converse-mcp-server 2.26.1 → 2.27.0

package/README.md

@@ -161,7 +161,7 @@ Get multiple AI models to analyze the same question simultaneously. Each model c
 // Asynchronous consensus (for complex analysis)
 {
   "prompt": "Review our system architecture and provide comprehensive recommendations",
-  "models": ["gpt-5", "gemini-2.5-pro", "claude-sonnet-4"],
+  "models": ["gpt-5", "gemini-2.5-pro", "claude-sonnet-4-6"],
   "files": ["/path/to/architecture-docs"],
   "async": true,          // Run in background
   "enable_cross_feedback": true
@@ -268,11 +268,12 @@ SUMMARIZATION_MODEL=gpt-5-nano        # Default: gpt-5-nano
 
 ### Anthropic Models
 
-- **claude-opus-4.1**: Highest intelligence with extended thinking (200K context)
-- **claude-sonnet-4**: Balanced performance with extended thinking (200K context)
-- **claude-3.7-sonnet**: Enhanced 3.x generation with thinking (200K context)
-- **claude-3.5-sonnet**: Fast and intelligent (200K context)
-- **claude-3.5-haiku**: Fastest model for simple queries (200K context)
+- **claude-fable-5** (alias: `fable`): Most capable model for demanding reasoning and long-horizon agentic work (1M context, 128K output)
+- **claude-opus-4-8** (alias: `opus`): Most capable Opus for complex reasoning and agentic coding (200K context, 1M via beta, 128K output)
+- **claude-opus-4-7** / **claude-opus-4-6**: Previous Opus generations with adaptive thinking (128K output)
+- **claude-opus-4-5** / **claude-opus-4-1**: Legacy Opus models with extended thinking (64K / 32K output)
+- **claude-sonnet-4-6** (alias: `sonnet`): Best combination of speed and intelligence with adaptive thinking (64K output)
+- **claude-haiku-4-5** (alias: `haiku`): Fast and intelligent for simple queries (64K output)
 
 ### Mistral Models
 
@@ -300,6 +301,14 @@ SUMMARIZATION_MODEL=gpt-5-nano        # Default: gpt-5-nano
   - Requires ChatGPT login or CODEX_API_KEY
   - See [Configuration](#configuration) for sandbox and approval settings
 
+### Claude Agent SDK Models
+
+- **claude** (aliases: `claude-sdk`, `claude-code`): Claude via the Claude Agent SDK
+  - Defaults to Claude Fable 5; pick a specific model with `claude:fable` or `claude:opus`
+  - Uses Claude Code CLI authentication (`claude login`) - no API key needed
+  - Direct filesystem access from working directory
+  - Unknown `claude:`-prefixed names pass through to the SDK (e.g. `claude:claude-sonnet-4-6`)
+
 ## 📚 Help & Documentation
 
 ### Built-in Help
@@ -398,6 +407,13 @@ Use `"auto"` for automatic model selection, or specify exact models:
 "pro"; // -> gemini-2.5-pro
 "grok"; // -> grok-4-0709
 "grok-4"; // -> grok-4-0709
+"fable"; // -> claude-fable-5 (Anthropic API)
+"opus"; // -> claude-opus-4-8 (Anthropic API)
+
+// SDK providers (subscription-based, no API key)
+"claude"; // -> Claude Agent SDK (Claude Fable 5)
+"claude:opus"; // -> Claude Agent SDK (Claude Opus 4.8)
+"copilot:codex"; // -> GitHub Copilot SDK (gpt-5.3-codex)
 ```
 
 **Auto Model Behavior:**
@@ -405,17 +421,21 @@ Use `"auto"` for automatic model selection, or specify exact models:
 - **Chat Tool**: Selects the first available provider and uses its default model
 - **Consensus Tool**: When using `["auto"]`, automatically expands to the first 3 available providers
 
-Provider priority order (requires corresponding API key):
-
-1. OpenAI (`gpt-5`)
-2. Google (`gemini-2.5-pro`)
-3. XAI (`grok-4`)
-4. Anthropic (`claude-sonnet-4-20250514`)
-5. Mistral (`magistral-medium-2506`)
-6. DeepSeek (`deepseek-reasoner`)
-7. OpenRouter (`qwen/qwen3-coder`)
-
-The system will use the first 3 providers that have valid API keys configured. This enables automatic multi-model consensus without manually specifying models.
+Provider priority order (subscription-based SDK providers first, then API-key providers):
+
+1. Codex (`codex`)
+2. Gemini CLI (`gemini`)
+3. Claude Agent SDK (`claude` → Claude Fable 5)
+4. Copilot (`copilot`)
+5. OpenAI (`gpt-5`)
+6. Google (`gemini-pro`)
+7. XAI (`grok-4-0709`)
+8. Anthropic (`claude-sonnet-4-20250514`)
+9. Mistral (`magistral-medium-2506`)
+10. DeepSeek (`deepseek-reasoner`)
+11. OpenRouter (`qwen/qwen3-coder`)
+
+The system will use the first 3 providers that are available (authenticated SDK or valid API key). This enables automatic multi-model consensus without manually specifying models.
 
 ### Advanced Configuration
 

package/docs/API.md

@@ -547,11 +547,17 @@ When complete, `check_status` for the continuation_id renders the full lap trans
 
 | Model | Alias | Context | Tokens | Features | Use Cases |
 |-------|-------|---------|--------|----------|-----------|
-| `claude-opus-4-5-20250220` | `opus-4.5`, `opus` | 200K | 32K | Extended thinking, images, caching | Most capable reasoning |
-| `claude-opus-4-1-20250805` | `opus-4.1`, `opus-4` | 200K | 32K | Extended thinking, images, caching | Complex reasoning tasks |
-| `claude-sonnet-4-5-20250929` | `sonnet-4.5`, `sonnet` | 200K | 64K | Extended thinking, images, caching | Enhanced reasoning |
-| `claude-sonnet-4-20250514` | `sonnet-4` | 200K | 64K | Extended thinking, images, caching | High performance, balanced |
-| `claude-haiku-4-5-20251001` | `haiku-4.5`, `haiku` | 200K | 64K | Extended thinking, caching | Fast and intelligent |
+| `claude-fable-5` | `fable`, `fable-5` | 1M | 128K | Adaptive thinking, effort, images, caching, compaction | Most demanding reasoning, long-horizon agentic work |
+| `claude-opus-4-8` | `opus`, `opus-4.8` | 200K (1M beta) | 128K | Adaptive thinking, effort, images, caching, compaction | Complex reasoning, agentic coding |
+| `claude-opus-4-7` | `opus-4.7` | 200K (1M beta) | 128K | Adaptive thinking, effort, images, caching, compaction | Previous Opus generation |
+| `claude-opus-4-6` | `opus-4.6` | 200K (1M beta) | 128K | Adaptive thinking, effort, images, caching, compaction | Previous Opus generation |
+| `claude-opus-4-5-20251101` | `opus-4.5` | 200K | 64K | Extended thinking, effort (beta), images, caching | Legacy Opus |
+| `claude-opus-4-1-20250805` | `opus-4.1`, `opus-4` | 200K | 32K | Extended thinking, images, caching | Legacy Opus |
+| `claude-sonnet-4-6` | `sonnet`, `sonnet-4.6` | 200K (1M beta) | 64K | Adaptive thinking, effort, images, caching, compaction | Best speed/intelligence balance |
+| `claude-sonnet-4-5-20250929` | `sonnet-4.5` | 200K (1M beta) | 64K | Extended thinking, images, caching | Legacy Sonnet |
+| `claude-haiku-4-5-20251001` | `haiku`, `haiku-4.5` | 200K | 64K | Extended thinking, images, caching | Fast and intelligent |
+
+**Note:** Claude Fable 5 does not accept the `temperature` parameter (it is silently omitted). Models with adaptive thinking control thinking depth via `reasoning_effort`, which maps to Anthropic's `effort` parameter.
 
 **Prompt Caching (Always Enabled):**
 - System prompts are automatically cached for 1 hour using Anthropic's prompt caching
@@ -592,6 +598,16 @@ When complete, `check_status` for the continuation_id renders the full lap trans
 - **Response times**: 6-20 seconds typical (complex tasks may take minutes)
 - **Authentication**: Requires ChatGPT login OR `CODEX_API_KEY` environment variable
 
+### Claude Agent SDK Models
+
+**Claude** is also available through the Claude Agent SDK, using Claude Code CLI authentication instead of an API key:
+
+- **Model**: `claude` (aliases: `claude-sdk`, `claude-code`) - defaults to Claude Fable 5
+- **Model selection**: `claude:fable` (Claude Fable 5) or `claude:opus` (Claude Opus 4.8); unknown `claude:`-prefixed names pass through to the SDK (e.g. `claude:claude-sonnet-4-6`)
+- **Authentication**: Claude Code login (`claude login`) - no `ANTHROPIC_API_KEY` needed
+- **Direct file access**: Reads files from working directory
+- **Note**: `temperature`, `use_websearch`, and `reasoning_effort` are managed by the SDK (ignored if specified)
+
 ### Gemini CLI Models (OAuth-based)
 
 **Gemini CLI** provides subscription-based access to Gemini models through OAuth:
@@ -657,6 +673,13 @@ Use `"auto"` for automatic selection or specify exact models:
 {"model": "pro"}    // -> gemini-2.5-pro  
 {"model": "grok"}   // -> grok-4-0709
 {"model": "grok-4"}  // -> grok-4-0709
+{"model": "fable"}  // -> claude-fable-5 (Anthropic API)
+{"model": "opus"}   // -> claude-opus-4-8 (Anthropic API)
+
+// SDK providers (subscription-based)
+{"model": "claude"}        // -> Claude Agent SDK (Claude Fable 5)
+{"model": "claude:opus"}   // -> Claude Agent SDK (Claude Opus 4.8)
+{"model": "copilot:codex"} // -> GitHub Copilot SDK (gpt-5.3-codex)
 ```
 
 ## Configuration