feat: Add Qwen model recognition tests and documentation

CLAUDE.md

@@ -1,6 +1,6 @@
 ## Project Overview
 
-**Remote Agentic Coding Platform**: Control AI coding assistants (Claude Code SDK, Codex SDK) remotely from Slack, Telegram, and GitHub. Built with **Bun + TypeScript + SQLite/PostgreSQL**, single-developer tool for AI-assisted development practitioners. Architecture prioritizes simplicity, flexibility, and user control.
+**Remote Agentic Coding Platform**: Control AI coding assistants (Claude Code SDK, Codex SDK, Qwen Code SDK) remotely from Slack, Telegram, and GitHub. Built with **Bun + TypeScript + SQLite/PostgreSQL**, single-developer tool for AI-assisted development practitioners. Architecture prioritizes simplicity, flexibility, and user control.
 
 ## Core Principles
 
@@ -443,8 +443,18 @@ import type { DagNode, WorkflowDefinition } from '@/lib/api';
 - Implement `IAssistantClient` interface
 - **ClaudeClient**: `@anthropic-ai/claude-agent-sdk`
 - **CodexClient**: `@openai/codex-sdk`
+- **QwenClient**: `@qwen-code/sdk`
 - Streaming: `for await (const event of events) { await platform.send(event) }`
 
+**Qwen Support Boundaries**
+- Archon uses `@qwen-code/sdk` and can select Qwen as the active assistant when configured.
+- `pathToQwenExecutable` is optional and only needed when pointing at a custom Qwen binary.
+- Archon does not provision Qwen credentials; by default it follows the Qwen Code auth/config
+  already present on the machine unless `assistants.qwen.authType` is set explicitly.
+- Claude-only options do not apply to Qwen: `settingSources`, `effort`, `thinking`,
+  `maxBudgetUsd`, `fallbackModel`, `betas`, `sandbox`, `hooks`, and `skills`.
+- Structured `output_format` is not currently forwarded to Qwen workflow nodes.
+
 ### Configuration
 
 **Environment Variables:**
@@ -470,6 +480,12 @@ assistants:
     additionalDirectories:
       - /absolute/path/to/other/repo
     codexBinaryPath: /usr/local/bin/codex  # Optional: custom Codex CLI binary path
+  qwen:
+    model: qwen-max
+    pathToQwenExecutable: qwen  # Optional: custom Qwen CLI executable path
+    permissionMode: yolo  # 'default' | 'plan' | 'auto-edit' | 'yolo'
+    authType: openai  # 'openai' | 'qwen-oauth'
+    includePartialMessages: true
 
 # docs:
 #   path: docs  # Optional: default is docs/
@@ -483,6 +499,7 @@ assistants:
 **Model Validation:**
 - Workflows are validated at load time for provider/model compatibility
 - Claude models: `sonnet`, `opus`, `haiku`, `claude-*`, `inherit`
+- Qwen models: `qwen-*`, `qwq-*`, `qvq-*`
 - Codex models: Any model except Claude-specific aliases
 - Invalid combinations fail workflow loading with clear error messages
 
@@ -681,7 +698,7 @@ async function createSession(conversationId: string, codebaseId: string) {
 2. **Workflows** (YAML-based):
    - Stored in `.archon/workflows/` (searched recursively)
    - Multi-step AI execution chains, discovered at runtime
-   - **`nodes:` (DAG format)**: Nodes with explicit `depends_on` edges; independent nodes in the same topological layer run concurrently. Node types: `command:` (named command file), `prompt:` (inline prompt), `bash:` (shell script, stdout captured as `$nodeId.output`, no AI), `loop:` (iterative AI prompt until completion signal), `approval:` (human gate; pauses until user approves or rejects; `capture_response: true` stores the user's comment as `$<node-id>.output` for downstream nodes, default false), `script:` (inline TypeScript/Python or named script from `.archon/scripts/`, runs via `bun` or `uv`, stdout captured as `$nodeId.output`, no AI, supports `deps:` for dependency installation and `timeout:` in ms, requires `runtime: bun` or `runtime: uv`) . Supports `when:` conditions, `trigger_rule` join semantics, `$nodeId.output` substitution, `output_format` for structured JSON output (Claude and Codex), `allowed_tools`/`denied_tools` for per-node tool restrictions (Claude only), `hooks` for per-node SDK hook callbacks (Claude only), `mcp` for per-node MCP server config files (Claude only, env vars expanded at execution time), and `skills` for per-node skill preloading via AgentDefinition wrapping (Claude only), and `effort`/`thinking`/`maxBudgetUsd`/`systemPrompt`/`fallbackModel`/`betas`/`sandbox` for Claude SDK advanced options (Claude only, also settable at workflow level)
+   - **`nodes:` (DAG format)**: Nodes with explicit `depends_on` edges; independent nodes in the same topological layer run concurrently. Node types: `command:` (named command file), `prompt:` (inline prompt), `bash:` (shell script, stdout captured as `$nodeId.output`, no AI), `loop:` (iterative AI prompt until completion signal), `approval:` (human gate; pauses until user approves or rejects; `capture_response: true` stores the user's comment as `$<node-id>.output` for downstream nodes, default false), `script:` (inline TypeScript/Python or named script from `.archon/scripts/`, runs via `bun` or `uv`, stdout captured as `$nodeId.output`, no AI, supports `deps:` for dependency installation and `timeout:` in ms, requires `runtime: bun` or `runtime: uv`) . Supports `when:` conditions, `trigger_rule` join semantics, `$nodeId.output` substitution, `output_format` for structured JSON output (Claude and Codex; Qwen support is currently limited), `allowed_tools`/`denied_tools` for per-node tool restrictions (Claude/Codex/Qwen), `hooks` for per-node SDK hook callbacks (Claude only), `mcp` for per-node MCP server config files (Claude/Codex/Qwen), and `skills` for per-node skill preloading via AgentDefinition wrapping (Claude only), and `effort`/`thinking`/`maxBudgetUsd`/`systemPrompt`/`fallbackModel`/`betas`/`sandbox` for Claude SDK advanced options (Claude only, also settable at workflow level)
    - Provider inherited from `.archon/config.yaml` unless explicitly set; per-node `provider` and `model` overrides supported
    - Model and options can be set per workflow or inherited from config defaults
    - `interactive: true` at the workflow level forces foreground execution on web (required for approval-gate workflows in the web UI)

README.md

@@ -268,7 +268,7 @@ The Web UI and CLI work out of the box. Optionally connect a chat platform for r
       ▼                ▼          ▼                ▼
 ┌───────────┐  ┌────────────┐  ┌──────────────────────────┐
 │  Command  │  │  Workflow  │  │    AI Assistant Clients  │
-│  Handler  │  │  Executor  │  │      (Claude / Codex)    │
+│  Handler  │  │  Executor  │  │   (Claude / Codex / Qwen)│
 │  (Slash)  │  │  (YAML)    │  │                          │
 └───────────┘  └────────────┘  └──────────────────────────┘
       │              │                      │
@@ -294,11 +294,72 @@ Full documentation is available at **[archon.diy](https://archon.diy)**.
 | [Authoring Workflows](https://archon.diy/guides/authoring-workflows/) | Create custom YAML workflows |
 | [Authoring Commands](https://archon.diy/guides/authoring-commands/) | Create reusable AI commands |
 | [Configuration](https://archon.diy/reference/configuration/) | All config options, env vars, YAML settings |
-| [AI Assistants](https://archon.diy/getting-started/ai-assistants/) | Claude and Codex setup details |
+| [AI Assistants](https://archon.diy/getting-started/ai-assistants/) | Claude, Codex, and Qwen setup details |
 | [Deployment](https://archon.diy/deployment/) | Docker, VPS, production setup |
 | [Architecture](https://archon.diy/reference/architecture/) | System design and internals |
 | [Troubleshooting](https://archon.diy/reference/troubleshooting/) | Common issues and fixes |
 
+### Supported AI Assistants
+
+Archon supports three first-class AI assistants:
+
+| Assistant | SDK | Best For | Setup Guide |
+|-----------|-----|----------|-------------|
+| **Claude** | Claude Code SDK | General-purpose coding, complex reasoning | [Claude Setup](https://archon.diy/getting-started/ai-assistants/) |
+| **Codex** | Codex CLI | OpenAI ecosystem integration | [Codex Setup](https://archon.diy/getting-started/ai-assistants/) |
+| **Qwen** | Qwen Code SDK | Code generation, cost-effective workflows | [Qwen Setup](./docs/qwen-setup.md) |
+
+All three assistants support:
+- ✅ Full workflow execution (plan, implement, review, PR creation)
+- ✅ Streaming responses with real-time output
+- ✅ Tool use and MCP server integration
+- ✅ Worktree isolation for parallel development
+- ✅ Session resume capability
+- ✅ AI-powered title generation
+
+You can mix assistants within a single workflow:
+
+```yaml
+nodes:
+  - id: plan
+    prompt: "Create implementation plan"
+    model: qwen-max  # Use Qwen for planning
+
+  - id: implement
+    depends_on: [plan]
+    prompt: "Implement the plan"
+    model: claude-sonnet-4-5  # Use Claude for coding
+
+  - id: review
+    depends_on: [implement]
+    prompt: "Review changes"
+    model: qwen-coder  # Use Qwen for code review
+```
+
+### Qwen Support Details
+
+For comprehensive Qwen setup and configuration, see the **[Qwen Setup Guide](./docs/qwen-setup.md)**.
+
+Quick highlights:
+- **Authentication**: Follows your existing Qwen Code CLI auth (OAuth or API key)
+- **Models**: Auto-detects `qwen-max`, `qwen-coder`, `qwen-turbo`, `qwen-plus`, `qwq-*`, `qvq-*`, and more
+- **Configuration**: Set via `~/.archon/config.yaml` under `assistants.qwen`
+- **Per-workflow override**: Specify `model: qwen-max` in any workflow node
+
+### Qwen Support Boundaries
+
+Archon integrates Qwen through `@qwen-code/sdk` and can select it as the active assistant
+when configured globally, per repository, or per workflow node.
+
+- Qwen works with Archon's normal assistant selection and MCP wiring.
+- `pathToQwenExecutable` is optional; use it only if you want to point Archon at a specific
+  Qwen CLI binary.
+- Archon does not provision Qwen credentials for you; by default it follows your existing
+  Qwen Code auth/config unless you explicitly override `assistants.qwen.authType`.
+- Claude-only workflow features do not apply to Qwen: `settingSources`, `effort`,
+  `thinking`, `maxBudgetUsd`, `fallbackModel`, `betas`, `sandbox`, `hooks`, and `skills`.
+- Structured `output_format` is not currently forwarded to Qwen workflow nodes.
+
 ## Contributing
 
 Contributions welcome! See the open [issues](https://github.com/coleam00/Archon/issues) for things to work on.