gencode-ai 0.2.0 → 0.3.0

package/docs/proposals/0006-memory-system.md

@@ -1,4 +1,4 @@
-# Proposal: Memory System (AGENT.md) with /init Command
+# Proposal: Memory System (GEN.md) with /init Command
 
 - **Proposal ID**: 0006
 - **Author**: mycode team
@@ -10,13 +10,13 @@
 ## Summary
 
 Implement a comprehensive memory system inspired by Claude Code, including:
-1. **AGENT.md files** - Project-specific context that persists across sessions (with CLAUDE.md fallback for compatibility)
+1. **GEN.md files** - Project-specific context that persists across sessions (with CLAUDE.md fallback for compatibility)
 2. **Hierarchical memory loading** - User → Project → Local with rules directories
-3. **/init command** - Automatic project analysis and AGENT.md generation
+3. **/init command** - Automatic project analysis and GEN.md generation
 4. **# prefix** - Quick memory adds during sessions (`# note` for project, `## note` for user)
 5. **/memory command** - View loaded memory files
 6. **@import syntax** - Include other files into memory (max 5 levels)
-7. **.gencode/rules/ directory** - Modular, path-scoped rules with frontmatter
+7. **.gen/rules/ directory** - Modular, path-scoped rules with frontmatter
 
 ## Motivation
 

package/docs/proposals/0008-checkpointing.md

@@ -2,9 +2,9 @@
 
 - **Proposal ID**: 0008
 - **Author**: mycode team
-- **Status**: Draft
+- **Status**: Partially Implemented
 - **Created**: 2025-01-15
-- **Updated**: 2025-01-15
+- **Updated**: 2025-01-17
 
 ## Summary
 
@@ -325,3 +325,110 @@ No breaking changes to existing functionality.
 
 - [Claude Code Checkpointing](https://code.claude.com/docs/en/checkpointing)
 - [Git Reset and Revert](https://git-scm.com/docs/git-reset)
+
+## Implementation Status
+
+### ✅ Implemented (Phase 1-2)
+
+**Core Checkpointing System**:
+- ✅ `CheckpointManager` class with full API
+  - `recordChange()` - Record file changes
+  - `getCheckpoints()` - Get all checkpoints
+  - `getFileHistory()` - Get file-specific history
+  - `rewind()` - Rewind with multiple options (checkpointId, path, all, count)
+  - `getSummary()` - Get change summary
+  - `clearCheckpoints()` - Clear checkpoints
+  - `formatCheckpointList()` - Format for display
+
+**Tool Integration**:
+- ✅ Automatic tracking in `ToolRegistry`
+  - Pre-execution file state capture
+  - Post-execution checkpoint recording
+  - Support for Write and Edit tools
+  - Automatic detection of create/modify/delete operations
+
+**CLI Commands**:
+- ✅ `/changes` - List all file changes in session
+- ✅ `/rewind [n]` - Rewind specific checkpoint by index
+- ✅ `/rewind all` - Rewind all changes
+- ✅ `/rewind` - Show changes list with usage info
+
+**Type System**:
+- ✅ Complete type definitions (`src/checkpointing/types.ts`)
+- ✅ All interfaces from the proposal
+
+**Testing**:
+- ✅ Test example (`examples/test-checkpointing.ts`)
+
+### ❌ Not Implemented (Phase 3-5)
+
+**Session Persistence**:
+- ❌ Checkpoints not saved to session files
+- ❌ No checkpoint restoration when resuming sessions
+- ❌ Session type doesn't include checkpoint data
+- **Impact**: Checkpoints are lost when the session ends
+
+**User Experience Enhancements**:
+- ❌ Confirmation prompt for `/rewind all`
+  - Currently executes immediately without confirmation
+  - Proposal shows interactive confirmation UI
+- ❌ Inline change indicators
+  - Proposal shows checkpoint saved message after each file modification
+  - Currently no visual feedback when checkpoint is created
+- ❌ Formatted change list UI with boxes/borders
+  - Current implementation uses simple text list
+  - Proposal shows fancy bordered display
+
+**Git Integration** (Phase 3):
+- ❌ Optional git-based versioning
+- ❌ Git commits as checkpoints
+- ❌ Integration with git workflow
+
+**Advanced Features** (Phase 4-5):
+- ❌ Selective rewind by time range
+- ❌ Checkpoint browsing UI
+- ❌ Diff viewing between checkpoints
+- ❌ Storage limits and cleanup policies
+- ❌ Large file optimization
+
+### 📋 Remaining Work
+
+To complete this proposal, the following tasks are needed:
+
+1. **Session Persistence** (High Priority):
+   - Add `checkpoints` field to `SessionMetadata` or `Session` type
+   - Save/load checkpoints in `SessionManager`
+   - Restore checkpoint manager state when resuming sessions
+
+2. **Confirmation UI** (Medium Priority):
+   - Add confirmation prompt for `/rewind all` command
+   - Show list of files that will be affected
+   - Allow user to confirm or cancel
+
+3. **Visual Feedback** (Medium Priority):
+   - Show checkpoint saved message after Write/Edit operations
+   - Improve `/changes` display with better formatting
+   - Add color coding for different change types
+
+4. **Git Integration** (Low Priority):
+   - Optional: Use git for checkpoint storage
+   - Optional: Create git commits as checkpoints
+   - Optional: Integrate with existing git workflow
+
+5. **Advanced Features** (Future):
+   - Time-based rewind
+   - Diff viewing
+   - Storage optimization
+
+### 📁 Implementation Files
+
+| File | Status | Notes |
+|------|--------|-------|
+| `src/checkpointing/types.ts` | ✅ Complete | All types defined |
+| `src/checkpointing/checkpoint-manager.ts` | ✅ Complete | Core logic implemented |
+| `src/checkpointing/index.ts` | ✅ Complete | Module exports |
+| `src/tools/registry.ts` | ✅ Modified | Checkpoint tracking added |
+| `src/cli/components/App.tsx` | ✅ Modified | `/changes` and `/rewind` commands |
+| `examples/test-checkpointing.ts` | ✅ Complete | Test coverage |
+| `src/session/types.ts` | ❌ Not Modified | Missing checkpoint fields |
+| `src/session/manager.ts` | ❌ Not Modified | No persistence logic |

package/docs/proposals/0011-custom-commands.md

@@ -3,8 +3,9 @@
 - **Proposal ID**: 0011
 - **Author**: mycode team
 - **Status**: Draft
+- **Priority**: P2 (Enhanced Feature)
 - **Created**: 2025-01-15
-- **Updated**: 2025-01-15
+- **Updated**: 2025-01-17
 
 ## Summary
 

package/docs/proposals/0021-skills-system.md

@@ -3,8 +3,9 @@
 - **Proposal ID**: 0021
 - **Author**: mycode team
 - **Status**: Draft
+- **Priority**: P1 (Core Feature)
 - **Created**: 2025-01-15
-- **Updated**: 2025-01-15
+- **Updated**: 2025-01-17
 
 ## Summary
 

package/docs/proposals/0023-permission-enhancements.md

@@ -487,7 +487,7 @@ Backward compatible with existing rules.
 
 1. **Pattern-Based Rules**: Supports Claude Code format like `Bash(git add:*)` with glob-style wildcards
 2. **Prompt-Based Permissions**: ExitPlanMode style with semantic matching for common operations (run tests, install dependencies, etc.)
-3. **Multi-Scope Permissions**: Session (in-memory), Project (.gencode/permissions.json), Global (~/.gencode/permissions.json)
+3. **Multi-Scope Permissions**: Session (in-memory), Project (.gen/permissions.json), Global (~/.gen/permissions.json)
 4. **Approval Options**: Allow once, Allow for session, Always allow (persistent), Deny
 5. **Audit Logging**: In-memory audit trail with optional file persistence
 6. **CLI Commands**: `/permissions` shows rules, `/permissions audit` shows decision history, `/permissions stats` shows statistics
@@ -507,7 +507,7 @@ Backward compatible with existing rules.
 
 ### Settings Configuration
 
-Add to `~/.gencode/settings.json`:
+Add to `~/.gen/settings.json`:
 ```json
 {
   "permissions": {

package/docs/proposals/0033-enterprise-deployment.md

@@ -179,7 +179,7 @@ Enterprise Mode: Acme Corp
 Authenticating via SSO...
 
 Opening browser for authentication...
-✓ Authenticated as john.doe@acme.com
+✓ Authenticated as user@company.com
 
 Session policies:
 - Allowed providers: anthropic, openai

package/docs/proposals/0041-configuration-system.md

@@ -49,9 +49,7 @@ Claude Code supports extensive environment variables:
 
 **Cloud Providers:**
 - `CLAUDE_CODE_USE_BEDROCK` - Enable AWS Bedrock
-- `CLAUDE_CODE_USE_VERTEX` - Enable Google Vertex AI
 - `CLAUDE_CODE_SKIP_BEDROCK_AUTH` - Bypass AWS auth
-- `CLAUDE_CODE_SKIP_VERTEX_AUTH` - Bypass Vertex auth
 
 **Operational:**
 - `CLAUDE_CODE_MAX_OUTPUT_TOKENS` - Token limit
@@ -121,7 +119,7 @@ OpenCode provides additional patterns worth adopting:
 ### Directory Structure
 
 ```
-~/.gencode/                          # User-level configuration
+~/.gen/                          # User-level configuration
 ├── settings.json                    # Main user config
 ├── settings.local.json              # User local overrides (gitignored pattern)
 ├── GENCODE.md                       # User context (like CLAUDE.md)
@@ -135,7 +133,7 @@ OpenCode provides additional patterns worth adopting:
 └── sessions/                        # Session data
 
 ./gencode.json                       # Project config (like opencode.json)
-./.gencode/                          # Project directory
+./.gen/                          # Project directory
 ├── settings.local.json              # Project local overrides (gitignored)
 ├── GENCODE.md                       # Project context
 ├── rules/                           # Path-scoped rules
@@ -146,10 +144,10 @@ OpenCode provides additional patterns worth adopting:
 
 1. **Environment variables** (`GENCODE_*`, provider API keys)
 2. **CLI arguments** (`--model`, `--provider`)
-3. **Project local** (`./.gencode/settings.local.json`)
+3. **Project local** (`./.gen/settings.local.json`)
 4. **Project shared** (`./gencode.json`)
-5. **User local** (`~/.gencode/settings.local.json`)
-6. **User global** (`~/.gencode/settings.json`)
+5. **User local** (`~/.gen/settings.local.json`)
+6. **User global** (`~/.gen/settings.json`)
 7. **Defaults**
 
 ### API Design
@@ -227,8 +225,8 @@ interface EnvHandler {
 ### Environment Variables
 
 **Provider Selection:**
-- `GENCODE_PROVIDER` - Provider name (anthropic, openai, gemini, bedrock, vertex)
-- `GENCODE_MODEL` - Model ID
+- `GEN_PROVIDER` - Provider name (anthropic, openai, gemini, bedrock, vertex)
+- `GEN_MODEL` - Model ID
 - `GENCODE_CONFIG` - Custom config file path
 
 **Provider API Keys (Auto-detect):**
@@ -300,7 +298,7 @@ const ConfigSchema = z.object({
 });
 
 export class ConfigLoader {
-  private userConfigDir = path.join(os.homedir(), '.gencode');
+  private userConfigDir = path.join(os.homedir(), '.gen');
 
   async load(): Promise<GencodeConfig> {
     const configs = await Promise.all([
@@ -349,7 +347,7 @@ export class EnvHandler {
   };
 
   getProviderFromEnv(): string | undefined {
-    return process.env.GENCODE_PROVIDER;
+    return process.env.GEN_PROVIDER;
   }
 
   getApiKey(provider: string): string | undefined {
@@ -442,7 +440,7 @@ $ gencode
 # Uses OpenAI automatically
 
 # Or specify explicitly
-$ export GENCODE_PROVIDER=anthropic
+$ export GEN_PROVIDER=anthropic
 $ export ANTHROPIC_API_KEY=sk-ant-...
 $ gencode
 ```
@@ -480,7 +478,7 @@ $ gencode config deny "Read(.env)"
 
 ### Single Config File
 
-A single `~/.gencoderc` file would be simpler but lacks:
+A single `~/.genrc` file would be simpler but lacks:
 - Project-specific overrides
 - Team-shareable settings
 - Local-only sensitive settings
@@ -526,7 +524,7 @@ Simpler implementation but:
 
 ## Migration Path
 
-1. **From mycode**: Migrate `~/.mycode/` to `~/.gencode/`
+1. **From mycode**: Migrate `~/.mycode/` to `~/.gen/`
 2. **Version Detection**: Check for legacy config locations and prompt migration
 3. **Backward Compatibility**: Support reading old config format for one major version
 
@@ -570,9 +568,9 @@ Simpler implementation but:
 
 ### Key Implementation Details
 
-1. **Multi-level Loading**: User (`~/.gencode/`) → Project (`./.gencode/`) with proper merge
+1. **Multi-level Loading**: User (`~/.gen/`) → Project (`./.gen/`) with proper merge
 2. **Claude Code Compatibility**: Supports both `GENCODE.md` and `CLAUDE.md` for memory files
-3. **Environment Variables**: `GENCODE_PROVIDER`, `GENCODE_MODEL`, and provider API key auto-detection
+3. **Environment Variables**: `GEN_PROVIDER`, `GEN_MODEL`, and provider API key auto-detection
 4. **Deep Merge**: Arrays concatenate, objects merge recursively
 5. **JSONC Support**: Configuration files support comments and trailing commas
 
@@ -580,8 +578,8 @@ Simpler implementation but:
 
 1. Environment variables (`GENCODE_*`)
 2. CLI arguments
-3. Project local (`./.gencode/settings.local.json`)
+3. Project local (`./.gen/settings.local.json`)
 4. Project shared (`./gencode.json`)
-5. User local (`~/.gencode/settings.local.json`)
-6. User global (`~/.gencode/settings.json`)
+5. User local (`~/.gen/settings.local.json`)
+6. User global (`~/.gen/settings.json`)
 7. Defaults

package/docs/proposals/0042-prompt-optimization.md

@@ -199,7 +199,7 @@ src/prompts/
 
 ### 4.2 Prompt Loading Strategy
 
-The prompt loading system uses a **model → provider → prompt** flow, leveraging the `~/.gencode/providers.json` configuration to automatically determine which prompt to use.
+The prompt loading system uses a **model → provider → prompt** flow, leveraging the `~/.gen/providers.json` configuration to automatically determine which prompt to use.
 
 #### 4.2.1 Loading Flow
 
@@ -209,7 +209,7 @@ model ID (e.g., "claude-sonnet-4-5@20250929")
     ▼
 ┌─────────────────────────────────────┐
 │ Look up provider in providers.json  │
-│ (search models.{provider}.list)     │
+│ (search models[provider:*].list)    │
 └──────────────┬──────────────────────┘
                │
     ┌──────────┴──────────┐
@@ -242,18 +242,26 @@ model ID (e.g., "claude-sonnet-4-5@20250929")
 
 #### 4.2.2 Provider Lookup from providers.json
 
-The `~/.gencode/providers.json` file stores cached models for each connected provider:
+The `~/.gen/providers.json` file stores cached models for each provider using `"provider:authMethod"` keys:
 
 ```json
 {
   "models": {
-    "anthropic": {
+    "anthropic:api_key": {
+      "cachedAt": "2025-01-17T10:00:00Z",
       "list": [
         { "id": "claude-sonnet-4-5@20250929", "name": "Claude Sonnet 4.5" },
         { "id": "claude-opus-4-1@20250805", "name": "Claude Opus 4.1" }
       ]
     },
-    "gemini": {
+    "anthropic:vertex": {
+      "cachedAt": "2025-01-17T09:00:00Z",
+      "list": [
+        { "id": "claude-3-5-sonnet@20241022", "name": "Claude 3.5 Sonnet" }
+      ]
+    },
+    "gemini:api_key": {
+      "cachedAt": "2025-01-17T08:00:00Z",
       "list": [
         { "id": "gemini-2.5-pro", "name": "Gemini 2.5 Pro" }
       ]
@@ -262,7 +270,7 @@ The `~/.gencode/providers.json` file stores cached models for each connected pro
 }
 ```
 
-Given a model ID, the system searches through each provider's model list to find the owning provider.
+Given a model ID, the system searches through each provider's model list to find the owning provider. The key format is `"provider:authMethod"` to support multiple authentication methods for the same provider.
 
 #### 4.2.3 Provider to Prompt Type Mapping
 
@@ -293,7 +301,7 @@ This ensures the system always has a valid prompt to use.
 // 5.1.1 Provider Type Definition
 export type ProviderType = 'anthropic' | 'openai' | 'gemini' | 'generic';
 
-// 5.1.2 Look up provider for a model from ~/.gencode/providers.json
+// 5.1.2 Look up provider for a model from ~/.gen/providers.json
 export function getProviderForModel(model: string): string | null {
   const config = loadProvidersConfig();
   if (!config?.models) return null;
@@ -462,7 +470,7 @@ User starts GenCode CLI
            │
            ▼
 ┌─────────────────────┐     ┌─────────────────────────────┐
-│ 3. Memory Loading    │────▶│ Load ~/.gencode/GENCODE.md  │
+│ 3. Memory Loading    │────▶│ Load ~/.gen/GENCODE.md  │
 │    (if exists)       │     │ Load ./CLAUDE.md (fallback) │
 └──────────┬──────────┘     └─────────────────────────────┘
            │
@@ -480,7 +488,7 @@ User starts GenCode CLI
 │                                                         │
 │    ┌─────────────────────────────────────────────────┐  │
 │    │ a. getPromptTypeForModel(model, fallback)       │  │
-│    │    - Look up provider in ~/.gencode/providers.json│
+│    │    - Look up provider in ~/.gen/providers.json│
 │    │    - Map provider → prompt type                  │  │
 │    │    - Fallback to 'generic' if not found         │  │
 │    └─────────────────────────────────────────────────┘  │

package/docs/proposals/README.md

@@ -19,15 +19,14 @@ This directory contains enhancement proposals for the gencode project. Each prop
 | [0001](./0001-web-fetch-tool.md) | WebFetch Tool | Implemented |
 | [0002](./0002-web-search-tool.md) | WebSearch Tool | Implemented |
 | [0003](./0003-task-subagents.md) | Task Tool & Subagents | Draft |
-| [0004](./0004-plan-mode.md) | Plan Mode | Draft |
+| [0004](./0004-plan-mode.md) | Plan Mode | Implemented |
 | [0005](./0005-todo-system.md) | Todo System | Implemented |
 | [0006](./0006-memory-system.md) | Memory System (MYCODE.md) | Implemented |
 | [0007](./0007-context-management.md) | Context Management | Draft |
-| [0008](./0008-checkpointing.md) | Checkpointing | Draft |
+| [0008](./0008-checkpointing.md) | Checkpointing | Partially Implemented |
 | [0009](./0009-hooks-system.md) | Hooks System | Draft |
 | [0010](./0010-mcp-integration.md) | MCP Integration | Draft |
-| [0011](./0011-custom-commands.md) | Custom Commands | Draft |
-| [0012](./0012-ask-user-question.md) | AskUserQuestion Tool | Implemented |
+| [0021](./0021-skills-system.md) | Skills System | Draft |
 | [0041](./0041-configuration-system.md) | Configuration System | Implemented |
 | [0042](./0042-prompt-optimization.md) | Prompt System Optimization | Implemented |
 
@@ -35,6 +34,8 @@ This directory contains enhancement proposals for the gencode project. Each prop
 
 | ID | Title | Status |
 |----|-------|--------|
+| [0011](./0011-custom-commands.md) | Custom Commands | Draft |
+| [0012](./0012-ask-user-question.md) | AskUserQuestion Tool | Implemented |
 | [0013](./0013-multi-edit-tool.md) | MultiEdit Tool | Draft |
 | [0014](./0014-lsp-tool.md) | LSP Tool | Draft |
 | [0015](./0015-ls-tool.md) | LS Tool | Draft |
@@ -43,7 +44,6 @@ This directory contains enhancement proposals for the gencode project. Each prop
 | [0018](./0018-parallel-tool-execution.md) | Parallel Tool Execution | Draft |
 | [0019](./0019-session-enhancements.md) | Session Enhancements | Draft |
 | [0020](./0020-session-summarization.md) | Session Summarization | Draft |
-| [0021](./0021-skills-system.md) | Skills System | Draft |
 | [0022](./0022-plugin-system.md) | Plugin System | Draft |
 | [0023](./0023-permission-enhancements.md) | Permission Enhancements | Draft |
 | [0024](./0024-keyboard-shortcuts.md) | Keyboard Shortcuts | Draft |

package/docs/providers.md

@@ -131,16 +131,25 @@ To use Claude models via Google Vertex AI:
 ### 1. Set Environment Variables
 
 ```bash
+# Required: Enable Vertex AI
+export CLAUDE_CODE_USE_VERTEX=1
+
 # Required: Your GCP project ID
 export ANTHROPIC_VERTEX_PROJECT_ID="your-project-id"
 
-# Optional: Region (defaults to us-east5)
-export ANTHROPIC_VERTEX_REGION="us-east5"
+# Required: Region (use 'global' or specific region)
+export CLOUD_ML_REGION=global
 ```
 
-Alternative variables also supported:
-- `GOOGLE_CLOUD_PROJECT` - GCP project ID
-- `CLOUD_ML_REGION` - GCP region
+**Environment Variable Details:**
+
+| Variable | Required | Default | Description |
+|----------|----------|---------|-------------|
+| `CLAUDE_CODE_USE_VERTEX` | Yes | - | Set to `1` to enable Vertex AI |
+| `ANTHROPIC_VERTEX_PROJECT_ID` | Yes | - | Your GCP project ID (also accepts `GCLOUD_PROJECT` or `GOOGLE_CLOUD_PROJECT`) |
+| `CLOUD_ML_REGION` | Yes | `us-east5` | Region (use `global` or specific region like `us-east5`) |
+
+**Documentation:** https://code.claude.com/docs/en/google-vertex-ai
 
 ### 2. Authenticate with Google Cloud
 
@@ -169,12 +178,24 @@ gcloud services enable aiplatform.googleapis.com
 # Press Enter to connect
 ```
 
+## Architecture
+
+GenCode uses a two-layer provider architecture:
+
+- **Layer 1: Provider** (Semantic layer) - `anthropic` | `openai` | `gemini`
+- **Layer 2: AuthMethod** (Implementation layer) - `api_key` | `vertex` | `bedrock` | `azure`
+
+Each provider can support multiple authentication methods. For example, Anthropic supports:
+- `api_key` - Direct API access
+- `vertex` - Google Vertex AI
+- `bedrock` - Amazon Bedrock (coming soon)
+
 ## Configuration Storage
 
 Provider connections and cached models are stored in:
 
 ```
-~/.gencode/
+~/.gen/
 ├── settings.json      # Current model and provider settings
 └── providers.json     # Provider connections and model cache
 ```
@@ -185,23 +206,87 @@ Provider connections and cached models are stored in:
 {
   "connections": {
     "anthropic": {
-      "method": "vertex",
+      "authMethod": "vertex",
+      "method": "Google Vertex AI",
       "connectedAt": "2025-01-15T10:00:00Z"
+    },
+    "openai": {
+      "authMethod": "api_key",
+      "method": "Direct API",
+      "connectedAt": "2025-01-15T09:00:00Z"
     }
   },
   "models": {
-    "anthropic": {
+    "anthropic:vertex": {
       "cachedAt": "2025-01-15T10:00:00Z",
+      "list": [
+        { "id": "claude-3-5-sonnet@20241022", "name": "Claude 3.5 Sonnet" }
+      ]
+    },
+    "anthropic:api_key": {
+      "cachedAt": "2025-01-15T11:00:00Z",
       "list": [
         { "id": "claude-sonnet-4-5@20250929", "name": "Claude Sonnet 4.5" }
       ]
+    },
+    "openai:api_key": {
+      "cachedAt": "2025-01-15T09:00:00Z",
+      "list": [
+        { "id": "gpt-4", "name": "GPT-4" }
+      ]
     }
   },
   "searchProvider": "exa"
 }
 ```
 
-The `searchProvider` field stores the selected search provider. Valid values: `exa`, `serper`, `brave`. If not set, defaults to `exa`.
+**Key points:**
+- `connections` stores the active connection for each provider
+  - `authMethod` - Authentication method being used
+  - `method` - Display name (optional, legacy field)
+- `models` uses `"provider:authMethod"` as the key
+  - Supports multiple auth methods for the same provider
+  - Each auth method has its own cached model list
+- `searchProvider` - Selected search provider (`exa`, `serper`, `brave`)
+
+### Migration from Old Format
+
+If you're upgrading from an older version, run the migration script to update your configuration:
+
+```bash
+npm run migrate
+```
+
+The script will:
+1. Convert old format models keys (e.g., `"anthropic"`) to new format (e.g., `"anthropic:vertex"`)
+2. Add `authMethod` field to connections
+3. Create a backup of your old configuration
+
+**Old format:**
+```json
+{
+  "models": {
+    "anthropic": {
+      "provider": "anthropic",
+      "authMethod": "vertex",
+      "cachedAt": "...",
+      "list": [...]
+    }
+  }
+}
+```
+
+**New format:**
+```json
+{
+  "models": {
+    "anthropic:vertex": {
+      "cachedAt": "...",
+      "list": [...]
+    }
+  }
+}
+```
 
 ## Troubleshooting
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gencode-ai",
-  "version": "0.2.0",
+  "version": "0.3.0",
   "description": "An open-source AI assistant for your terminal",
   "type": "module",
   "main": "dist/index.js",
@@ -14,7 +14,8 @@
     "start": "node dist/cli/index.js",
     "start:dev": "npx tsx src/cli/index.tsx",
     "test": "node --experimental-vm-modules node_modules/jest/bin/jest.js",
-    "example": "npx tsx examples/basic.ts"
+    "example": "npx tsx examples/basic.ts",
+    "migrate": "npx tsx scripts/migrate.ts"
   },
   "keywords": [
     "agent",