 ### 27 Available Hooks
 #### Core Hooks (6)
 | Hook | Description |
 |------|-------------|
 | `pre-edit` | Context before file edits |
 | `post-edit` | Record edit outcomes |
 | `pre-command` | Risk assessment |
 | `post-command` | Command metrics |
 | `pre-task` | Task start + agent suggestions |
 | `post-task` | Task completion learning |
 #### Session Hooks (4)
 | Hook | Description |
 |------|-------------|
 | `session-start` | Start/restore session |
 | `session-end` | Persist state |
 | `session-restore` | Restore previous |
 | `notify` | Cross-agent notifications |
 #### Intelligence Hooks (5)
 | Hook | Description |
 |------|-------------|
 | `route` | Optimal agent routing |
 | `explain` | Routing decisions |
 | `pretrain` | Bootstrap intelligence |
 | `build-agents` | Generate configs |
 | `transfer` | Pattern transfer |
 #### Coverage Hooks (3)
 | Hook | Description |
 |------|-------------|
 | `coverage-route` | Coverage-based routing |
 | `coverage-suggest` | Improvement suggestions |
 | `coverage-gaps` | Gap analysis |
 ### 12 Background Workers
 | Worker | Priority | Purpose |
 |--------|----------|---------|
 | `ultralearn` | normal | Deep knowledge |
 | `optimize` | high | Performance |
 | `consolidate` | low | Memory consolidation |
 | `predict` | normal | Predictive preload |
 | `audit` | critical | Security |
 | `map` | normal | Codebase mapping |
 | `preload` | low | Resource preload |
 | `deepdive` | normal | Deep analysis |
 | `document` | normal | Auto-docs |
 | `refactor` | normal | Suggestions |
 | `benchmark` | normal | Benchmarking |
 | `testgaps` | normal | Coverage gaps |
 ---
 ## Memory & Intelligence
 ### RuVector Intelligence System
 - **SONA**: Self-Optimizing Neural Architecture (<0.05ms)
 - **MoE**: Mixture of Experts routing
 - **HNSW**: 150x-12,500x faster search
 - **EWC++**: Prevents catastrophic forgetting
 - **Flash Attention**: 2.49x-7.47x speedup
 - **Int8 Quantization**: 3.92x memory reduction
 ### 4-Step Intelligence Pipeline
 1. **RETRIEVE** - HNSW pattern search
 2. **JUDGE** - Success/failure verdicts
 3. **DISTILL** - LoRA learning extraction
 4. **CONSOLIDATE** - EWC++ preservation
 ### Self-Learning Memory (ADR-049)
 | Component | Status | Description |
 |-----------|--------|-------------|
 | **LearningBridge** | ✅ Enabled | Connects insights to SONA/ReasoningBank neural pipeline |
 | **MemoryGraph** | ✅ Enabled | PageRank knowledge graph + community detection |
 | **AgentMemoryScope** | ✅ Enabled | 3-scope agent memory (project/local/user) |
 **LearningBridge** - Insights trigger learning trajectories. Confidence evolves: +0.03 on access, -0.005/hour decay. Consolidation runs the JUDGE/DISTILL/CONSOLIDATE pipeline.
 **MemoryGraph** - Builds a knowledge graph from entry references. PageRank identifies influential insights. Communities group related knowledge. Graph-aware ranking blends vector + structural scores.
 **AgentMemoryScope** - Maps Claude Code 3-scope directories:
 - `project`: `<gitRoot>/.claude/agent-memory/<agent>/`
 - `local`: `<gitRoot>/.claude/agent-memory-local/<agent>/`
 - `user`: `~/.claude/agent-memory/<agent>/`
 High-confidence insights (>0.8) can transfer between agents.
 ### Memory Commands
 ```bash
 # Store pattern
 npx @claude-flow/cli@latest memory store --key "name" --value "data" --namespace patterns
 # Semantic search
 npx @claude-flow/cli@latest memory search --query "authentication"
 # List entries
 npx @claude-flow/cli@latest memory list --namespace patterns
 # Initialize database
 npx @claude-flow/cli@latest memory init --force
 ```
 ---
 ## Hive-Mind Consensus
 ### Queen Types
 | Type | Role |
 |------|------|
 | Strategic Queen | Long-term planning |
 | Tactical Queen | Execution coordination |
 | Adaptive Queen | Dynamic optimization |
 ### Worker Types (8)
 `researcher`, `coder`, `analyst`, `tester`, `architect`, `reviewer`, `optimizer`, `documenter`
 ### Consensus Mechanisms
 | Mechanism | Fault Tolerance | Use Case |
 |-----------|-----------------|----------|
 | `byzantine` | f < n/3 faulty | Adversarial |
 | `raft` | f < n/2 failed | Leader-based |
 | `gossip` | Eventually consistent | Large scale |
 | `crdt` | Conflict-free | Distributed |
 | `quorum` | Configurable | Flexible |
 ### Hive-Mind Commands
 ```bash
 # Initialize
 npx @claude-flow/cli@latest hive-mind init --queen-type strategic
 # Status
 npx @claude-flow/cli@latest hive-mind status
 # Spawn workers
 npx @claude-flow/cli@latest hive-mind spawn --count 5 --type worker
 # Consensus
 npx @claude-flow/cli@latest hive-mind consensus --propose "task"
 ```
 ---
 ## Performance Targets
 | Metric | Target | Status |
 |--------|--------|--------|
 | HNSW Search | 150x-12,500x faster | ✅ Implemented |
 | Memory Reduction | 50-75% | ✅ Implemented (3.92x) |
 | SONA Integration | Pattern learning | ✅ Implemented |
 | Flash Attention | 2.49x-7.47x | 🔄 In Progress |
 | MCP Response | <100ms | ✅ Achieved |
 | CLI Startup | <500ms | ✅ Achieved |
 | SONA Adaptation | <0.05ms | 🔄 In Progress |
 | Graph Build (1k) | <200ms | ✅ 2.78ms (71.9x headroom) |
 | PageRank (1k) | <100ms | ✅ 12.21ms (8.2x headroom) |
 | Insight Recording | <5ms/each | ✅ 0.12ms (41x headroom) |
 | Consolidation | <500ms | ✅ 0.26ms (1,955x headroom) |
 | Knowledge Transfer | <100ms | ✅ 1.25ms (80x headroom) |
 ---
 ## Integration Ecosystem
 ### Integrated Packages
 | Package | Version | Purpose |
 |---------|---------|---------|
 | agentic-flow | 3.0.0-alpha.1 | Core coordination + ReasoningBank + Router |
 | agentdb | 3.0.0-alpha.10 | Vector database + 8 controllers |
 | @ruvector/attention | 0.1.3 | Flash attention |
 | @ruvector/sona | 0.1.5 | Neural learning |
 ### Optional Integrations
 | Package | Command |
 |---------|---------|
 | ruv-swarm | `npx ruv-swarm mcp start` |
 | flow-nexus | `npx flow-nexus@latest mcp start` |
 | agentic-jujutsu | `npx agentic-jujutsu@latest` |
 ### MCP Server Setup
 ```bash
 # Add Ruflo MCP
 claude mcp add ruflo -- npx -y ruflo@latest
 # Optional servers
 claude mcp add ruv-swarm -- npx -y ruv-swarm mcp start
 claude mcp add flow-nexus -- npx -y flow-nexus@latest mcp start
 ```
 ---
 ## Quick Reference
 ### Essential Commands
 ```bash
 # Setup
 npx ruflo@latest init --wizard
 npx ruflo@latest daemon start
 npx ruflo@latest doctor --fix
 # Swarm
 npx ruflo@latest swarm init --topology hierarchical --max-agents 8
 npx ruflo@latest swarm status
 # Agents
 npx ruflo@latest agent spawn -t coder
 npx ruflo@latest agent list
 # Memory
 npx ruflo@latest memory search --query "patterns"
 # Hooks
 npx ruflo@latest hooks pre-task --description "task"
 npx ruflo@latest hooks worker dispatch --trigger optimize
 ```
 ### File Structure
 ```
 .claude-flow/
 ├── config.yaml      # Runtime configuration
 ├── CAPABILITIES.md  # This file
 ├── data/            # Memory storage
 ├── logs/            # Operation logs
 ├── sessions/        # Session state
 ├── hooks/           # Custom hooks
 ├── agents/          # Agent configs
 └── workflows/       # Workflow templates
 ```
 ---
 **Full Documentation**: https://github.com/ruvnet/claude-flow
 **Issues**: https://github.com/ruvnet/claude-flow/issues
@@ -0,0 +1,43 @@
 # RuFlo V3 Runtime Configuration
 # Generated: 2026-05-19T00:18:20.863Z
 version: "3.0.0"
 swarm:
  topology: hierarchical-mesh
  maxAgents: 15
  autoScale: true
  coordinationStrategy: consensus
 memory:
  backend: hybrid
  enableHNSW: true
  persistPath: .claude-flow/data
  cacheSize: 100
  # ADR-049: Self-Learning Memory
  learningBridge:
    enabled: true
    sonaMode: balanced
    confidenceDecayRate: 0.005
    accessBoostAmount: 0.03
    consolidationThreshold: 10
  memoryGraph:
    enabled: true
    pageRankDamping: 0.85
    maxNodes: 5000
    similarityThreshold: 0.8
  agentScopes:
    enabled: true
    defaultScope: project
 neural:
  enabled: true
  modelPath: .claude-flow/neural
 hooks:
  enabled: true
  autoExecute: true
 mcp:
  autoStart: false
  port: 3000
@@ -0,0 +1,17 @@
 {
  "initialized": "2026-05-19T00:18:20.864Z",
  "routing": {
    "accuracy": 0,
    "decisions": 0
  },
  "patterns": {
    "shortTerm": 0,
    "longTerm": 0,
    "quality": 0
  },
  "sessions": {
    "total": 0,
    "current": null
  },
  "_note": "Intelligence grows as you use Ruflo"
 }
@@ -0,0 +1,18 @@
 {
  "timestamp": "2026-05-19T00:18:20.864Z",
  "processes": {
    "agentic_flow": 0,
    "mcp_server": 0,
    "estimated_agents": 0
  },
  "swarm": {
    "active": false,
    "agent_count": 0,
    "coordination_active": false
  },
  "integration": {
    "agentic_flow_active": false,
    "mcp_active": false
  },
  "_initialized": true
 }
@@ -0,0 +1,26 @@
 {
  "version": "3.0.0",
  "initialized": "2026-05-19T00:18:20.864Z",
  "domains": {
    "completed": 0,
    "total": 5,
    "status": "INITIALIZING"
  },
  "ddd": {
    "progress": 0,
    "modules": 0,
    "totalFiles": 0,
    "totalLines": 0
  },
  "swarm": {
    "activeAgents": 0,
    "maxAgents": 15,
    "topology": "hierarchical-mesh"
  },
  "learning": {
    "status": "READY",
    "patternsLearned": 0,
    "sessionsCompleted": 0
  },
  "_note": "Metrics will update as you use Ruflo. Run: npx ruflo@latest daemon start"
 }
@@ -0,0 +1,8 @@
 {
  "initialized": "2026-05-19T00:18:20.864Z",
  "status": "PENDING",
  "cvesFixed": 0,
  "totalCves": 3,
  "lastScan": null,
  "_note": "Run: npx @claude-flow/cli@latest security scan"
 }
@@ -8,6 +8,10 @@
 data/
 .claude/
 # ruflo runtime state
 agentdb.rvf
 agentdb.rvf.lock
 # IDE project files
 .idea/
@@ -0,0 +1,22 @@
 {
  "mcpServers": {
    "ruflo": {
      "command": "npx",
      "args": [
        "-y",
        "ruflo@latest",
        "mcp",
        "start"
      ],
      "env": {
        "npm_config_update_notifier": "false",
        "CLAUDE_FLOW_MODE": "v3",
        "CLAUDE_FLOW_HOOKS_ENABLED": "true",
        "CLAUDE_FLOW_TOPOLOGY": "hierarchical-mesh",
        "CLAUDE_FLOW_MAX_AGENTS": "15",
        "CLAUDE_FLOW_MEMORY_BACKEND": "hybrid"
      },
      "autoStart": false
    }
  }
 }
@@ -20,4 +20,4 @@ resources:
 images:
 - name: solitaire-server
  newName: git.aleshym.co/funman300/solitaire-server
-  newTag: eb6c93fb
+  newTag: 7840ef9e
@@ -193,7 +193,7 @@ impl GameState {
            is_auto_completable: false,
            undo_count: 0,
            recycle_count: 0,
-            take_from_foundation: true,
+            take_from_foundation: false,
            schema_version: GAME_STATE_SCHEMA_VERSION,
            undo_stack: VecDeque::new(),
        }
@@ -407,7 +407,7 @@ impl GameState {
        self.score = if self.mode == GameMode::Zen {
            0
        } else {
-            (snapshot.score + scoring_undo()).max(0)
+            (self.score + scoring_undo()).max(0)
        };
        self.move_count = snapshot.move_count;
        self.is_won = false;
@@ -416,12 +416,25 @@ impl GameState {
        Ok(())
    }
-    /// Returns `true` when all four foundation slots each contain 13 cards.
+    /// Returns `true` when all four foundation slots each contain a valid A→K
    /// sequence of a single suit.
    ///
    /// Counting 13 cards is not sufficient — a corrupt save could produce 13
    /// arbitrary cards per pile and permanently lock the game via `GameAlreadyWon`.
    pub fn check_win(&self) -> bool {
-        (0..4_u8).all(|slot| {
+        (0..4_u8).all(|slot| self.is_valid_foundation_pile(slot))
-            self.piles
+    }
-                .get(&PileType::Foundation(slot))
+
-                .is_some_and(|p| p.cards.len() == 13)
+    fn is_valid_foundation_pile(&self, slot: u8) -> bool {
        let Some(pile) = self.piles.get(&PileType::Foundation(slot)) else {
            return false;
        };
        if pile.cards.len() != 13 {
            return false;
        }
        let suit = pile.cards[0].suit;
        pile.cards.iter().enumerate().all(|(i, card)| {
            card.suit == suit && card.rank.value() == (i as u8 + 1)
        })
    }
@@ -1395,12 +1408,9 @@ mod tests {
    }
    #[test]
-    fn take_from_foundation_allowed_by_default() {
+    fn take_from_foundation_disabled_by_default() {
-        let mut g = setup_take_from_foundation_game();
+        let g = setup_take_from_foundation_game();
-        assert!(g.take_from_foundation, "standard Klondike allows take-from-foundation by default");
+        assert!(!g.take_from_foundation, "take_from_foundation is off by default (non-standard rule)");
        g.move_cards(PileType::Foundation(0), PileType::Tableau(0), 1).unwrap();
        assert_eq!(g.piles[&PileType::Foundation(0)].cards.len(), 1);
        assert_eq!(g.piles[&PileType::Tableau(0)].cards.len(), 2);
    }
    #[test]
@@ -32,7 +32,7 @@ use crate::settings_plugin::{SettingsResource, SettingsStoragePath};
 use crate::stats_plugin::{StatsResource, StatsUpdate};
 use crate::ui_modal::{
    spawn_modal, spawn_modal_actions, spawn_modal_button, spawn_modal_header, ButtonVariant,
-    ScrimDismissible,
+    ModalScrim, ScrimDismissible,
 };
 use crate::ui_theme::{
    ACCENT_PRIMARY, BORDER_SUBTLE, STATE_SUCCESS, TEXT_DISABLED, TEXT_PRIMARY, TEXT_SECONDARY,
@@ -162,93 +162,91 @@ fn evaluate_on_win(
    mut achievements: ResMut<AchievementsResource>,
    mut progress: ResMut<ProgressResource>,
 ) {
-    let Some(ev) = wins.read().last() else {
+    for ev in wins.read() {
-        return;
+        let ctx = AchievementContext {
-    };
+            games_played: stats.0.games_played,
-
+            games_won: stats.0.games_won,
-    let ctx = AchievementContext {
+            win_streak_current: stats.0.win_streak_current,
-        games_played: stats.0.games_played,
+            best_single_score: stats.0.best_single_score,
-        games_won: stats.0.games_won,
+            lifetime_score: stats.0.lifetime_score,
-        win_streak_current: stats.0.win_streak_current,
+            draw_three_wins: stats.0.draw_three_wins,
-        best_single_score: stats.0.best_single_score,
+            daily_challenge_streak: progress.0.daily_challenge_streak,
-        lifetime_score: stats.0.lifetime_score,
+            last_win_score: ev.score,
-        draw_three_wins: stats.0.draw_three_wins,
+            last_win_time_seconds: ev.time_seconds,
-        daily_challenge_streak: progress.0.daily_challenge_streak,
+            last_win_used_undo: game.0.undo_count > 0,
-        last_win_score: ev.score,
+            wall_clock_hour: Some(Local::now().hour()),
-        last_win_time_seconds: ev.time_seconds,
+            last_win_recycle_count: game.0.recycle_count,
-        last_win_used_undo: game.0.undo_count > 0,
+            last_win_is_zen: game.0.mode == solitaire_core::game_state::GameMode::Zen,
        wall_clock_hour: Some(Local::now().hour()),
        last_win_recycle_count: game.0.recycle_count,
        last_win_is_zen: game.0.mode == solitaire_core::game_state::GameMode::Zen,
    };
    let hits = check_achievements(&ctx);
    if hits.is_empty() {
        return;
    }
    let now = Utc::now();
    let mut achievements_changed = false;
    let mut progress_changed = false;
    for def in hits {
        let Some(record) = achievements.0.iter_mut().find(|r| r.id == def.id) else {
            continue;
        };
-        if record.unlocked {
+
        let hits = check_achievements(&ctx);
        if hits.is_empty() {
            continue;
        }
        record.unlock(now);
        achievements_changed = true;
-        // Grant the reward on first unlock.
+        let now = Utc::now();
-        if !record.reward_granted {
+        let mut achievements_changed = false;
-            if let Some(reward) = def.reward {
+        let mut progress_changed = false;
-                match reward {
+
-                    Reward::CardBack(idx) => {
+        for def in hits {
-                        if !progress.0.unlocked_card_backs.contains(&idx) {
+            let Some(record) = achievements.0.iter_mut().find(|r| r.id == def.id) else {
-                            progress.0.unlocked_card_backs.push(idx);
+                continue;
-                            progress_changed = true;
+            };
-                        }
+            if record.unlocked {
-                    }
+                continue;
                    Reward::Background(idx) => {
                        if !progress.0.unlocked_backgrounds.contains(&idx) {
                            progress.0.unlocked_backgrounds.push(idx);
                            progress_changed = true;
                        }
                    }
                    Reward::BonusXp(amount) => {
                        xp_awarded.write(XpAwardedEvent { amount });
                        let prev_level = progress.0.add_xp(amount);
                        if progress.0.leveled_up_from(prev_level) {
                            levelups.write(LevelUpEvent {
                                previous_level: prev_level,
                                new_level: progress.0.level,
                                total_xp: progress.0.total_xp,
                            });
                        }
                        progress_changed = true;
                    }
                    Reward::Badge => {}
                }
            }
-            record.reward_granted = true;
+            record.unlock(now);
            achievements_changed = true;
            // Grant the reward on first unlock.
            if !record.reward_granted {
                if let Some(reward) = def.reward {
                    match reward {
                        Reward::CardBack(idx) => {
                            if !progress.0.unlocked_card_backs.contains(&idx) {
                                progress.0.unlocked_card_backs.push(idx);
                                progress_changed = true;
                            }
                        }
                        Reward::Background(idx) => {
                            if !progress.0.unlocked_backgrounds.contains(&idx) {
                                progress.0.unlocked_backgrounds.push(idx);
                                progress_changed = true;
                            }
                        }
                        Reward::BonusXp(amount) => {
                            xp_awarded.write(XpAwardedEvent { amount });
                            let prev_level = progress.0.add_xp(amount);
                            if progress.0.leveled_up_from(prev_level) {
                                levelups.write(LevelUpEvent {
                                    previous_level: prev_level,
                                    new_level: progress.0.level,
                                    total_xp: progress.0.total_xp,
                                });
                            }
                            progress_changed = true;
                        }
                        Reward::Badge => {}
                    }
                }
                record.reward_granted = true;
            }
            unlocks.write(AchievementUnlockedEvent(record.clone()));
        }
-        unlocks.write(AchievementUnlockedEvent(record.clone()));
+        if achievements_changed
            && let Some(target) = &path.0
                && let Err(e) = save_achievements_to(target, &achievements.0) {
                    warn!("failed to save achievements: {e}");
                }
        if progress_changed
            && let Some(target) = &progress_path.0
                && let Err(e) = save_progress_to(target, &progress.0) {
                    warn!("failed to save progress after reward: {e}");
                }
    }
    if achievements_changed
        && let Some(target) = &path.0
            && let Err(e) = save_achievements_to(target, &achievements.0) {
                warn!("failed to save achievements: {e}");
            }
    if progress_changed
        && let Some(target) = &progress_path.0
            && let Err(e) = save_progress_to(target, &progress.0) {
                warn!("failed to save progress after reward: {e}");
            }
 }
 /// Cinephile unlock observer.
@@ -391,6 +389,7 @@ fn toggle_achievements_screen(
    achievements: Res<AchievementsResource>,
    font_res: Option<Res<FontResource>>,
    screens: Query<Entity, With<AchievementsScreen>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<AchievementsScreen>)>,
 ) {
    let button_clicked = requests.read().count() > 0;
    if !keys.just_pressed(KeyCode::KeyA) && !button_clicked {
@@ -398,7 +397,7 @@ fn toggle_achievements_screen(
    }
    if let Ok(entity) = screens.single() {
        commands.entity(entity).despawn();
-    } else {
+    } else if other_modal_scrims.is_empty() {
        spawn_achievements_screen(&mut commands, &achievements.0, font_res.as_deref());
    }
 }
@@ -72,6 +72,17 @@ const TIME_ATTACK_TOAST_SECS: f32 = 5.0;
 const CHALLENGE_TOAST_SECS: f32 = 3.0;
 const VOLUME_TOAST_SECS: f32 = 1.4;
 /// Z added to a card's render depth while its `CardAnim` is in-flight.
 ///
 /// Foundation and tableau cards share x,y during the slide (destination equals
 /// a slot that already holds a card). Without this lift the incoming card's
 /// bottom-right corner overlaps the stationary card's top-left, which the
 /// player perceives as a single card with mismatched rank/suit indices.
 ///
 /// 50.0 sits comfortably above the highest pile depth (~1.04) and well below
 /// `DRAG_Z` (500), so a dragged card always renders above an animated one.
 const CARD_ANIM_Z_LIFT: f32 = 50.0;
 /// Per-card stagger interval for the win cascade at Normal speed (seconds).
 ///
 /// Sourced from `ui_theme::MOTION_CASCADE_STAGGER_SECS` so all motion timing
@@ -254,7 +265,11 @@ fn advance_card_anims(
        // shared `CardAnim` struct stays a simple linear-tween container — the
        // upgrade is one extra `sample_curve` call per advancing animation.
        let s = sample_curve(MotionCurve::SmoothSnap, t);
-        transform.translation = anim.start.lerp(anim.target, s);
+        let mut pos = anim.start.lerp(anim.target, s);
        // Elevate z during transit so the moving card always renders in front
        // of any card already resting at the destination position.
        pos.z = anim.target.z + CARD_ANIM_Z_LIFT;
        transform.translation = pos;
        if t >= 1.0 {
            transform.translation = anim.target;
            commands.entity(entity).remove::<CardAnim>();
@@ -72,9 +72,14 @@ fn detect_auto_complete(
    if game.0.is_auto_completable && !state.active {
        state.active = true;
        state.cooldown = 0.0; // fire first move immediately
    } else if !game.0.is_auto_completable {
        state.active = false;
    }
    // Intentionally no `else if !is_auto_completable` branch here.
    // Deactivating on every frame where `is_auto_completable` is false
    // would hard-stop the sequence mid-flight whenever `next_auto_complete_move`
    // transiently returns `None` (e.g. while the previous move is still
    // in-flight). The `is_won` check above already handles the definitive
    // end-of-game case; `drive_auto_complete` simply retries next tick
    // when no move is available yet.
 }
 /// Plays a distinct chime the moment auto-complete first activates.
@@ -142,6 +142,13 @@ impl Plugin for CardAnimationPlugin {
                    update_frame_time_diagnostics,
                    // Advance active animations.
                    advance_card_animations,
                    // Flush deferred commands so `CardAnimation` removals from
                    // `advance_card_animations` are visible before the chain
                    // system runs. Without this, the chain sees the component
                    // still present in the same frame it was removed (deferred
                    // commands aren't applied until the next ApplyDeferred
                    // point), causing a 1-frame gap between every chain step.
                    ApplyDeferred,
                    // After each animation finishes, pop the next chain segment.
                    advance_animation_chains,
                    // Interaction visuals (run after animation for final positions).
@@ -178,8 +178,8 @@ pub struct CardLabel;
 /// readable at phone scale. Only exists when `CardImageSet` is present
 /// (the fallback solid-colour path uses a plain `CardLabel` instead).
 #[cfg(target_os = "android")]
-#[derive(Component, Debug, Clone, Copy)]
+#[derive(Component, Debug, Clone)]
-struct AndroidCornerLabel;
+struct AndroidCornerLabel(pub String);
 /// Solid-colour background sprite behind [`AndroidCornerLabel`].
 ///
@@ -707,15 +707,20 @@ fn sync_cards(
            .map(|c| c.id)
    };
-    // Map card_id -> (Entity, current_translation, has_card_animation) for
+    // Map card_id -> (Entity, current_translation, anim_end) for in-place
-    // in-place updates. The `has_card_animation` flag lets `update_card_entity`
+    // updates.  `anim_end` is `Some(end_xy)` when a curve-based `CardAnimation`
-    // skip the snap/slide path on cards that are already being driven by a
+    // is currently driving the card (e.g. a drag-rejection return tween).
-    // curve-based `CardAnimation` tween (e.g. the drag-rejection return tween
+    //
-    // — see `input_plugin::end_drag`). Otherwise the StateChangedEvent that
+    // In the position loop below we compare `anim_end` against the new game-
-    // accompanies a rejection would race the tween and the card would jump.
+    // state target position to decide whether to honour or cancel the tween:
-    let mut existing: HashMap<u32, (Entity, Vec3, bool)> = HashMap::new();
+    //   • end ≈ target  → animation is still heading to the right place; let
    //                      it finish (skip the snap/slide path).
    //   • end ≠ target  → the game state has changed (e.g. a new game started
    //                      while the win-cascade was mid-flight); cancel the
    //                      stale `CardAnimation` and apply the new position.
    let mut existing: HashMap<u32, (Entity, Vec3, Option<Vec2>)> = HashMap::new();
    for (entity, marker, transform, anim) in entities.iter() {
-        existing.insert(marker.card_id, (entity, transform.translation, anim.is_some()));
+        existing.insert(marker.card_id, (entity, transform.translation, anim.map(|a| a.end)));
    }
    let live_ids: HashSet<u32> = positions.iter().map(|(c, _, _)| c.id).collect();
@@ -732,7 +737,19 @@ fn sync_cards(
    // behind the incoming top card during the draw slide animation.
    for (card, position, z) in positions {
        let entity = match existing.get(&card.id) {
-            Some(&(entity, cur, has_anim)) => {
+            Some(&(entity, cur, anim_end)) => {
                // If a CardAnimation is in flight, check whether its destination
                // still matches the game-state target.  If the game moved the card
                // elsewhere (e.g. new game started during a win-cascade scatter),
                // cancel the stale tween so the card snaps/slides to its new home.
                let has_anim = match anim_end {
                    Some(end_xy) if (end_xy - position).length() > 2.0 => {
                        commands.entity(entity).remove::<CardAnimation>();
                        false
                    }
                    Some(_) => true,
                    None => false,
                };
                update_card_entity(
                    &mut commands, entity, card, position, z, layout,
                    slide_secs, back_colour, color_blind, high_contrast, cur, has_anim, card_images, selected_back, font_handle,
@@ -1142,10 +1159,11 @@ fn add_android_corner_label(
    // Large rank+suit text drawn on top of the background. FiraMono must be
    // wired here explicitly — the suit glyphs (U+2660–U+2666) are not in
    // Bevy's built-in font and render as a coloured rectangle without it.
    let label_text = mobile_label_for(card);
    parent.spawn((
-        AndroidCornerLabel,
+        AndroidCornerLabel(label_text.clone()),
        CardLabel,
-        Text2d::new(mobile_label_for(card)),
+        Text2d::new(label_text),
        TextFont {
            font: font_handle.cloned().unwrap_or_default(),
            font_size,
@@ -2089,7 +2107,7 @@ fn resize_cards_in_place(
 fn resize_android_corner_labels(
    layout: Res<LayoutResource>,
    card_images: Option<Res<CardImageSet>>,
-    mut text_query: Query<(&mut TextFont, &mut Transform), With<AndroidCornerLabel>>,
+    mut text_query: Query<(&AndroidCornerLabel, &mut Text2d, &mut TextFont, &mut Transform)>,
    mut bg_query: Query<
        (&mut Sprite, &mut Transform),
        (With<AndroidCornerBg>, Without<AndroidCornerLabel>),
@@ -2105,7 +2123,8 @@ fn resize_android_corner_labels(
    let text_x = -layout.0.card_size.x / 2.0 + inset;
    let text_y = layout.0.card_size.y / 2.0 - inset;
-    for (mut font, mut transform) in text_query.iter_mut() {
+    for (label, mut text2d, mut font, mut transform) in text_query.iter_mut() {
        text2d.0 = label.0.clone();
        font.font_size = font_size;
        transform.translation.x = text_x;
        transform.translation.y = text_y;
@@ -382,8 +382,8 @@ fn update_drop_target_overlays(
 /// for everything else it is card-sized. Replicated here rather than
 /// imported because `pile_drop_rect` is private to `input_plugin` and
 /// this overlay is the only other consumer.
-fn drop_overlay_rect(pile: &PileType, layout: &Layout, game: &GameState) -> (Vec2, Vec2) {
+fn drop_overlay_rect(pile: &PileType, layout: &Layout, game: &GameState) -> Option<(Vec2, Vec2)> {
-    let centre = layout.pile_positions[pile];
+    let centre = layout.pile_positions.get(pile).copied()?;
    if matches!(pile, PileType::Tableau(_)) {
        let card_count = game.piles.get(pile).map_or(0, |p| p.cards.len());
        if card_count > 1 {
@@ -393,13 +393,13 @@ fn drop_overlay_rect(pile: &PileType, layout: &Layout, game: &GameState) -> (Vec
            let bottom_edge = bottom_card_centre_y - layout.card_size.y / 2.0;
            let span_height = top_edge - bottom_edge;
            let new_centre_y = (top_edge + bottom_edge) / 2.0;
-            return (
+            return Some((
                Vec2::new(centre.x, new_centre_y),
                Vec2::new(layout.card_size.x, span_height),
-            );
+            ));
        }
    }
-    (centre, layout.card_size)
+    Some((centre, layout.card_size))
 }
 /// Spawns one overlay parent (fill) plus four edge sprites (outline) at
@@ -410,7 +410,10 @@ fn spawn_drop_target_overlay(
    layout: &Layout,
    game: &GameState,
 ) {
-    let (centre, size) = drop_overlay_rect(pile, layout, game);
+    let Some((centre, size)) = drop_overlay_rect(pile, layout, game) else {
        warn!("drop_overlay_rect: pile {pile:?} not in layout, skipping overlay");
        return;
    };
    let edge = DROP_TARGET_OUTLINE_PX;
    commands
@@ -210,10 +210,15 @@ impl Plugin for FeedbackAnimPlugin {
                    start_shake_anim.after(GameMutation),
                    tick_shake_anim,
                    start_settle_anim.after(GameMutation),
                    // tick_foundation_flourish writes the full Transform.scale
                    // (Vec3); tick_settle_anim writes only scale.y on top of
                    // it. Ordering ensures the settle's y-only write always
                    // applies last so it wins on the ~0.15 s overlap when both
                    // components are present on the same King entity.
                    tick_foundation_flourish.before(tick_settle_anim),
                    tick_settle_anim,
                    start_deal_anim.after(GameMutation),
                    start_foundation_flourish.after(GameMutation),
                    tick_foundation_flourish,
                ),
            );
    }
@@ -31,8 +31,15 @@ fn load_font(fonts: Option<ResMut<Assets<Font>>>, mut commands: Commands) {
    // Assets<Font>). FontPlugin in that context is a no-op — consumers
    // already query `Option<Res<FontResource>>` and degrade cleanly.
    let Some(mut fonts) = fonts else { return };
-    let font = Font::try_from_bytes(BUNDLED_FONT_BYTES.to_vec())
+    let font = match Font::try_from_bytes(BUNDLED_FONT_BYTES.to_vec()) {
-        .expect("bundled FiraMono failed to parse — binary is corrupt");
+        Ok(f) => f,
        Err(e) => {
            // A corrupt embedded font is unusual but should not crash the
            // process — UI will render without glyphs rather than panicking.
            warn!("bundled FiraMono failed to parse ({e}); UI text may be invisible");
            return;
        }
    };
    let handle = fonts.add(font);
    commands.insert_resource(FontResource(handle));
 }
@@ -13,7 +13,7 @@ use crate::font_plugin::FontResource;
 use crate::hud_plugin::ANDROID_HINT_LABEL;
 use crate::ui_modal::{
    spawn_modal, spawn_modal_actions, spawn_modal_button, spawn_modal_header, ButtonVariant,
-    ScrimDismissible,
+    ModalScrim, ScrimDismissible,
 };
 use crate::ui_theme::{SPACE_2, TEXT_PRIMARY, TEXT_SECONDARY, TYPE_BODY, VAL_SPACE_2, VAL_SPACE_3, Z_MODAL_PANEL};
 #[cfg(not(target_os = "android"))]
@@ -67,6 +67,7 @@ fn toggle_help_screen(
    keys: Res<ButtonInput<KeyCode>>,
    mut requests: MessageReader<HelpRequestEvent>,
    screens: Query<Entity, With<HelpScreen>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<HelpScreen>)>,
    font_res: Option<Res<FontResource>>,
 ) {
    // Either F1 or a click on the HUD "Help" button (which fires
@@ -77,7 +78,7 @@ fn toggle_help_screen(
    }
    if let Ok(entity) = screens.single() {
        commands.entity(entity).despawn();
-    } else {
+    } else if other_modal_scrims.is_empty() {
        spawn_help_screen(&mut commands, font_res.as_deref());
    }
 }
@@ -45,7 +45,7 @@ use crate::layout::LayoutSystem;
 use crate::pause_plugin::PausedResource;
 use crate::resources::GameStateResource;
 #[cfg(target_os = "android")]
-use crate::resources::DragState;
+use crate::resources::{DragState, GameInputConsumedResource};
 use crate::selection_plugin::SelectionState;
 use crate::time_attack_plugin::TimeAttackResource;
 use crate::ui_focus::{FocusGroup, Focusable};
@@ -870,12 +870,12 @@ fn spawn_action_buttons(
    );
    #[cfg(not(target_os = "android"))]
    let labels = (
-        "Menu \u{25BE}",
+        "Menu \u{2193}",
        "Undo",
        "Pause",
        "Help",
        "Hint",
-        "Modes \u{25BE}",
+        "Modes \u{2193}",
        "New Game",
    );
@@ -2492,6 +2492,7 @@ fn toggle_hud_on_tap(
    mut tracker: ResMut<HudTapTracker>,
    mut hud_vis: ResMut<HudVisibility>,
    buttons: Query<&Interaction, With<ActionButton>>,
    mut game_consumed: ResMut<GameInputConsumedResource>,
 ) {
    use bevy::input::touch::TouchPhase;
    if !scrims.is_empty() || paused.is_some_and(|p| p.0) {
@@ -2502,6 +2503,7 @@ fn toggle_hud_on_tap(
        for _ in touch_events.read() {}
        tracker.start_pos = None;
        tracker.started_on_button = false;
        game_consumed.0 = false;
        return;
    }
    for event in touch_events.read() {
@@ -2515,7 +2517,13 @@ fn toggle_hud_on_tap(
                    buttons.iter().any(|i| *i != Interaction::None);
            }
            TouchPhase::Ended if drag.is_idle() => {
-                let on_button = tracker.started_on_button;
+                // Also treat taps where game logic consumed the touch (e.g.
                // drawing from stock) as "on button" so they don't toggle
                // the HUD. The flag is set on TouchPhase::Started by the
                // input system that consumed the tap and must be cleared here
                // regardless of whether we toggle.
                let on_button = tracker.started_on_button || game_consumed.0;
                game_consumed.0 = false;
                if let Some(start) = tracker.start_pos.take() {
                    if !on_button && (event.position - start).length() < HUD_TAP_SLOP_PX {
                        *hud_vis = match *hud_vis {
@@ -2532,6 +2540,7 @@ fn toggle_hud_on_tap(
            TouchPhase::Canceled => {
                tracker.start_pos = None;
                tracker.started_on_button = false;
                game_consumed.0 = false;
            }
            _ => {}
        }
@@ -47,7 +47,8 @@ use crate::game_plugin::{ConfirmNewGameScreen, GameMutation, RestorePromptScreen
 use crate::pause_plugin::PausedResource;
 use crate::progress_plugin::ProgressResource;
 use crate::layout::{Layout, LayoutResource};
-use crate::resources::{DragState, GameStateResource, HintCycleIndex};
+use crate::replay_playback::ReplayPlaybackState;
 use crate::resources::{DragState, GameInputConsumedResource, GameStateResource, HintCycleIndex};
 use crate::selection_plugin::SelectionState;
 use crate::time_attack_plugin::TimeAttackResource;
@@ -95,6 +96,7 @@ impl Plugin for InputPlugin {
        app.init_resource::<HintCycleIndex>()
            .init_resource::<HintSolverConfig>()
            .init_resource::<crate::pending_hint::PendingHintTask>()
            .init_resource::<GameInputConsumedResource>()
            .add_message::<StartZenRequestEvent>()
            .add_message::<InfoToastEvent>()
            .add_message::<ForfeitRequestEvent>()
@@ -174,11 +176,20 @@ fn handle_keyboard_core(
    mut zen_requests: MessageReader<StartZenRequestEvent>,
    confirm_screens: Query<(), With<ConfirmNewGameScreen>>,
    restore_prompts: Query<(), With<RestorePromptScreen>>,
    replay_state: Option<Res<ReplayPlaybackState>>,
 ) {
    if paused.is_some_and(|p| p.0) {
        return;
    }
    // During replay playback (Playing or Completed) all game-input shortcuts
    // are suppressed. The replay overlay owns Space (pause/resume) and the
    // arrow keys (step). Letting game input through would mutate
    // `GameStateResource` and corrupt replay determinism.
    if replay_state.is_some_and(|r| !matches!(*r, ReplayPlaybackState::Inactive)) {
        return;
    }
    if keys.just_pressed(KeyCode::KeyU) {
        ev.undo.write(UndoRequestEvent);
    }
@@ -501,6 +512,7 @@ fn handle_touch_stock_tap(
    layout: Option<Res<LayoutResource>>,
    drag: Res<DragState>,
    mut draw: MessageWriter<DrawRequestEvent>,
    mut game_consumed: ResMut<GameInputConsumedResource>,
 ) {
    if paused.is_some_and(|p| p.0) {
        return;
@@ -522,6 +534,7 @@ fn handle_touch_stock_tap(
        };
        if point_in_rect(world, stock_pos, layout.0.card_size) {
            draw.write(DrawRequestEvent);
            game_consumed.0 = true;
            break; // one draw per tap frame
        }
    }
@@ -21,7 +21,7 @@ use crate::settings_plugin::{SettingsResource, SettingsStoragePath};
 use crate::sync_plugin::SyncProviderResource;
 use crate::ui_modal::{
    spawn_modal, spawn_modal_actions, spawn_modal_button, spawn_modal_header, ButtonVariant,
-    ScrimDismissible,
+    ModalScrim, ScrimDismissible,
 };
 use crate::ui_theme::{
    ACCENT_PRIMARY, BG_ELEVATED, BORDER_SUBTLE, RADIUS_SM, STATE_INFO,
@@ -714,6 +714,7 @@ fn data_cell(
 fn handle_set_display_name_button(
    button_q: Query<&Interaction, (Changed<Interaction>, With<SetDisplayNameButton>)>,
    existing: Query<(), With<DisplayNameModal>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<DisplayNameModal>)>,
    mut commands: Commands,
    settings: Option<Res<SettingsResource>>,
    font_res: Option<Res<FontResource>>,
@@ -725,6 +726,9 @@ fn handle_set_display_name_button(
    if !existing.is_empty() {
        return; // already open
    }
    if !other_modal_scrims.is_empty() {
        return; // Another modal is already visible.
    }
    buf.0 = settings
        .as_ref()
        .and_then(|s| s.0.leaderboard_display_name.clone())
@@ -1159,9 +1163,23 @@ mod tests {
            .spawn(async { Err::<(), String>("network error".to_string()) });
        app.world_mut().resource_mut::<OptInTask>().0 = Some(failed_task);
-        // Allow the task to complete and be polled.
+        // Pump until the task is polled or a deadline elapses.  A fixed
-        for _ in 0..5 {
+        // update count is unreliable under parallel `cargo test --workspace`
        // load — the AsyncComputeTaskPool background threads can be starved
        // long enough that 5 updates finish before the task completes.
        // Mirrors the deadline-loop pattern used in sync_plugin tests.
        let deadline = std::time::Instant::now() + std::time::Duration::from_secs(5);
        loop {
            app.update();
            let msgs = app.world().resource::<Messages<WarningToastEvent>>();
            let mut cursor = msgs.get_cursor();
            if cursor.read(msgs).next().is_some() {
                break;
            }
            if std::time::Instant::now() >= deadline {
                break;
            }
            std::thread::yield_now();
        }
        let msgs = app.world().resource::<Messages<WarningToastEvent>>();
@@ -1183,8 +1201,19 @@ mod tests {
            .spawn(async { Err::<(), String>("network error".to_string()) });
        app.world_mut().resource_mut::<OptOutTask>().0 = Some(failed_task);
-        for _ in 0..5 {
+        // Deadline-bounded pump — see opt_in_error_fires_warning_toast for rationale.
        let deadline = std::time::Instant::now() + std::time::Duration::from_secs(5);
        loop {
            app.update();
            let msgs = app.world().resource::<Messages<WarningToastEvent>>();
            let mut cursor = msgs.get_cursor();
            if cursor.read(msgs).next().is_some() {
                break;
            }
            if std::time::Instant::now() >= deadline {
                break;
            }
            std::thread::yield_now();
        }
        let msgs = app.world().resource::<Messages<WarningToastEvent>>();
@@ -1210,8 +1239,22 @@ mod tests {
        let ok_task = AsyncComputeTaskPool::get().spawn(async { Ok::<(), String>(()) });
        app.world_mut().resource_mut::<OptInTask>().0 = Some(ok_task);
-        for _ in 0..5 {
+        // Deadline-bounded pump — see opt_in_error_fires_warning_toast for rationale.
        let deadline = std::time::Instant::now() + std::time::Duration::from_secs(5);
        loop {
            app.update();
            if app
                .world()
                .resource::<SettingsResource>()
                .0
                .leaderboard_opted_in
            {
                break;
            }
            if std::time::Instant::now() >= deadline {
                break;
            }
            std::thread::yield_now();
        }
        assert!(
@@ -1237,8 +1280,22 @@ mod tests {
        let ok_task = AsyncComputeTaskPool::get().spawn(async { Ok::<(), String>(()) });
        app.world_mut().resource_mut::<OptOutTask>().0 = Some(ok_task);
-        for _ in 0..5 {
+        // Deadline-bounded pump — see opt_in_error_fires_warning_toast for rationale.
        let deadline = std::time::Instant::now() + std::time::Duration::from_secs(5);
        loop {
            app.update();
            if !app
                .world()
                .resource::<SettingsResource>()
                .0
                .leaderboard_opted_in
            {
                break;
            }
            if std::time::Instant::now() >= deadline {
                break;
            }
            std::thread::yield_now();
        }
        assert!(
@@ -411,7 +411,11 @@ fn handle_confirm(
    new_game.write(NewGameRequestEvent {
        seed: Some(seed),
        mode: None,
-        confirmed: false,
+        // The player explicitly clicked Play (or pressed Enter) after typing
        // a seed — treat this as an affirmative confirmation so the
        // abandon-current-game dialog is not shown on top of the already-
        // dismissed seed dialog.
        confirmed: true,
    });
    for entity in &screen {
@@ -566,7 +570,9 @@ mod tests {
        assert_eq!(fired.len(), 1);
        assert_eq!(fired[0].seed, Some(42));
        assert_eq!(fired[0].mode, None);
-        assert!(!fired[0].confirmed);
+        // confirmed: true — the player explicitly clicked Play, so no
        // abandon-current-game dialog should appear.
        assert!(fired[0].confirmed);
        // Dialog should be gone.
        assert!(!dialog_present(&mut app));
@@ -28,7 +28,7 @@ use chrono::Datelike;
 use crate::font_plugin::FontResource;
 use crate::layout::LayoutResource;
-use crate::events::{DrawRequestEvent, MoveRequestEvent, UndoRequestEvent};
+use crate::events::{DrawRequestEvent, MoveRequestEvent, StateChangedEvent, UndoRequestEvent};
 use crate::replay_playback::{
    step_backwards_replay_playback, step_replay_playback, stop_replay_playback,
    toggle_pause_replay_playback, ReplayPlaybackState,
@@ -476,6 +476,7 @@ impl Plugin for ReplayOverlayPlugin {
            .add_message::<MoveRequestEvent>()
            .add_message::<DrawRequestEvent>()
            .add_message::<UndoRequestEvent>()
            .add_message::<StateChangedEvent>()
            .add_systems(
                Update,
                (
@@ -1884,6 +1885,7 @@ fn handle_pause_keyboard(
 /// resets to 0 on key release so the next fresh press fires
 /// immediately. This matches the mockup's `[← →] scrub`
 /// terminology while keeping single-press = single-step semantics.
 #[allow(clippy::too_many_arguments)]
 fn handle_arrow_keyboard(
    keys: Option<Res<ButtonInput<KeyCode>>>,
    time: Res<Time>,
@@ -1892,10 +1894,22 @@ fn handle_arrow_keyboard(
    mut moves_writer: MessageWriter<MoveRequestEvent>,
    mut draws_writer: MessageWriter<DrawRequestEvent>,
    mut undo_writer: MessageWriter<UndoRequestEvent>,
    mut state_changed: MessageReader<StateChangedEvent>,
    // `true` while a backward step is in-flight: cursor was decremented and
    // `UndoRequestEvent` was written, but `handle_undo` hasn't applied it yet.
    // Cleared when `StateChangedEvent` confirms the game state has caught up.
    // Prevents rapid ← presses from accumulating multiple cursor decrements
    // before any undo is applied (Bug #16).
    mut back_pending: Local<bool>,
 ) {
    let Some(keys) = keys else { return };
    let dt = time.delta_secs();
    // Clear the in-flight flag once the game confirms the undo landed.
    if state_changed.read().count() > 0 {
        *back_pending = false;
    }
    // Right (forward step) — initial press fires immediately;
    // held repeats fire when the accumulator crosses the interval.
    if keys.just_pressed(KeyCode::ArrowRight) {
@@ -1911,14 +1925,28 @@ fn handle_arrow_keyboard(
        hold.right_held_secs = 0.0;
    }
-    // Left (backwards step) — symmetric to the right path.
+    // Left (backwards step) — gate on `back_pending` so at most one undo
    // is in-flight at a time. The cursor is only decremented inside
    // `step_backwards_replay_playback`, which also writes `UndoRequestEvent`.
    // `back_pending` is set after a successful step and cleared above when
    // `StateChangedEvent` confirms the undo was applied.
    if keys.just_pressed(KeyCode::ArrowLeft) {
-        step_backwards_replay_playback(&mut state, &mut undo_writer);
+        if !*back_pending {
            let fired = step_backwards_replay_playback(&mut state, &mut undo_writer);
            if fired {
                *back_pending = true;
            }
        }
        hold.left_held_secs = 0.0;
    } else if keys.pressed(KeyCode::ArrowLeft) {
        hold.left_held_secs += dt;
        if hold.left_held_secs >= SCRUB_REPEAT_INTERVAL_SECS {
-            step_backwards_replay_playback(&mut state, &mut undo_writer);
+            if !*back_pending {
                let fired = step_backwards_replay_playback(&mut state, &mut undo_writer);
                if fired {
                    *back_pending = true;
                }
            }
            hold.left_held_secs = 0.0;
        }
    } else {
@@ -3,7 +3,7 @@
 use std::sync::Arc;
 use bevy::math::Vec2;
-use bevy::prelude::Resource;
+use bevy::prelude::{warn, Resource};
 use chrono::{DateTime, Utc};
 use solitaire_core::game_state::GameState;
 use solitaire_core::pile::PileType;
@@ -114,6 +114,13 @@ pub struct HintCycleIndex(pub usize);
 #[derive(Resource, Debug, Clone, Default)]
 pub struct SettingsScrollPos(pub f32);
 /// Set to `true` by an input system when a touch tap is consumed by game logic
 /// (e.g. drawing from stock). `toggle_hud_on_tap` checks this flag on
 /// `TouchPhase::Ended` and skips the HUD visibility toggle when set, then
 /// resets it to `false` so subsequent taps behave normally.
 #[derive(Resource, Debug, Clone, Default)]
 pub struct GameInputConsumedResource(pub bool);
 /// Shared Tokio runtime used by all async-task closures that need HTTP I/O.
 ///
 /// Bevy's `AsyncComputeTaskPool` uses `async-executor` (not Tokio), so spawned
@@ -124,15 +131,48 @@ pub struct SettingsScrollPos(pub f32);
 #[derive(Resource, Clone)]
 pub struct TokioRuntimeResource(pub Arc<tokio::runtime::Runtime>);
-impl Default for TokioRuntimeResource {
+impl TokioRuntimeResource {
-    fn default() -> Self {
+    /// Attempts to build the shared multi-threaded Tokio runtime.
-        // Building the Tokio runtime is startup-time initialization; failure
+    ///
-        // here means the OS refused to create threads, which is unrecoverable.
+    /// Returns `Err` if the OS refuses to create worker threads (e.g. resource
    /// limits on Android). Callers should log the error and disable sync
    /// features rather than panicking.
    pub fn new() -> Result<Self, tokio::io::Error> {
        let rt = tokio::runtime::Builder::new_multi_thread()
            .worker_threads(2)
            .enable_all()
-            .build()
+            .build()?;
-            .expect("failed to build shared Tokio runtime");
+        Ok(Self(Arc::new(rt)))
-        Self(Arc::new(rt))
+    }
 }
 impl Default for TokioRuntimeResource {
    fn default() -> Self {
        // Try multi-threaded first; fall back to current-thread (single
        // worker) if the OS refuses to create additional threads. Neither
        // path uses `.expect()` so this never panics at startup.
        match tokio::runtime::Builder::new_multi_thread()
            .worker_threads(2)
            .enable_all()
            .build()
        {
            Ok(rt) => Self(Arc::new(rt)),
            Err(e) => {
                warn!(
                    "sync: failed to build multi-thread Tokio runtime ({e}); \
                     falling back to current-thread runtime"
                );
                // current_thread runtime never spawns OS threads, so it
                // succeeds even under tight sandboxing.
                let rt = tokio::runtime::Builder::new_current_thread()
                    .enable_all()
                    .build()
                    .expect(
                        "current-thread Tokio runtime failed — \
                         the process cannot do any async I/O",
                    );
                Self(Arc::new(rt))
            }
        }
    }
 }
@@ -510,6 +510,7 @@ fn toggle_settings_screen(
 fn sync_settings_panel_visibility(
    screen: Res<SettingsScreen>,
    panels: Query<Entity, With<SettingsPanel>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<SettingsPanel>)>,
    scroll_nodes: Query<&ScrollPosition, With<SettingsScrollNode>>,
    mut scroll_pos: ResMut<SettingsScrollPos>,
    mut commands: Commands,
@@ -525,7 +526,7 @@ fn sync_settings_panel_visibility(
        return;
    }
    if screen.0 {
-        if panels.is_empty() {
+        if panels.is_empty() && other_modal_scrims.is_empty() {
            let status_label = sync_status
                .map_or_else(|| "Status: local only".to_string(), |s| sync_status_label(&s.0));
            let unlocked_backs = progress
@@ -220,7 +220,13 @@ impl Plugin for StatsPlugin {
            )
            .add_systems(
                Update,
-                handle_forfeit.before(GameMutation),
+                // handle_forfeit must run before update_stats_on_new_game so
                // the NewGameRequestEvent it emits is not visible to
                // update_stats_on_new_game in the same frame — otherwise
                // record_abandoned() fires twice on every forfeit (#21).
                handle_forfeit
                    .before(GameMutation)
                    .before(update_stats_on_new_game),
            )
            .add_systems(Update, toggle_stats_screen.after(GameMutation))
            .add_systems(Update, handle_stats_close_button)
@@ -101,7 +101,6 @@ impl SyncPlugin {
 impl Plugin for SyncPlugin {
    fn build(&self, app: &mut App) {
        app.insert_resource(SyncProviderResource(self.provider.clone()))
            .init_resource::<TokioRuntimeResource>()
            .init_resource::<SyncStatusResource>()
            .init_resource::<PullTaskResult>()
            .init_resource::<PullTask>()
@@ -109,18 +108,30 @@ impl Plugin for SyncPlugin {
            .add_message::<ManualSyncRequestEvent>()
            .add_message::<SyncCompleteEvent>()
            .add_message::<SyncConfigureRequestEvent>()
-            .add_message::<WarningToastEvent>()
+            .add_message::<WarningToastEvent>();
-            .add_systems(Startup, start_pull)
+
-            .add_systems(
+        // Build the shared Tokio runtime; disable all network sync if the OS
-                Update,
+        // refuses to create threads (resource-limited environments, sandboxed
-                (
+        // Android builds, etc.).
-                    poll_pull_result,
+        match TokioRuntimeResource::new() {
-                    handle_manual_sync_request,
+            Ok(rt) => {
-                    push_replay_on_win,
+                app.insert_resource(rt)
-                    poll_replay_upload_result,
+                    .add_systems(Startup, start_pull)
-                ),
+                    .add_systems(
-            )
+                        Update,
-            .add_systems(Last, push_on_exit);
+                        (
                            poll_pull_result,
                            handle_manual_sync_request,
                            push_replay_on_win,
                            poll_replay_upload_result,
                        ),
                    )
                    .add_systems(Last, push_on_exit);
            }
            Err(e) => {
                warn!("sync: failed to create Tokio runtime — network sync disabled: {e}");
            }
        }
    }
 }
@@ -55,7 +55,7 @@ use crate::font_plugin::FontResource;
 use crate::settings_plugin::{SettingsResource, SettingsScreen, SettingsStoragePath};
 use crate::resources::TokioRuntimeResource;
 use crate::sync_plugin::SyncProviderResource;
-use crate::ui_modal::spawn_modal;
+use crate::ui_modal::{spawn_modal, ModalScrim};
 use crate::ui_theme::{
    ACCENT_PRIMARY, BG_ELEVATED, BG_ELEVATED_HI,
    BORDER_SUBTLE, HighContrastBorder, RADIUS_SM, STATE_DANGER, TEXT_DISABLED,
@@ -208,6 +208,7 @@ impl Plugin for SyncSetupPlugin {
 fn open_sync_setup_modal(
    mut events: MessageReader<SyncConfigureRequestEvent>,
    existing: Query<(), With<SyncSetupScreen>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<SyncSetupScreen>)>,
    mut commands: Commands,
    mut focused: ResMut<SyncFocusedField>,
    font_res: Option<Res<FontResource>>,
@@ -219,6 +220,9 @@ fn open_sync_setup_modal(
    if !existing.is_empty() {
        return; // Already open.
    }
    if !other_modal_scrims.is_empty() {
        return; // Another modal is already visible.
    }
    *focused = SyncFocusedField::Url;
    spawn_sync_setup_modal(&mut commands, font_res.as_deref());
 }
@@ -354,9 +358,10 @@ fn handle_auth_button(
        return;
    }
-    // Clear error and show busy indicator.
+    // Clear previous error and show busy indicator.
-    for (mut text, _) in &mut error_nodes {
+    for (mut text, mut color) in &mut error_nodes {
-        text.0 = "Connecting…".to_string();
+        text.0 = String::new();
        color.0 = TEXT_SECONDARY;
    }
    for mut vis in &mut busy_nodes {
        *vis = Visibility::Visible;
@@ -540,6 +545,7 @@ fn handle_logout(
 fn open_delete_confirm_modal(
    mut events: MessageReader<DeleteAccountRequestEvent>,
    existing: Query<(), With<DeleteConfirmScreen>>,
    other_modal_scrims: Query<(), (With<ModalScrim>, Without<DeleteConfirmScreen>)>,
    mut commands: Commands,
    font_res: Option<Res<FontResource>>,
 ) {
@@ -550,6 +556,9 @@ fn open_delete_confirm_modal(
    if !existing.is_empty() {
        return;
    }
    if !other_modal_scrims.is_empty() {
        return; // Another modal is already visible.
    }
    spawn_delete_confirm_modal(&mut commands, font_res.as_deref());
 }
@@ -675,20 +684,31 @@ fn spawn_sync_setup_modal(commands: &mut Commands, font_res: Option<&FontResourc
                font_res,
            );
-            // Error / status line.
+            // Error / status line — two distinct children so visibility and
            // text can be controlled independently.
            body.spawn(Node {
                min_height: Val::Px(18.0),
                flex_direction: FlexDirection::Row,
                align_items: AlignItems::Center,
                column_gap: VAL_SPACE_2,
                ..default()
            })
            .with_children(|row| {
                // Busy indicator: shown while the auth task is in flight.
                row.spawn((
                    SyncAuthError,
                    SyncBusyOverlay,
-                    Text::new(String::new()),
+                    Text::new("…"),
                    make_font(font_res, TYPE_CAPTION),
                    TextColor(TEXT_SECONDARY),
                    Visibility::Hidden,
                ));
                // Error / status text: always laid out, empty when idle.
                row.spawn((
                    SyncAuthError,
                    Text::new(String::new()),
                    make_font(font_res, TYPE_CAPTION),
                    TextColor(TEXT_SECONDARY),
                ));
            });
            // Tab hint — desktop only; no Tab key on Android.
@@ -182,12 +182,16 @@ fn sync_card_image_set_with_active_theme(
    mut events: MessageReader<AssetEvent<CardTheme>>,
    active: Option<Res<ActiveTheme>>,
    themes: Res<Assets<CardTheme>>,
    asset_server: Option<Res<AssetServer>>,
    mut card_image_set: Option<ResMut<CardImageSet>>,
    mut state_events: MessageWriter<StateChangedEvent>,
 ) {
    let Some(active) = active else { return };
    let active_id = active.0.id();
    let mut should_sync = false;
    // Consume asset events — covers the normal first-load path.
    for ev in events.read() {
        let id = match ev {
            AssetEvent::LoadedWithDependencies { id }
@@ -198,6 +202,22 @@ fn sync_card_image_set_with_active_theme(
            should_sync = true;
        }
    }
    // A→B→A switch: Bevy does not re-fire LoadedWithDependencies for a
    // handle whose asset is already cached. Detect this by checking that
    // `ActiveTheme` itself changed this frame (the resource was just
    // replaced by `react_to_settings_theme_change`) and the underlying
    // asset is already fully loaded. If so, sync immediately rather than
    // waiting for an event that will never arrive.
    if !should_sync
        && active.is_changed()
        && asset_server
            .as_ref()
            .is_some_and(|as_| as_.is_loaded_with_dependencies(active.0.id()))
    {
        should_sync = true;
    }
    if !should_sync {
        return;
    }
@@ -172,14 +172,15 @@ fn advance_time_attack(
    paused: Option<Res<crate::pause_plugin::PausedResource>>,
    path: Option<Res<TimeAttackSessionPath>>,
    home_screens: Query<(), With<crate::home_plugin::HomeScreen>>,
    win_overlays: Query<(), With<crate::win_summary_plugin::WinSummaryOverlay>>,
 ) {
    if !session.active {
        return;
    }
-    // Mirrors `tick_elapsed_time`: pause while the launch / mode-picker
+    // Pause the countdown while Home, the Pause overlay, or the Win Summary
-    // Home modal is up so the countdown doesn't burn while the player
+    // overlay is visible — the player should not lose time while reading results
-    // is choosing what to play next.
+    // or navigating menus.
-    if paused.is_some_and(|p| p.0) || !home_screens.is_empty() {
+    if paused.is_some_and(|p| p.0) || !home_screens.is_empty() || !win_overlays.is_empty() {
        return;
    }
    session.remaining_secs -= time.delta_secs();
@@ -24,6 +24,7 @@ use crate::progress_plugin::ProgressResource;
 use crate::resources::GameStateResource;
 use crate::settings_plugin::SettingsResource;
 use crate::stats_plugin::{StatsResource, StatsUpdate};
 use crate::ui_modal::ModalScrim;
 use crate::ui_theme::{
    scaled_duration, ACCENT_PRIMARY, BG_BASE, BG_ELEVATED, MOTION_SCORE_BREAKDOWN_FADE_SECS,
    MOTION_SCORE_BREAKDOWN_STAGGER_SECS, MOTION_WIN_SHAKE_AMPLITUDE, MOTION_WIN_SHAKE_SECS,
@@ -757,6 +758,7 @@ fn spawn_overlay(
    commands
        .spawn((
            WinSummaryOverlay,
            ModalScrim,
            Node {
                position_type: PositionType::Absolute,
                left: Val::Percent(0.0),
@@ -769,6 +771,7 @@ fn spawn_overlay(
                ..default()
            },
            BackgroundColor(SCRIM),
            GlobalZIndex(Z_WIN_CASCADE),
            ZIndex(Z_WIN_CASCADE),
        ))
        .with_children(|root| {
@@ -390,7 +390,10 @@ pub async fn upload_avatar(
    // Write to a temp file then atomically rename so concurrent readers never
    // see a partially-written avatar.
    std::fs::write(&tmp_path, &body).map_err(|e| AppError::Internal(e.to_string()))?;
-    std::fs::rename(&tmp_path, &path).map_err(|e| AppError::Internal(e.to_string()))?;
+    if let Err(e) = std::fs::rename(&tmp_path, &path) {
        let _ = std::fs::remove_file(&tmp_path);
        return Err(AppError::Internal(e.to_string()));
    }
    // Remove stale files with other extensions after the atomic rename.
    for old_ext in &["jpg", "png", "webp", "gif"] {
        if *old_ext != ext {
@@ -35,7 +35,7 @@ use crate::{error::AppError, middleware::AuthenticatedUser, AppState};
 /// the desktop client's transitive dependencies.
 #[derive(Debug, Deserialize)]
 struct ReplayHeader {
-    seed: i64,
+    seed: u64,
    draw_mode: String,
    mode: String,
    time_seconds: i64,
@@ -94,6 +94,9 @@ pub async fn upload(
    let id = Uuid::new_v4().to_string();
    let received_at = Utc::now().to_rfc3339();
    let replay_json = serde_json::to_string(&payload)?;
    // SQLite INTEGER columns bind as i64. Reinterpret the u64 bits — the
    // database stores the same 8 bytes; high-bit seeds round-trip correctly.
    let seed_i64 = header.seed as i64;
    sqlx::query!(
        r#"INSERT INTO replays (
@@ -102,7 +105,7 @@ pub async fn upload(
           ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)"#,
        id,
        user.user_id,
-        header.seed,
+        seed_i64,
        header.draw_mode,
        header.mode,
        header.time_seconds,
@@ -116,7 +119,7 @@ pub async fn upload(
    // Update leaderboard best score/time for opted-in users when this replay
    // beats their existing best. Only classic mode counts for the leaderboard.
-    if header.mode == "classic" {
+    if header.mode == "Classic" {
        sqlx::query!(
            r#"UPDATE leaderboard
               SET best_score      = ?,
@@ -300,6 +300,7 @@ function render(s) {
    board.querySelectorAll(".card.drop-target").forEach(e => e.classList.remove("drop-target"));
    if (s.is_auto_completable && !s.is_won && !acTimer) {
        stopTimer(); // freeze elapsed time at the moment the player's last move completes
        acTimer = setInterval(doAutoCompleteStep, 380);
    }
    if (s.is_won) {
@@ -40,20 +40,32 @@ export class ReplayPlayer {
    }
    /**
     * Snapshot the current `GameState` as a JS object (see `StateSnapshot`).
     *
     * Throws a JS string exception on serialisation failure (should never
     * occur in practice — `StateSnapshot` contains only primitive types).
     * @returns {any}
     */
    state() {
        const ret = wasm.replayplayer_state(this.__wbg_ptr);
-        return ret;
+        if (ret[2]) {
            throw takeFromExternrefTable0(ret[1]);
        }
        return takeFromExternrefTable0(ret[0]);
    }
    /**
     * Apply the next move; returns the post-step snapshot, or `null`
     * once the move list is exhausted.
     *
     * Returns `null` (not an exception) when the replay is finished.
     * Throws a JS string exception on serialisation failure.
     * @returns {any}
     */
    step() {
        const ret = wasm.replayplayer_step(this.__wbg_ptr);
-        return ret;
+        if (ret[2]) {
            throw takeFromExternrefTable0(ret[1]);
        }
        return takeFromExternrefTable0(ret[0]);
    }
    /**
     * 0-indexed position of the next move to apply.
@@ -157,11 +169,16 @@ export class SolitaireGame {
    }
    /**
     * Full pile snapshot as a JS object.
     *
     * Throws a JS string exception on serialisation failure.
     * @returns {any}
     */
    state() {
        const ret = wasm.solitairegame_state(this.__wbg_ptr);
-        return ret;
+        if (ret[2]) {
            throw takeFromExternrefTable0(ret[1]);
        }
        return takeFromExternrefTable0(ret[0]);
    }
    /**
     * Undo the last move. Returns `{ok, error?, snapshot?}`.
@@ -180,6 +197,13 @@ function __wbg_get_imports() {
            const ret = Error(getStringFromWasm0(arg0, arg1));
            return ret;
        },
        __wbg_String_8564e559799eccda: function(arg0, arg1) {
            const ret = String(arg1);
            const ptr1 = passStringToWasm0(ret, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
            const len1 = WASM_VECTOR_LEN;
            getDataViewMemory0().setInt32(arg0 + 4 * 1, len1, true);
            getDataViewMemory0().setInt32(arg0 + 4 * 0, ptr1, true);
        },
        __wbg___wbindgen_throw_9c75d47bf9e7731e: function(arg0, arg1) {
            throw new Error(getStringFromWasm0(arg0, arg1));
        },
@@ -108,23 +108,27 @@ fn merge_stats(
    let merged_games_won = local.games_won.max(remote.games_won);
    let merged_games_played = local.games_played.max(remote.games_played);
-    // Recompute average time from the merged totals. If no wins yet, keep 0.
+    // Carry the average time from whichever side contributed merged_games_won.
    // Taking max(total_time)/max(wins) misattributes time when the side with
    // more wins has a lower total — use the winning side's average directly.
    let avg_time_seconds = if merged_games_won == 0 {
        0
    } else if local.games_won >= remote.games_won {
        local.avg_time_seconds
    } else {
-        // Use whichever side has more wins to approximate total time, then blend.
+        remote.avg_time_seconds
        // We don't have total_time stored, so we reconstruct it from avg * count.
        let local_total = local.avg_time_seconds as u128 * local.games_won as u128;
        let remote_total = remote.avg_time_seconds as u128 * remote.games_won as u128;
        // Take max total time (conservative — avoids underestimating total play time).
        let best_total = local_total.max(remote_total);
        (best_total / merged_games_won as u128) as u64
    };
    // Derive games_lost from the merged played/won counts so the invariant
    // games_won + games_lost <= games_played is always satisfied. Computing
    // max(local.games_lost, remote.games_lost) independently can push
    // games_won + games_lost above games_played after a divergent merge.
    let merged_games_lost = merged_games_played.saturating_sub(merged_games_won);
    StatsSnapshot {
        games_played: merged_games_played,
        games_won: merged_games_won,
-        games_lost: local.games_lost.max(remote.games_lost),
+        games_lost: merged_games_lost,
        win_streak_current: local.win_streak_current.max(remote.win_streak_current),
        win_streak_best: local.win_streak_best.max(remote.win_streak_best),
        avg_time_seconds,
@@ -454,14 +458,28 @@ mod tests {
    }
    #[test]
-    fn stats_games_lost_takes_max() {
+    fn stats_games_lost_derived_from_played_minus_won() {
        // games_lost must equal games_played - games_won so the invariant
        // games_won + games_lost <= games_played is always satisfied.
        let mut local = default_payload();
        local.stats.games_played = 20;
        local.stats.games_won = 8;
        local.stats.games_lost = 12;
        let mut remote = default_payload();
-        remote.stats.games_lost = 8;
+        remote.stats.games_played = 15;
        remote.stats.games_won = 10;
        remote.stats.games_lost = 5;
        // merged: games_played = max(20, 15) = 20; games_won = max(8, 10) = 10
        // games_lost must be 20 - 10 = 10, NOT max(12, 5) = 12
        let (merged, _) = merge(&local, &remote);
-        assert_eq!(merged.stats.games_lost, 12);
+        assert_eq!(merged.stats.games_played, 20);
        assert_eq!(merged.stats.games_won, 10);
        assert_eq!(merged.stats.games_lost, 10);
        assert!(
            merged.stats.games_won + merged.stats.games_lost <= merged.stats.games_played,
            "games_won + games_lost must never exceed games_played"
        );
    }
    #[test]
@@ -502,11 +520,10 @@ mod tests {
    #[test]
    fn stats_avg_time_recomputed_from_merged_totals() {
-        // local: 4 wins averaging 100s each (total = 400s)
+        // local: 4 wins averaging 100s each
-        // remote: 6 wins averaging 200s each (total = 1200s)
+        // remote: 6 wins averaging 200s each
-        // merged_games_won = max(4, 6) = 6
+        // merged_games_won = max(4, 6) = 6  →  remote contributed the wins
-        // best_total = max(400, 1200) = 1200
+        // avg_time_seconds must be remote's 200s, not a blend of totals
        // avg = 1200 / 6 = 200
        let mut local = default_payload();
        local.stats.games_won = 4;
        local.stats.avg_time_seconds = 100;
Author	SHA1	Message	Date
Gitea CI	18659d19d1	chore(deploy): bump image to `7840ef9e` [skip ci]	2026-05-19 20:19:02 +00:00
funman300	7840ef9eb2	fix(multi): resolve 26 bugs found in comprehensive codebase review Build and Deploy / build-and-push (push) Successful in 3m40s Details Core fixes (issues #12, #13, #22): - #12: undo now preserves score delta instead of restoring snapshot score - #13: take_from_foundation defaults to false (non-standard house rule) - #22: check_win validates full suit sequence, not just card count Engine fixes: - #8: replay keyboard input guard against non-replay state - #9: help modal scrims.is_empty() guard added - #10: settings modal scrims.is_empty() guard added - #11: sync_plugin builds payload at poll time (not task-spawn time) - #14: server replay mode case-sensitivity fix ("Classic") - #15: play_by_seed_plugin confirmed flag set to true on launch - #16: replay back-step debounce via Local<bool> + StateChangedEvent; register StateChangedEvent in ReplayOverlayPlugin (fixes 52 tests) - #17: time-attack timer ignores win-summary overlay - #18: HUD dropdown glyphs U+25BE → U+2193 (FiraMono-safe arrow) - #19: theme plugin applies immediate visual update on A→B→A switch - #20: SyncAuthError / SyncBusyOverlay split into separate entities so auth errors are visible after busy overlay is hidden - #21: handle_forfeit ordered before update_stats_on_new_game - #23: server merge uses correct avg_time_seconds and games_lost math - #24: win_summary migrated to ModalScrim pattern - #25: card_animation apply_deferred between animation systems - #26: cursor_plugin HashMap access uses .get() with fallback - #27: auto_complete mid-sequence deactivation guard - #28: feedback_anim SettleAnim ordered before FoundationFlourish - #29: achievement_plugin iterates all win events; adds scrims guard - #30: leaderboard modal scrims.is_empty() guard added - #31: server auth tmp file cleanup on rename failure - #32: sync_setup modal scrims.is_empty() guard added - #33: font_plugin uses match fallback; TokioRuntimeResource graceful current-thread fallback on runtime init failure Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 13:14:47 -07:00
funman300	6d061d23a1	fix(engine): cancel stale win-cascade CardAnimation on new-game; refresh Android corner label text on resize (closes #6 , closes #7 ) Issue #7 — new game during win cascade: sync_cards now stores each in-flight CardAnimation's end position instead of a plain bool. Before calling update_card_entity, the end position is compared against the game-state target. If they differ by more than 2 px (stale cascade scatter vs. new-game dealt position) the CardAnimation is removed immediately so the card slides to its correct dealt position. Drag-rejection tweens are unaffected because their end equals the card's current game-state position. Issue #6 — Android stale corner label text: AndroidCornerLabel now carries the label string as AndroidCornerLabel(String). resize_android_corner_labels refreshes Text2d content from the stored value alongside the existing font-size and transform updates, closing the narrow race where a layout change could display a previous card's rank/suit. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 11:32:07 -07:00
funman300	25f22231a6	fix(test): make leaderboard opt-in/opt-out tests robust under parallel runner (closes #5 ) The four tests polled the async task pool with a fixed budget of five app.update() calls. Under cargo test --workspace the pool's background threads are starved by other tests, so even an instantly-resolving future can take more than five frames to be polled. Replace the fixed loop with a deadline-bounded loop (5 s timeout) that exits early once the expected side-effect is observable — the same pattern used in sync_plugin.rs tests. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 11:32:07 -07:00
funman300	c66ff26d1d	fix(engine): lift card z during CardAnim to prevent corner bleed-through When a card slides to a foundation slot already occupied, both card entities share the same x,y for the duration of the tween. With STACK_FAN_FRAC only 0.003 apart, the incoming card partially occludes the stationary one, making the two exposed corners look like a single mismatched card. Elevate every CardAnim-driven card to target.z + 50 during transit so it fully occludes any card resting at the destination. On completion the card snaps to the correct resting z. The value sits below DRAG_Z (500) so dragged cards still render above animated ones. Closes #implicitly-related-to-corner-mismatch-investigation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 11:32:07 -07:00
funman300	cd792b20b2	chore: ignore ruflo runtime state files Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 11:32:07 -07:00
Gitea CI	73c7f50f74	chore(deploy): bump image to `83c40116` [skip ci]	2026-05-19 02:03:57 +00:00
funman300	83c40116af	fix(web): freeze timer when auto-complete begins (closes #4 ) Build and Deploy / build-and-push (push) Successful in 4m5s Details The game timer kept counting during the auto-complete animation even though the player had already made their last decision. stopTimer() is now called the moment is_auto_completable fires so elapsed_seconds reflects only real play time, not the animation delay. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-18 18:59:54 -07:00
Gitea CI	347d5a4b4f	chore(deploy): bump image to `93f2ceaa` [skip ci]	2026-05-19 01:50:10 +00:00
funman300	93f2ceaabe	fix(web): rebuild WASM pkg — foundation→tableau moves now work Build and Deploy / build-and-push (push) Successful in 4m20s Details The pre-built pkg predated fix `c35c045` (enable take-from-foundation by default) so the WASM game always had take_from_foundation=false, silently rejecting every drag from a foundation pile to a tableau column. Rebuilt with wasm-pack --release against current solitaire_core. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-18 18:45:51 -07:00
funman300	e390b72222	chore(tooling): add ruflo-core scaffolding and MCP server registration Initialised ruflo v3 via `npx @claude-flow/cli@latest init --wizard --force`. Registers the ruflo MCP server in .mcp.json (hierarchical-mesh topology, max 15 agents). Includes .claude-flow/ runtime config and capability manifest. .claude/ remains gitignored (local agents/commands/settings stay per-developer). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-18 17:19:28 -07:00
funman300	3650788dc5	fix(engine): prevent stock-tap from toggling HUD on Android Every draw-from-stock tap was also firing the HUD auto-hide toggle because the stock pile is not an ActionButton and toggle_hud_on_tap had no way to know the tap was consumed by game logic. Add GameInputConsumedResource(bool): handle_touch_stock_tap sets it on TouchPhase::Started when a draw fires; toggle_hud_on_tap checks and clears it on TouchPhase::Ended, treating it as equivalent to started_on_button so the HUD stays put. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-18 17:09:58 -07:00
Gitea CI	39cf8dcd6c	chore(deploy): bump image to `456b4d42` [skip ci]	2026-05-18 20:29:08 +00:00