Add pluggable parser system, perception loop, and project infrastructure

Phase 0 (housekeeping):
- Add .gitignore to protect credentials and runtime files
- Track all project source code, docs, and config

Phase 1 (perception loop):
- Implement pluggable parser architecture (BaseMudParser ABC + ParserFactory)
- Add DeadSoulsParser with regex for HP/MP/SP prompt, room descriptions,
  exits (multiple formats), NPC/item detection, and combat events
- Add OutputBuffer for accumulating text between prompts
- Update GameState with SP support and update_from_prompt/room/combat methods
- Rewire main loop to use parser + buffer for real-time game state updates
- Add reconnection logic to TelnetClient (retry with backoff)
- Remove old TextParser (replaced by pluggable system)
- Add mud_type config field for selecting parser at runtime
- Add comprehensive test suite (50 tests, all passing)

TODO.md

@@ -0,0 +1,82 @@
+# MUD AI Player - Implementation Roadmap
+
+This document outlines the tasks required to build a complete, autonomous AI player for LPMUDs.
+
+## 🟢 Phase 1: Core Connectivity & Basic Setup (Current Status)
+
+- [x] **Telnet Client**: Basic connection and disconnection handling.
+- [x] **Configuration**: YAML-based config for credentials and server info.
+- [x] **Automatic Login**: 
+  - [x] Username/Password automtation.
+  - [x] "Take over existing session" handling.
+- [x] **KeepAlive Manager**: Prevent idle disconnects.
+- [ ] **Logging**: 
+  - [x] Setup basic logging.
+  - [ ] Implement strict `game.log` rotation and archiving structure.
+
+## 🟡 Phase 2: Perception (Parsing & Sensing)
+
+The bot needs to "see" the world effectively.
+
+- [ ] **ANSI Parsing**:
+  - [ ] Refine `ANSIParser` to fully strip color codes for logic but preserve them for UI if needed.
+  - [ ] Detect prompt lines vs. content lines.
+- [ ] **Status Parsing (Vitals)**:
+  - [ ] Implement regex to capture HP/MP/SP from prompt (e.g., `hp: 350/350`).
+  - [ ] Update `GameState.character` with real-time stats.
+- [ ] **Room Parsing (Vision)**:
+  - [ ] Identify Room Title (usually colored/bold).
+  - [ ] Identify Room Description (multiline text).
+  - [ ] Identify Exits (e.g., `[ Exits: n s w ]`).
+  - [ ] Identify NPCs/Mobs/Items in the room.
+- [ ] **Inventory & Equipment**:
+  - [ ] Implement `inv` command parser.
+  - [ ] Implement `eq` command parser.
+
+## 🟠 Phase 3: World Modeling (Mapping)
+
+The bot needs to know where it is and where it has been.
+
+- [ ] **Room Hashing**:
+  - [ ] Create unique hash for rooms based on Title + Description + Exits (to handle non-unique IDs).
+- [ ] **Graph Database**:
+  - [ ] Integrate `NetworkX` or SQLite for storing the world map.
+  - [ ] Store nodes (Rooms) and edges (Exits).
+- [ ] **Pathfinding**:
+  - [ ] Implement A* algorithm to find shortest path between known rooms.
+  - [ ] Implement "Breadth-First Search" (BFS) for exploration mode.
+
+## 🔴 Phase 4: Decision Engine (The Brain)
+
+The bot needs to decide what to do next.
+
+- [ ] **Action Queue System**:
+  - [ ] Priority queue for actions (e.g., `HEAL` > `ATTACK` > `LOOT` > `MOVE`).
+- [ ] **Behavior Trees / State Machine**:
+  - [ ] **IdleState**: Heal, Buff, Check Inventory.
+  - [ ] **ExploreState**: Pick unknown exit, Move.
+  - [ ] **CombatState**:
+    - [ ] Auto-target hostile mobs.
+    - [ ] Execute combat rotation (skills/spells).
+    - [ ] Flee trigger (if HP < Threshold).
+  - [ ] **LootState**: reliable looting of corpses.
+- [ ] **LLM Integration (Optional/Hybrid)**:
+  - [ ] Use LLM for "roleplay" responses to other players.
+  - [ ] Use LLM for "quest parsing" (understanding vague instructions).
+
+## 🔵 Phase 5: Humanization
+
+Make the bot undetectable and less robotic.
+
+- [ ] **Reaction Time Jitter**:
+  - [ ] Randomize delays between commands (e.g., 200ms - 1500ms).
+- [ ] **Typo Generator**:
+  - [ ] Occasionally misspell commands and correct them (optional).
+- [ ] **Schedule System**:
+  - [ ] Define "active hours" and "sleep hours".
+
+## 🟣 Phase 6: Testing & Metrics
+
+- [ ] **Unit Tests**: Coverage for parsers (critical).
+- [ ] **Integration Tests**: Mock MUD server for offline testing.
+- [ ] **Performance Profiling**: Ensure graph traversal doesn't lag the bot.