refactor: restrict fallback_models to preset-only and clean up provider factory

- Restrict fallback_models to only reference preset names in model_presets.
- Add schema validation to reject unknown preset names in fallback_models.
- Remove build_provider_for_model() since bare model fallback is no longer supported.
- Simplify make_provider_factory() to only look up presets by name.
- Update onboard UI to remove "Add custom model" option from fallback chain.
- Update tests to use preset names instead of bare model strings in fallback chains.
- Fix test imports referencing deleted _make_provider function.

README.md

@@ -123,6 +123,7 @@
 - **Ultra-lightweight**: stable long-running agent behavior with a small, readable core.
 - **Research-ready**: the codebase is intentionally simple enough to study, modify, and extend.
 - **Practical**: chat channels, API, memory, MCP, and deployment paths are already built in.
+- **Runtime model switching**: define [model presets](docs/configuration.md#model-presets) and switch between cheap/fast and powerful models mid-conversation — no restart required.
 - **Hackable**: you can start fast, then go deeper through repo docs instead of a monolithic landing page.
 
 ## 📦 Install

docs/configuration.md

@@ -656,6 +656,146 @@ That's it! Environment variables, model routing, config matching, and `nanobot s
 
 </details>
 
+## Agent Settings
+
+### Model Presets
+
+Model presets let you define **named bundles** of model + generation parameters and switch between them instantly — no restart required.
+
+> [!NOTE]
+> Config fields in `config.json` use **camelCase** (`modelPreset`, `contextWindowTokens`).
+> The [`my` tool](./my-tool.md) uses **snake_case** (`model_preset`, `context_window_tokens`).
+> Both refer to the same thing — just different naming conventions for config vs. runtime API.
+
+**Why use presets?**
+- Switch between a cheap/fast model and a powerful model mid-conversation.
+- Share the same config across different tasks without manually editing `model`, `provider`, `temperature`, etc.
+- Runtime switching via the [`my` tool](./my-tool.md).
+
+> [!TIP]
+> The easiest way to set up presets and fallback models is through the interactive wizard:
+> ```bash
+> nanobot onboard --wizard
+> ```
+> Choose **"[M] Model Presets"** to create, edit, or delete presets interactively.
+
+**Configuration example:**
+
+```json
+{
+  "modelPresets": {
+    "fast": {
+      "model": "gpt-4.1-mini",
+      "provider": "openai",
+      "maxTokens": 4096,
+      "contextWindowTokens": 128000,
+      "temperature": 0.3
+    },
+    "deep": {
+      "model": "claude-opus-4-7",
+      "provider": "anthropic",
+      "maxTokens": 8192,
+      "contextWindowTokens": 200000,
+      "temperature": 0.1,
+      "reasoningEffort": "high"
+    }
+  },
+  "agents": {
+    "defaults": {
+      "modelPreset": "fast"
+    }
+  }
+}
+```
+
+**Preset fields:**
+
+| Field | Type | Default | Description |
+|-------|------|---------|-------------|
+| `model` | string | *(required)* | Model identifier, e.g. `anthropic/claude-opus-4-7` or `gpt-4.1` |
+| `provider` | string | `"auto"` | Provider name or `"auto"` to infer from the model string |
+| `maxTokens` | integer | `8192` | Max completion tokens per turn |
+| `contextWindowTokens` | integer | `65536` | Context window size for token budgeting |
+| `temperature` | float | `0.1` | Sampling temperature |
+| `reasoningEffort` | string or null | `null` | Thinking mode: `low`, `medium`, `high`, `adaptive` |
+
+**How it works:**
+- When `modelPreset` is set, the preset **completely overrides** all model-specific fields in `agents.defaults`.
+- When `modelPreset` is omitted, nanobot automatically creates an implicit `"default"` preset from your existing `agents.defaults.model`, `provider`, `temperature`, etc. — **zero migration required** for existing configs.
+
+**Runtime switching** (requires `tools.my.allowSet: true`):
+
+```text
+my(action="set", key="model_preset", value="deep")
+```
+
+This atomically swaps the model, provider, generation parameters, and context window for the next turn.
+
+If the preset name does not exist, the agent receives an error such as `model_preset 'unknown' not found. Available: fast, deep`.
+
+> [!NOTE]
+> Directly modifying `model` or `contextWindowTokens` via `my(action="set", key="model", ...)` still works, but it automatically clears the active preset because the live state no longer matches the preset bundle. Use `model_preset` for atomic switches instead.
+
+See [`my-tool.md`](./my-tool.md) for more runtime examples.
+
+---
+
+### Fallback Models
+
+When the primary model returns a transient error (rate limit, server overload, quota exhausted), nanobot can automatically fail over to a chain of backup models.
+
+**Configuration example:**
+
+```json
+{
+  "agents": {
+    "defaults": {
+      "modelPreset": "fast",
+      "fallbackModels": ["deep", "backup"]
+    }
+  }
+}
+```
+
+**How it works:**
+1. nanobot tries the primary model first (the one from the active preset).
+2. The provider retries transient errors internally (e.g. 3 attempts with exponential backoff for 503/429).
+3. Only after the provider's own retries are exhausted and the final response still has `finish_reason == "error"` with a retryable error kind, nanobot moves to the next candidate in `fallbackModels`.
+4. Each candidate must be a preset name defined in `modelPresets`. The preset's full config (model, provider, generation params) is used.
+5. If all candidates are exhausted, the final error is returned to the user.
+
+**Failover triggers on:**
+- `server_error` (503, 502, 500)
+- `rate_limit` (429)
+- `insufficient_quota` / `quota_exhausted` (429)
+
+**Failover does NOT trigger on:**
+- Authentication errors (401) — rotating to another model with the same key won't help
+- Invalid request errors (400) — the request itself is malformed
+
+> [!TIP]
+> Fallback models must reference preset names defined in `modelPresets`. Define a preset for each fallback model you want to use: `["cheap-preset", "backup", "emergency"]`.
+
+---
+
+### Other Agent Defaults
+
+| Option | Type | Default | Description |
+|--------|------|---------|-------------|
+| `agents.defaults.model` | string | `"anthropic/claude-opus-4-5"` | Default model when no preset is active |
+| `agents.defaults.provider` | string | `"auto"` | Default provider when no preset is active |
+| `agents.defaults.maxTokens` | integer | `8192` | Max completion tokens when no preset is active |
+| `agents.defaults.temperature` | float | `0.1` | Sampling temperature when no preset is active |
+| `agents.defaults.reasoningEffort` | string or null | `null` | Thinking mode when no preset is active |
+| `agents.defaults.maxToolIterations` | integer | `200` | Max tool calls per conversation turn |
+| `agents.defaults.maxToolResultChars` | integer | `16000` | Max characters per tool result |
+| `agents.defaults.providerRetryMode` | string | `"standard"` | `"standard"` or `"persistent"` — how aggressively to retry provider-level errors |
+| `agents.defaults.timezone` | string | `"UTC"` | IANA timezone for runtime context |
+| `agents.defaults.unifiedSession` | boolean | `false` | Share one session across all channels |
+| `agents.defaults.sessionTtlMinutes` | integer | `0` | Auto-compact idle threshold (0 = disabled) |
+| `agents.defaults.maxMessages` | integer | `120` | Max messages to replay from session history |
+| `agents.defaults.consolidationRatio` | float | `0.5` | Target ratio retained after context compression |
+
 ## Channel Settings
 
 Global settings that apply to all channels. Configure under the `channels` section in `~/.nanobot/config.json`:

docs/my-tool.md

@@ -12,6 +12,11 @@ My tool fills this gap. With it, the agent can:
 - **Adapt on the fly**: Complex task? Expand the context window. Simple chat? Switch to a faster model.
 - **Remember across turns**: Store notes in your scratchpad that persist into the next conversation turn.
 
+> [!NOTE]
+> This tool uses **snake_case** keys (`model_preset`, `context_window_tokens`).
+> The matching config fields in `config.json` are **camelCase** (`modelPreset`, `contextWindowTokens`).
+> See [`configuration.md`](./configuration.md#model-presets) for how to define presets in your config.
+
 ## Configuration
 
 Enabled by default (read-only mode). The agent can check its state but not set it.
@@ -39,8 +44,7 @@ Without parameters, returns a key config overview:
 ```text
 my(action="check")
 # → max_iterations: 40
-#   context_window_tokens: 65536
+#   model_preset: 'fast'
-#   model: 'anthropic/claude-sonnet-4-20250514'
 #   workspace: PosixPath('/tmp/workspace')
 #   provider_retry_mode: 'standard'
 #   max_tool_result_chars: 16000
@@ -55,8 +59,13 @@ With a key parameter, drill into a specific config:
 my(action="check", key="_last_usage.prompt_tokens")
 # → How many prompt tokens I've used so far
 
-my(action="check", key="model")
+my(action="check", key="model_preset")
-# → What model I'm currently running on
+# → Current active preset name (e.g. 'fast')
+
+my(action="check", key="model_presets")
+# → Lists all preset names and their models, e.g.:
+#   fast → gpt-4.1-mini (openai)
+#   deep → claude-opus-4-7 (anthropic)
 
 my(action="check", key="web_config.enable")
 # → Whether web search is enabled
@@ -66,7 +75,7 @@ my(action="check", key="web_config.enable")
 
 | Scenario | How |
 |----------|-----|
-| "What model are you using?" | `check("model")` |
+| "What model are you using?" | `check("model_preset")` |
 | "How many more tool calls can you make?" | `check("max_iterations")` minus `check("_current_iteration")` |
 | "How many tokens has this conversation used?" | `check("_last_usage")` — cumulative across all turns |
 | "Where is your working directory?" | `check("workspace")` |
@@ -83,8 +92,11 @@ Changes take effect immediately, no restart required.
 my(action="set", key="max_iterations", value=80)
 # → Bump iteration limit from 40 to 80
 
-my(action="set", key="model", value="fast-model")
+my(action="set", key="model_preset", value="fast")
-# → Switch to a faster model
+# → Switch to the 'fast' preset (model, provider, temperature, etc. all at once)
+#
+# If the preset name does not exist:
+# → Error: model_preset 'unknown' not found. Available: fast, deep
 
 my(action="set", key="context_window_tokens", value=131072)
 # → Expand context window for long documents
@@ -101,15 +113,17 @@ my(action="set", key="task_complexity", value="high")
 
 ### Protected parameters
 
-These parameters have type and range validation — invalid values are rejected:
+These parameters have validation — invalid values are rejected:
 
-| Parameter | Type | Range | Purpose |
+| Parameter | Type | Range / Constraint | Purpose |
-|-----------|------|-------|---------|
+|-----------|------|-------------------|---------|
 | `max_iterations` | int | 1–100 | Max tool calls per conversation turn |
-| `context_window_tokens` | int | 4,096–1,000,000 | Context window size |
+| `model_preset` | str | must exist in `model_presets` | Switch to a named preset bundle |
-| `model` | str | non-empty | LLM model to use |
 
-Other parameters (e.g. `workspace`, `provider_retry_mode`, `max_tool_result_chars`) can be set freely, as long as the value is JSON-safe.
+Other parameters (e.g. `model`, `context_window_tokens`, `workspace`, `provider_retry_mode`, `max_tool_result_chars`) can be set freely, as long as the value is JSON-safe.
+
+> [!NOTE]
+> Setting `model` or `context_window_tokens` directly automatically clears the active `model_preset`, because the live state no longer matches the preset bundle. Use `model_preset` for atomic switches instead.
 
 ---
 
@@ -125,8 +139,8 @@ Agent: This codebase is large, let me expand my context window to handle it.
 ### "Simple question, don't waste compute"
 
 ```text
-Agent: This is a straightforward question, let me switch to a faster model.
+Agent: This is a straightforward question, let me switch to the fast preset.
-→ my(action="set", key="model", value="fast-model")
+→ my(action="set", key="model_preset", value="fast")
 ```
 
 ### "Remember user preferences across turns"

docs/quick-start.md

@@ -95,6 +95,8 @@ Configure these **two parts** in your config (other options have defaults).
 }
 ```
 
+*Want to switch models mid-conversation?* Define [`modelPresets`](./configuration.md#model-presets) and switch instantly with `my(action="set", key="model_preset", value="fast")`.
+
 **3. Chat**
 
 ```bash

nanobot/agent/loop.py

@@ -41,7 +41,7 @@ from nanobot.agent.tools.web import WebFetchTool, WebSearchTool
 from nanobot.bus.events import InboundMessage, OutboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.command import CommandContext, CommandRouter, register_builtin_commands
-from nanobot.config.schema import AgentDefaults
+from nanobot.config.schema import AgentDefaults, ModelPresetConfig
 from nanobot.providers.base import LLMProvider
 from nanobot.providers.factory import ProviderSnapshot
 from nanobot.session.manager import Session, SessionManager
@@ -188,6 +188,50 @@ class AgentLoop:
     _RUNTIME_CHECKPOINT_KEY = "runtime_checkpoint"
     _PENDING_USER_TURN_KEY = "pending_user_turn"
 
+    @classmethod
+    def from_config(
+        cls,
+        config: Any,
+        bus: MessageBus | None = None,
+        **extra: Any,
+    ) -> AgentLoop:
+        """Create an AgentLoop from config with the common parameter set."""
+        from nanobot.providers.factory import build_provider_for_preset, make_provider_factory
+
+        if bus is None:
+            bus = MessageBus()
+        defaults = config.agents.defaults
+        resolved_preset = config.resolve_preset()
+        provider = build_provider_for_preset(config, resolved_preset)
+        return cls(
+            bus=bus,
+            provider=provider,
+            workspace=config.workspace_path,
+            model=resolved_preset.model,
+            max_iterations=defaults.max_tool_iterations,
+            context_window_tokens=resolved_preset.context_window_tokens,
+            context_block_limit=defaults.context_block_limit,
+            max_tool_result_chars=defaults.max_tool_result_chars,
+            provider_retry_mode=defaults.provider_retry_mode,
+            fallback_presets=defaults.fallback_presets,
+            provider_factory=make_provider_factory(config),
+            web_config=config.tools.web,
+            exec_config=config.tools.exec,
+            restrict_to_workspace=config.tools.restrict_to_workspace,
+            mcp_servers=config.tools.mcp_servers,
+            channels_config=config.channels,
+            timezone=defaults.timezone,
+            unified_session=defaults.unified_session,
+            disabled_skills=defaults.disabled_skills,
+            session_ttl_minutes=defaults.session_ttl_minutes,
+            consolidation_ratio=defaults.consolidation_ratio,
+            max_messages=defaults.max_messages,
+            tools_config=config.tools,
+            model_presets=config.model_presets,
+            model_preset=defaults.model_preset,
+            **extra,
+        )
+
     def __init__(
         self,
         bus: MessageBus,
@@ -200,6 +244,8 @@ class AgentLoop:
         max_tool_result_chars: int | None = None,
         provider_retry_mode: str = "standard",
         tool_hint_max_length: int | None = None,
+        fallback_presets: list[str] | None = None,
+        provider_factory: Callable[[str], LLMProvider] | None = None,
         web_config: WebToolsConfig | None = None,
         exec_config: ExecToolConfig | None = None,
         cron_service: CronService | None = None,
@@ -217,6 +263,8 @@ class AgentLoop:
         tools_config: ToolsConfig | None = None,
         provider_snapshot_loader: Callable[[], ProviderSnapshot] | None = None,
         provider_signature: tuple[object, ...] | None = None,
+        model_presets: dict[str, ModelPresetConfig] | None = None,
+        model_preset: str | None = None,
     ):
         from nanobot.config.schema import ExecToolConfig, ToolsConfig, WebToolsConfig
 
@@ -224,7 +272,12 @@ class AgentLoop:
         defaults = AgentDefaults()
         self.bus = bus
         self.channels_config = channels_config
-        self.provider = provider
+        self.provider_factory = provider_factory
+        self.fallback_presets = fallback_presets or []
+        wrapped_provider = self._wrap_with_failover(
+            provider, model or provider.get_default_model()
+        )
+        self.provider = wrapped_provider
         self._provider_snapshot_loader = provider_snapshot_loader
         self._provider_signature = provider_signature
         self.workspace = workspace
@@ -262,9 +315,9 @@ class AgentLoop:
         # One file-read/write tracker per logical session. The tool registry is
         # shared by this loop, so tools resolve the active state via contextvars.
         self._file_state_store = FileStateStore()
-        self.runner = AgentRunner(provider)
+        self.runner = AgentRunner(wrapped_provider)
         self.subagents = SubagentManager(
-            provider=provider,
+            provider=wrapped_provider,
             workspace=workspace,
             bus=bus,
             model=self.model,
@@ -296,13 +349,13 @@ class AgentLoop:
         )
         self.consolidator = Consolidator(
             store=self.context.memory,
-            provider=provider,
+            provider=wrapped_provider,
             model=self.model,
             sessions=self.sessions,
             context_window_tokens=self.context_window_tokens,
             build_messages=self.context.build_messages,
             get_tool_definitions=self.tools.get_definitions,
-            max_completion_tokens=provider.generation.max_tokens,
+            max_completion_tokens=wrapped_provider.generation.max_tokens,
             consolidation_ratio=consolidation_ratio,
         )
         self.auto_compact = AutoCompact(
@@ -312,9 +365,13 @@ class AgentLoop:
         )
         self.dream = Dream(
             store=self.context.memory,
-            provider=provider,
+            provider=wrapped_provider,
             model=self.model,
         )
+        self.model_presets: dict[str, ModelPresetConfig] = model_presets or {}
+        self._active_preset: str | None = (
+            model_preset if model_preset in self.model_presets else None
+        )
         self._register_default_tools()
         if _tc.my.enable:
             self.tools.register(MyTool(loop=self, modify_allowed=_tc.my.allow_set))
@@ -327,6 +384,38 @@ class AgentLoop:
         """Keep subagent runtime limits aligned with mutable loop settings."""
         self.subagents.max_iterations = self.max_iterations
 
+    def _wrap_with_failover(self, provider: LLMProvider, model: str) -> LLMProvider:
+        """Wrap provider with failover router when fallback_presets are configured."""
+        if not self.fallback_presets or not self.provider_factory:
+            return provider
+        from nanobot.providers.failover import ModelRouter
+
+        if isinstance(provider, ModelRouter):
+            return provider
+
+        return ModelRouter(
+            primary_provider=provider,
+            primary_model=model,
+            fallback_presets=self.fallback_presets,
+            provider_factory=self.provider_factory,
+        )
+
+    def _apply_provider_state(
+        self,
+        provider: LLMProvider,
+        model: str,
+        context_window_tokens: int,
+    ) -> None:
+        """Push provider/model/context_window to all LLM-consuming subsystems."""
+        self.provider = provider
+        # Bypass property setters so internal updates don't clear _active_preset.
+        object.__setattr__(self, "_model", model)
+        object.__setattr__(self, "_context_window_tokens", context_window_tokens)
+        self.runner.provider = provider
+        self.subagents.set_provider(provider, model)
+        self.consolidator.set_provider(provider, model, context_window_tokens)
+        self.dream.set_provider(provider, model)
+
     def _apply_provider_snapshot(self, snapshot: ProviderSnapshot) -> None:
         """Swap model/provider for future turns without disturbing an active one."""
         provider = snapshot.provider
@@ -335,14 +424,13 @@ class AgentLoop:
         if self.provider is provider and self.model == model:
             return
         old_model = self.model
-        self.provider = provider
+        provider = self._wrap_with_failover(provider, model)
-        self.model = model
+        self._apply_provider_state(provider, model, context_window_tokens)
-        self.context_window_tokens = context_window_tokens
-        self.runner.provider = provider
-        self.subagents.set_provider(provider, model)
-        self.consolidator.set_provider(provider, model, context_window_tokens)
-        self.dream.set_provider(provider, model)
         self._provider_signature = snapshot.signature
+        if self._active_preset:
+            preset = self.model_presets.get(self._active_preset)
+            if preset and preset.model != model:
+                self._active_preset = None
         logger.info("Runtime model switched for next turn: {} -> {}", old_model, model)
 
     def _refresh_provider_snapshot(self) -> None:
@@ -357,6 +445,58 @@ class AgentLoop:
             return
         self._apply_provider_snapshot(snapshot)
 
+    # -- model / context_window_tokens properties with preset invalidation --
+
+    @property
+    def model(self) -> str:
+        return self._model
+
+    @model.setter
+    def model(self, value: str) -> None:
+        self._model = value
+        if hasattr(self, "_active_preset"):
+            self._active_preset = None
+
+    @property
+    def context_window_tokens(self) -> int:
+        return self._context_window_tokens
+
+    @context_window_tokens.setter
+    def context_window_tokens(self, value: int) -> None:
+        self._context_window_tokens = value
+        if hasattr(self, "_active_preset"):
+            self._active_preset = None
+
+    # -- model_preset property --
+
+    @property
+    def model_preset(self) -> str | None:
+        return self._active_preset
+
+    @model_preset.setter
+    def model_preset(self, name: str) -> None:
+        """Resolve a preset by name and apply all fields."""
+        if not isinstance(name, str) or not name.strip():
+            raise ValueError("model_preset must be a non-empty string")
+        if name not in self.model_presets:
+            raise KeyError(
+                f"model_preset {name!r} not found. Available: {', '.join(self.model_presets) or '(none)'}"
+            )
+        if self.provider_factory is None:
+            raise ValueError("provider_factory is not configured; cannot switch model preset")
+
+        p = self.model_presets[name]
+        new_provider = self._wrap_with_failover(self.provider_factory(name), p.model)
+
+        # Preserve dream model_override if it differs from the current loop model.
+        old_dream_model = self.dream.model
+        dream_had_override = old_dream_model != self.model
+
+        self._apply_provider_state(new_provider, p.model, p.context_window_tokens)
+        if dream_had_override:
+            self.dream.model = old_dream_model
+        self._active_preset = name
+
     def _register_default_tools(self) -> None:
         """Register the default set of tools."""
         allowed_dir = (

nanobot/agent/tools/self.py

@@ -76,8 +76,6 @@ class MyTool(Tool):
 
     RESTRICTED: dict[str, dict[str, Any]] = {
         "max_iterations":        {"type": int, "min": 1,   "max": 100},
-        "context_window_tokens": {"type": int, "min": 4096, "max": 1_000_000},
-        "model":                 {"type": str, "min_len": 1},
     }
 
     _MAX_RUNTIME_KEYS = 64
@@ -118,13 +116,14 @@ class MyTool(Tool):
             "Scratchpad keys persist across turns but not restarts.\n"
             "Key values: _current_iteration (current progress), "
             "max_iterations - _current_iteration = remaining iterations.\n"
+            "Use 'model_preset' to switch the active model preset.\n"
             "Note: web_config and exec_config are readable but read-only.\n"
             "\n"
             "When to use:\n"
             "- User asks about your model, settings, or token usage → check that key.\n"
             "- A tool fails or behaves unexpectedly → check the related config to diagnose.\n"
             "- User asks you to remember a preference for this session → set to store it in your scratchpad.\n"
-            "- About to start a large task → check context_window_tokens and max_iterations first."
+            "- About to start a large task → check max_iterations and model_preset first."
         )
         if not self._modify_allowed:
             base += "\nREAD-ONLY MODE: set is disabled."
@@ -132,7 +131,7 @@ class MyTool(Tool):
             base += (
                 "\nIMPORTANT: Before setting state, predict the potential impact. "
                 "If the operation could cause crashes or instability "
-                "(e.g. changing model), warn the user first."
+                "(e.g. changing model_preset), warn the user first."
             )
         return base
 
@@ -148,7 +147,7 @@ class MyTool(Tool):
                 },
                 "key": {
                     "type": "string",
-                    "description": "Dot-path for check/set. Examples: 'max_iterations', 'workspace', 'provider_retry_mode'. "
+                    "description": "Dot-path for check/set. Examples: 'max_iterations', 'model_preset', 'provider_retry_mode'. "
                     "For check without key, shows all config values.",
                 },
                 "value": {"description": "New value (for set). Type must match target (int for max_iterations/context_window_tokens, str for model)."},
@@ -330,6 +329,8 @@ class MyTool(Tool):
         # RESTRICTED keys
         for k in self.RESTRICTED:
             parts.append(self._format_value(getattr(loop, k, None), k))
+        # model_preset (property on AgentLoop)
+        parts.append(self._format_value(loop.model_preset, "model_preset"))
         # Other useful top-level keys shown in description
         for k in ("workspace", "provider_retry_mode", "max_tool_result_chars", "_current_iteration", "web_config", "exec_config", "subagents"):
             if _has_real_attr(loop, k):
@@ -386,6 +387,8 @@ class MyTool(Tool):
                 value = expected(value)
             except (ValueError, TypeError):
                 return f"Error: '{key}' must be {expected.__name__}, got {type(value).__name__}"
+
+        # --- existing restricted key logic ---
         old = getattr(self._loop, key)
         if "min" in spec and value < spec["min"]:
             return f"Error: '{key}' must be >= {spec['min']}"
@@ -412,7 +415,14 @@ class MyTool(Tool):
                         f"REJECTED type mismatch {key}: expects {old_t.__name__}, got {new_t.__name__}",
                     )
                     return f"Error: '{key}' expects {old_t.__name__}, got {new_t.__name__}"
-            setattr(self._loop, key, value)
+            # When a model-specific field is set directly, it no longer matches any preset
+            if key in ("model", "context_window_tokens"):
+                self._loop._active_preset = None
+            try:
+                setattr(self._loop, key, value)
+            except (AttributeError, TypeError, ValueError, KeyError) as e:
+                self._audit("modify", f"REJECTED {key}: {e}")
+                return f"Error: {e}"
             self._audit("modify", f"{key}: {old!r} -> {value!r}")
             return f"Set {key} = {value!r} (was {old!r})"
         if callable(value):

nanobot/channels/websocket.py

@@ -160,7 +160,7 @@ def _read_webui_model_name() -> str | None:
     try:
         from nanobot.config.loader import load_config
 
-        model = load_config().agents.defaults.model.strip()
+        model = load_config().resolve_preset().model.strip()
         return model or None
     except Exception as e:
         logger.debug("webui bootstrap could not load model name: {}", e)

nanobot/channels/weixin.py

@@ -11,13 +11,13 @@ from __future__ import annotations
 
 import asyncio
 import base64
+import copy
 import hashlib
 import json
 import os
 import random
 import re
 import time
-import uuid
 from collections import OrderedDict
 from contextlib import suppress
 from pathlib import Path
@@ -54,7 +54,7 @@ MESSAGE_TYPE_BOT = 2
 MESSAGE_STATE_FINISH = 2
 
 WEIXIN_MAX_MESSAGE_LEN = 4000
-WEIXIN_CHANNEL_VERSION = "2.1.1"
+WEIXIN_CHANNEL_VERSION = "2.1.7"
 ILINK_APP_ID = "bot"
 
 
@@ -80,6 +80,36 @@ BASE_INFO: dict[str, str] = {"channel_version": WEIXIN_CHANNEL_VERSION}
 ERRCODE_SESSION_EXPIRED = -14
 SESSION_PAUSE_DURATION_S = 60 * 60
 
+# iLink rate-limit / stale-session errcode
+RATE_LIMIT_ERRCODE = -2
+
+
+def _is_stale_session_ret(
+    ret: int | None,
+    errcode: int | None,
+    errmsg: str | None,
+) -> bool:
+    """True when iLink returns ret=-2 / errcode=-2 that is likely a stale
+    context_token rather than a genuine rate limit.
+
+    Empirically iLink signals these two scenarios weakly:
+    - stale session:      ret=-2, errmsg="unknown error" OR errmsg empty/None
+    - genuine rate limit: ret=-2 with a populated errmsg such as
+      "frequency limit" / "too frequently" / similar
+
+    Treating "unknown error" and empty/None errmsg as stale-session signals
+    lets the caller attempt one tokenless retry. A true rate limit still
+    falls through to the existing retry/backoff path if the tokenless
+    attempt also fails.
+    """
+    if ret != RATE_LIMIT_ERRCODE and errcode != RATE_LIMIT_ERRCODE:
+        return False
+    msg = (errmsg or "").strip().lower()
+    if not msg:
+        return True
+    return msg == "unknown error"
+
+
 # Retry constants (matching the reference plugin's monitor.ts)
 MAX_CONSECUTIVE_FAILURES = 3
 BACKOFF_DELAY_S = 30
@@ -486,6 +516,7 @@ class WeixinChannel(BaseChannel):
             except Exception:
                 if not self._running:
                     break
+                self.logger.exception("WeChat poll loop error")
                 consecutive_failures += 1
                 if consecutive_failures >= MAX_CONSECUTIVE_FAILURES:
                     consecutive_failures = 0
@@ -525,6 +556,22 @@ class WeixinChannel(BaseChannel):
                 f"WeChat session paused, {remaining_min} min remaining (errcode {ERRCODE_SESSION_EXPIRED})"
             )
 
+    def _check_response_error(self, data: dict, operation: str, *, body: dict | None = None) -> None:
+        """Check both ``ret`` and ``errcode`` like the reference TS code.
+
+        The iLink API may signal failure through either field (or both).
+        ``_poll_once`` already checks both; outbound send helpers must do
+        the same to avoid silent drops.
+        """
+        ret = data.get("ret", 0)
+        errcode = data.get("errcode", 0)
+        is_error = (ret is not None and ret != 0) or (errcode is not None and errcode != 0)
+        if not is_error:
+            return
+        raise RuntimeError(
+            f"WeChat {operation} error (ret={ret}, errcode={errcode}): {data.get('errmsg', '')}"
+        )
+
     async def _poll_once(self) -> None:
         remaining = self._session_pause_remaining_s()
         if remaining > 0:
@@ -575,8 +622,10 @@ class WeixinChannel(BaseChannel):
         # Process messages (WeixinMessage[] from types.ts)
         msgs: list[dict] = data.get("msgs", []) or []
         for msg in msgs:
-            with suppress(Exception):
+            try:
                 await self._process_message(msg)
+            except Exception:
+                self.logger.exception("Failed to process WeChat message")
 
     # ------------------------------------------------------------------
     # Inbound message processing  (matches inbound.ts + process-message.ts)
@@ -1089,6 +1138,14 @@ class WeixinChannel(BaseChannel):
         except Exception as e:
             self.logger.debug("typing clear failed for {}: {}", chat_id, e)
 
+    @staticmethod
+    def _generate_client_id() -> str:
+        """Generate a client_id matching the reference plugin format.
+
+        openclaw-weixin uses ``{prefix}:{timestamp}-{8-char hex}``.
+        """
+        return f"nanobot:{int(time.time() * 1000)}-{os.urandom(4).hex()}"
+
     async def _send_text(
         self,
         to_user_id: str,
@@ -1096,7 +1153,7 @@ class WeixinChannel(BaseChannel):
         context_token: str,
     ) -> None:
         """Send a text message matching the exact protocol from send.ts."""
-        client_id = f"nanobot-{uuid.uuid4().hex[:12]}"
+        client_id = self._generate_client_id()
 
         item_list: list[dict] = []
         if text:
@@ -1120,11 +1177,47 @@ class WeixinChannel(BaseChannel):
         }
 
         data = await self._api_post("ilink/bot/sendmessage", body)
+        ret = data.get("ret", 0)
         errcode = data.get("errcode", 0)
-        if errcode and errcode != 0:
+        errmsg = data.get("errmsg", "")
-            raise RuntimeError(
+
-                f"WeChat send text error (code {errcode}): {data.get('errmsg', '')}"
+        # The iLink sendmessage API may return ret=-2 / errcode=-2 for two
+        # different reasons:
+        #   - stale context_token: errmsg is empty/None or "unknown error"
+        #   - genuine rate limit:  errmsg is populated (e.g. "frequency limit")
+        # Per hermes-agent#17228 / #18100, the empty/None variant is a stale
+        # session signal.  Retry once without context_token (iLink accepts
+        # tokenless sends as a degraded fallback).  If the tokenless attempt
+        # also fails, let _check_response_error raise so ChannelManager can
+        # retry with backoff — do NOT swallow the error.
+        if _is_stale_session_ret(ret, errcode, errmsg) and context_token:
+            self.logger.warning(
+                "WeChat send text returned stale-session signal for {} (client_id={}); "
+                "retrying without context_token",
+                to_user_id,
+                client_id,
             )
+            body_no_ctx = copy.deepcopy(body)
+            body_no_ctx["msg"].pop("context_token", None)
+            data = await self._api_post("ilink/bot/sendmessage", body_no_ctx)
+            ret = data.get("ret", 0)
+            errcode = data.get("errcode", 0)
+            errmsg = data.get("errmsg", "")
+            if ret == 0 and (errcode == 0 or errcode is None):
+                self.logger.warning(
+                    "WeChat send text succeeded WITHOUT context_token for {}; "
+                    "clearing expired token from cache",
+                    to_user_id,
+                )
+                self._context_tokens.pop(to_user_id, None)
+                self._save_state()
+                self.logger.debug(
+                    "WeChat text sent to {} (client_id={})", to_user_id, client_id
+                )
+                return
+
+        self._check_response_error(data, "send text", body=body)
+        self.logger.debug("WeChat text sent to {} (client_id={})", to_user_id, client_id)
 
     async def _send_media_file(
         self,
@@ -1250,7 +1343,7 @@ class WeixinChannel(BaseChannel):
             media_item["len"] = str(raw_size)
 
         # Send each media item as its own message (matching reference plugin)
-        client_id = f"nanobot-{uuid.uuid4().hex[:12]}"
+        client_id = self._generate_client_id()
         item_list: list[dict] = [{"type": item_type, item_key: media_item}]
 
         weixin_msg: dict[str, Any] = {
@@ -1270,11 +1363,35 @@ class WeixinChannel(BaseChannel):
         }
 
         data = await self._api_post("ilink/bot/sendmessage", body)
+        ret = data.get("ret", 0)
         errcode = data.get("errcode", 0)
-        if errcode and errcode != 0:
+        errmsg = data.get("errmsg", "")
-            raise RuntimeError(
+
-                f"WeChat send media error (code {errcode}): {data.get('errmsg', '')}"
+        # Same stale-session handling as _send_text (hermes-agent#17228 / #18100).
+        if _is_stale_session_ret(ret, errcode, errmsg) and context_token:
+            self.logger.warning(
+                "WeChat send media returned stale-session signal for {} (client_id={}); "
+                "retrying without context_token",
+                to_user_id,
+                client_id,
             )
+            body_no_ctx = copy.deepcopy(body)
+            body_no_ctx["msg"].pop("context_token", None)
+            data = await self._api_post("ilink/bot/sendmessage", body_no_ctx)
+            ret = data.get("ret", 0)
+            errcode = data.get("errcode", 0)
+            errmsg = data.get("errmsg", "")
+            if ret == 0 and (errcode == 0 or errcode is None):
+                self.logger.warning(
+                    "WeChat send media succeeded WITHOUT context_token for {}; "
+                    "clearing expired token from cache",
+                    to_user_id,
+                )
+                self._context_tokens.pop(to_user_id, None)
+                self._save_state()
+                return
+
+        self._check_response_error(data, "send media", body=body)
 
 
 # ---------------------------------------------------------------------------

nanobot/cli/commands.py

@@ -48,6 +48,7 @@ from rich.table import Table
 from rich.text import Text
 
 from nanobot import __logo__, __version__
+from nanobot.agent.loop import AgentLoop
 
 
 class SafeFileHistory(FileHistory):
@@ -437,20 +438,6 @@ def _onboard_plugins(config_path: Path) -> None:
         json.dump(data, f, indent=2, ensure_ascii=False)
 
 
-def _make_provider(config: Config):
-    """Create the appropriate LLM provider from config.
-
-    Routing is driven by ``ProviderSpec.backend`` in the registry.
-    """
-    from nanobot.providers.factory import make_provider
-
-    try:
-        return make_provider(config)
-    except ValueError as exc:
-        console.print(f"[red]Error: {exc}[/red]")
-        raise typer.Exit(1) from exc
-
-
 def _load_runtime_config(config: str | None = None, workspace: str | None = None) -> Config:
     """Load config and optionally override the active workspace."""
     from nanobot.config.loader import load_config, resolve_config_env_vars, set_config_path
@@ -528,7 +515,6 @@ def serve(
         raise typer.Exit(1)
 
     from loguru import logger
-    from nanobot.agent.loop import AgentLoop
     from nanobot.api.server import create_app
     from nanobot.bus.queue import MessageBus
     from nanobot.session.manager import SessionManager
@@ -545,38 +531,20 @@ def serve(
     timeout = timeout if timeout is not None else api_cfg.timeout
     sync_workspace_templates(runtime_config.workspace_path)
     bus = MessageBus()
-    provider = _make_provider(runtime_config)
+    defaults = runtime_config.agents.defaults
     session_manager = SessionManager(runtime_config.workspace_path)
-    agent_loop = AgentLoop(
+    resolved_preset = runtime_config.resolve_preset()
-        bus=bus,
+    agent_loop = AgentLoop.from_config(
-        provider=provider,
+        runtime_config, bus,
-        workspace=runtime_config.workspace_path,
-        model=runtime_config.agents.defaults.model,
-        max_iterations=runtime_config.agents.defaults.max_tool_iterations,
-        context_window_tokens=runtime_config.agents.defaults.context_window_tokens,
-        context_block_limit=runtime_config.agents.defaults.context_block_limit,
-        max_tool_result_chars=runtime_config.agents.defaults.max_tool_result_chars,
-        provider_retry_mode=runtime_config.agents.defaults.provider_retry_mode,
-        tool_hint_max_length=runtime_config.agents.defaults.tool_hint_max_length,
-        web_config=runtime_config.tools.web,
-        exec_config=runtime_config.tools.exec,
-        restrict_to_workspace=runtime_config.tools.restrict_to_workspace,
         session_manager=session_manager,
-        mcp_servers=runtime_config.tools.mcp_servers,
-        channels_config=runtime_config.channels,
-        timezone=runtime_config.agents.defaults.timezone,
-        unified_session=runtime_config.agents.defaults.unified_session,
-        disabled_skills=runtime_config.agents.defaults.disabled_skills,
-        session_ttl_minutes=runtime_config.agents.defaults.session_ttl_minutes,
-        consolidation_ratio=runtime_config.agents.defaults.consolidation_ratio,
-        max_messages=runtime_config.agents.defaults.max_messages,
-        tools_config=runtime_config.tools,
     )
 
-    model_name = runtime_config.agents.defaults.model
+    model_name = resolved_preset.model
+    preset_name = defaults.model_preset
+    preset_tag = f" (preset: {preset_name})" if preset_name else ""
     console.print(f"{__logo__} Starting OpenAI-compatible API server")
     console.print(f"  [cyan]Endpoint[/cyan] : http://{host}:{port}/v1/chat/completions")
-    console.print(f"  [cyan]Model[/cyan]    : {model_name}")
+    console.print(f"  [cyan]Model[/cyan]    : {model_name}{preset_tag}")
     console.print("  [cyan]Session[/cyan]  : api:default")
     console.print(f"  [cyan]Timeout[/cyan]  : {timeout}s")
     if host in {"0.0.0.0", "::"}:
@@ -638,7 +606,6 @@ def _run_gateway(
     open_browser_url: str | None = None,
 ) -> None:
     """Shared gateway runtime; ``open_browser_url`` opens a tab once channels are up."""
-    from nanobot.agent.loop import AgentLoop
     from nanobot.agent.tools.cron import CronTool
     from nanobot.agent.tools.message import MessageTool
     from nanobot.bus.queue import MessageBus
@@ -659,7 +626,6 @@ def _run_gateway(
     except ValueError as exc:
         console.print(f"[red]Error: {exc}[/red]")
         raise typer.Exit(1) from exc
-    provider = provider_snapshot.provider
     session_manager = SessionManager(config.workspace_path)
 
     # Preserve existing single-workspace installs, but keep custom workspaces clean.
@@ -671,31 +637,10 @@ def _run_gateway(
     cron = CronService(cron_store_path)
 
     # Create agent with cron service
-    agent = AgentLoop(
+    agent = AgentLoop.from_config(
-        bus=bus,
+        config, bus,
-        provider=provider,
-        workspace=config.workspace_path,
-        model=provider_snapshot.model,
-        max_iterations=config.agents.defaults.max_tool_iterations,
-        context_window_tokens=provider_snapshot.context_window_tokens,
-        web_config=config.tools.web,
-        context_block_limit=config.agents.defaults.context_block_limit,
-        max_tool_result_chars=config.agents.defaults.max_tool_result_chars,
-        provider_retry_mode=config.agents.defaults.provider_retry_mode,
-        tool_hint_max_length=config.agents.defaults.tool_hint_max_length,
-        exec_config=config.tools.exec,
         cron_service=cron,
-        restrict_to_workspace=config.tools.restrict_to_workspace,
         session_manager=session_manager,
-        mcp_servers=config.tools.mcp_servers,
-        channels_config=config.channels,
-        timezone=config.agents.defaults.timezone,
-        unified_session=config.agents.defaults.unified_session,
-        disabled_skills=config.agents.defaults.disabled_skills,
-        session_ttl_minutes=config.agents.defaults.session_ttl_minutes,
-        consolidation_ratio=config.agents.defaults.consolidation_ratio,
-        max_messages=config.agents.defaults.max_messages,
-        tools_config=config.tools,
         provider_snapshot_loader=load_provider_snapshot,
         provider_signature=provider_snapshot.signature,
     )
@@ -798,7 +743,7 @@ def _run_gateway(
 
         if job.payload.deliver and job.payload.to and response:
             should_notify = await evaluate_response(
-                response, reminder_note, provider, agent.model,
+                response, reminder_note, agent.provider, agent.model,
             )
             if should_notify:
                 await _deliver_to_channel(
@@ -888,7 +833,7 @@ def _run_gateway(
     hb_cfg = config.gateway.heartbeat
     heartbeat = HeartbeatService(
         workspace=config.workspace_path,
-        provider=provider,
+        provider=agent.provider,
         model=agent.model,
         on_execute=on_heartbeat_execute,
         on_notify=on_heartbeat_notify,
@@ -1041,7 +986,6 @@ def agent(
     """Interact with the agent directly."""
     from loguru import logger
 
-    from nanobot.agent.loop import AgentLoop
     from nanobot.bus.queue import MessageBus
     from nanobot.cron.service import CronService
 
@@ -1049,8 +993,6 @@ def agent(
     sync_workspace_templates(config.workspace_path)
 
     bus = MessageBus()
-    provider = _make_provider(config)
-
     # Preserve existing single-workspace installs, but keep custom workspaces clean.
     if is_default_workspace(config.workspace_path):
         _migrate_cron_store(config)
@@ -1064,30 +1006,10 @@ def agent(
     else:
         logger.disable("nanobot")
 
-    agent_loop = AgentLoop(
+    resolved_preset = config.resolve_preset()
-        bus=bus,
+    agent_loop = AgentLoop.from_config(
-        provider=provider,
+        config, bus,
-        workspace=config.workspace_path,
-        model=config.agents.defaults.model,
-        max_iterations=config.agents.defaults.max_tool_iterations,
-        context_window_tokens=config.agents.defaults.context_window_tokens,
-        web_config=config.tools.web,
-        context_block_limit=config.agents.defaults.context_block_limit,
-        max_tool_result_chars=config.agents.defaults.max_tool_result_chars,
-        provider_retry_mode=config.agents.defaults.provider_retry_mode,
-        tool_hint_max_length=config.agents.defaults.tool_hint_max_length,
-        exec_config=config.tools.exec,
         cron_service=cron,
-        restrict_to_workspace=config.tools.restrict_to_workspace,
-        mcp_servers=config.tools.mcp_servers,
-        channels_config=config.channels,
-        timezone=config.agents.defaults.timezone,
-        unified_session=config.agents.defaults.unified_session,
-        disabled_skills=config.agents.defaults.disabled_skills,
-        session_ttl_minutes=config.agents.defaults.session_ttl_minutes,
-        consolidation_ratio=config.agents.defaults.consolidation_ratio,
-        max_messages=config.agents.defaults.max_messages,
-        tools_config=config.tools,
     )
     restart_notice = consume_restart_notice_from_env()
     if restart_notice and should_show_cli_restart_notice(restart_notice, session_id):
@@ -1131,7 +1053,7 @@ def agent(
         # Interactive mode — route through bus like other channels
         from nanobot.bus.events import InboundMessage
         _init_prompt_session()
-        console.print(f"{__logo__} Interactive mode [bold blue]({config.agents.defaults.model})[/bold blue] — type [bold]exit[/bold] or [bold]Ctrl+C[/bold] to quit\n")
+        console.print(f"{__logo__} Interactive mode [bold blue]({resolved_preset.model})[/bold blue] — type [bold]exit[/bold] or [bold]Ctrl+C[/bold] to quit\n")
 
         if ":" in session_id:
             cli_channel, cli_chat_id = session_id.split(":", 1)
@@ -1489,7 +1411,10 @@ def status():
     if config_path.exists():
         from nanobot.providers.registry import PROVIDERS
 
-        console.print(f"Model: {config.agents.defaults.model}")
+        resolved_preset = config.resolve_preset()
+        preset = config.agents.defaults.model_preset
+        preset_tag = f" (preset: {preset})" if preset else ""
+        console.print(f"Model: {resolved_preset.model}{preset_tag}")
 
         # Check API keys from registry
         for spec in PROVIDERS:

nanobot/cli/onboard.py

@@ -22,7 +22,7 @@ from nanobot.cli.models import (
     get_model_suggestions,
 )
 from nanobot.config.loader import get_config_path, load_config
-from nanobot.config.schema import Config
+from nanobot.config.schema import Config, ModelPresetConfig
 
 console = Console()
 
@@ -49,6 +49,16 @@ _SELECT_FIELD_HINTS: dict[str, tuple[list[str], str]] = {
 
 _BACK_PRESSED = object()  # Sentinel value for back navigation
 
+# Cache of model-preset names populated at runtime so that field handlers can
+# offer existing presets as choices (e.g. AgentDefaults.model_preset).
+#
+# Lifecycle: populated by _sync_preset_cache(config), which must be called
+# after every config mutation that changes model_presets (add, delete, edit).
+# Cleared between tests via _MODEL_PRESET_CACHE.clear(). In long-running
+# processes (gateway) the cache is refreshed each time the preset management
+# screen is entered, so staleness is bounded by user interaction.
+_MODEL_PRESET_CACHE: set[str] = set()
+
 
 def _get_questionary():
     """Return questionary or raise a clear error when wizard deps are unavailable."""
@@ -191,13 +201,13 @@ def _get_field_type_info(field_info) -> FieldTypeInfo:
             origin = get_origin(annotation)
             args = get_args(annotation)
 
-    _SIMPLE_TYPES: dict[type, str] = {bool: "bool", int: "int", float: "float"}
+    _simple_types: dict[type, str] = {bool: "bool", int: "int", float: "float"}
 
     if origin is list or (hasattr(origin, "__name__") and origin.__name__ == "List"):
         return FieldTypeInfo("list", args[0] if args else str)
     if origin is dict or (hasattr(origin, "__name__") and origin.__name__ == "Dict"):
         return FieldTypeInfo("dict", None)
-    for py_type, name in _SIMPLE_TYPES.items():
+    for py_type, name in _simple_types.items():
         if annotation is py_type:
             return FieldTypeInfo(name, None)
     if isinstance(annotation, type) and issubclass(annotation, BaseModel):
@@ -403,7 +413,7 @@ def _input_text(display_name: str, current: Any, field_type: str, field_info=Non
 
     value = _get_questionary().text(f"{display_name}:", default=default).ask()
 
-    if value is None or value == "":
+    if value is None:
         return None
 
     if field_type == "int":
@@ -507,7 +517,7 @@ def _input_model_with_autocomplete(
         qmark=">",
     ).ask()
 
-    return value if value else None
+    return value if value is not None else None
 
 
 def _input_context_window_with_recommendation(
@@ -588,12 +598,112 @@ def _handle_context_window_field(
         setattr(working_model, field_name, new_value)
 
 
+def _handle_model_preset_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'model_preset' field with a list of existing presets."""
+    # model_preset lives on AgentDefaults, but the preset list is on Config.
+    # We can't easily access Config here, so we read from the global config
+    # via a module-level cache set by _configure_model_presets / run_onboard.
+    preset_names = sorted(_MODEL_PRESET_CACHE)
+    choices = ["(clear/unset)"] + preset_names
+    default_choice = str(current_value) if current_value else "(clear/unset)"
+    new_value = _select_with_back(field_display, choices, default=default_choice)
+    if new_value is _BACK_PRESSED:
+        return
+    if new_value == "(clear/unset)":
+        setattr(working_model, field_name, None)
+    elif new_value is not None:
+        setattr(working_model, field_name, new_value)
+
+
+def _handle_provider_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'provider' field with a list of registered providers."""
+    provider_names = sorted(_get_provider_names().keys())
+    choices = ["auto"] + provider_names
+    default_choice = str(current_value) if current_value else "auto"
+    new_value = _select_with_back(field_display, choices, default=default_choice)
+    if new_value is _BACK_PRESSED:
+        return
+    if new_value is not None:
+        setattr(working_model, field_name, new_value)
+
+
+def _handle_fallback_presets_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'fallback_presets' field with preset-aware multi-select."""
+    items: list[str] = list(current_value) if isinstance(current_value, list) else []
+    preset_names = sorted(_MODEL_PRESET_CACHE)
+
+    while True:
+        console.clear()
+        console.print(f"[bold]{field_display}[/bold]")
+        if items:
+            for idx, item in enumerate(items, 1):
+                console.print(f"  {idx}. {item}")
+        else:
+            console.print("  [dim](empty)[/dim]")
+        console.print()
+
+        choices = ["[+] Add preset"]
+        if items:
+            choices.append("[-] Remove last")
+            choices.append("[X] Clear all")
+        choices.append("[Done]")
+        choices.append("<- Back")
+
+        answer = _get_questionary().select(
+            "Manage fallback chain:",
+            choices=choices,
+            qmark=">",
+        ).ask()
+
+        if answer is None or answer == "<- Back":
+            return
+        if answer == "[Done]":
+            setattr(working_model, field_name, items)
+            return
+        if answer == "[+] Add preset":
+            if not preset_names:
+                console.print("[yellow]! No presets defined yet.[/yellow]")
+                _get_questionary().press_any_key_to_continue().ask()
+                continue
+            add_choices = [p for p in preset_names if p not in items]
+            if not add_choices:
+                console.print("[yellow]! All presets already added.[/yellow]")
+                _get_questionary().press_any_key_to_continue().ask()
+                continue
+            picked = _select_with_back("Select preset:", add_choices)
+            if picked is _BACK_PRESSED or picked is None:
+                continue
+            items.append(picked)
+        elif answer == "[-] Remove last" and items:
+            items.pop()
+        elif answer == "[X] Clear all" and items:
+            items.clear()
+
+
 _FIELD_HANDLERS: dict[str, Any] = {
     "model": _handle_model_field,
     "context_window_tokens": _handle_context_window_field,
+    "model_preset": _handle_model_preset_field,
+    "provider": _handle_provider_field,
+    "fallback_presets": _handle_fallback_presets_field,
 }
 
 
+def _is_str_or_none(annotation: Any) -> bool:
+    """Check whether a field annotation is ``str | None`` (or ``Optional[str]``)."""
+    origin = get_origin(annotation)
+    if origin is None:
+        return False
+    args = get_args(annotation)
+    return str in args and type(None) in args
+
+
 def _configure_pydantic_model(
     model: BaseModel,
     display_name: str,
@@ -626,11 +736,20 @@ def _configure_pydantic_model(
             items.append(f"{display}: {formatted}")
         return items + ["[Done]"]
 
+    last_field_name: str | None = None
     while True:
         console.clear()
         _show_config_panel(display_name, working_model, fields)
         choices = get_choices()
-        answer = _select_with_back("Select field to configure:", choices)
+        default_choice = None
+        if last_field_name:
+            for idx, (fname, _) in enumerate(fields):
+                if fname == last_field_name:
+                    default_choice = choices[idx]
+                    break
+        answer = _select_with_back(
+            "Select field to configure:", choices, default=default_choice
+        )
 
         if answer is _BACK_PRESSED or answer is None:
             return None
@@ -641,6 +760,8 @@ def _configure_pydantic_model(
         if field_idx < 0 or field_idx >= len(fields):
             return None
 
+        last_field_name = fields[field_idx][0]
+
         field_name, field_info = fields[field_idx]
         current_value = getattr(working_model, field_name, None)
         ftype = _get_field_type_info(field_info)
@@ -697,6 +818,10 @@ def _configure_pydantic_model(
         else:
             new_value = _input_with_existing(field_display, current_value, ftype.type_name, field_info=field_info)
         if new_value is not None:
+            # Normalize empty string to None for optional string fields so that
+            # clearing an api_key / api_base actually removes the value.
+            if new_value == "" and _is_str_or_none(field_info.annotation):
+                new_value = None
             setattr(working_model, field_name, new_value)
 
 
@@ -733,6 +858,113 @@ def _try_auto_fill_context_window(model: BaseModel, new_model_name: str) -> None
         console.print("[dim](i) Could not auto-fill context window (model not in database)[/dim]")
 
 
+# --- Model Preset Configuration ---
+
+
+def _sync_preset_cache(config: Config) -> None:
+    """Synchronise the module-level preset name cache from config."""
+    _MODEL_PRESET_CACHE.clear()
+    _MODEL_PRESET_CACHE.update(config.model_presets.keys())
+
+
+def _configure_model_presets(config: Config) -> None:
+    """Configure model presets (CRUD)."""
+    _sync_preset_cache(config)
+
+    def get_preset_choices() -> list[str]:
+        choices: list[str] = []
+        for name, preset in config.model_presets.items():
+            choices.append(f"{name} ({preset.model})")
+        choices.append("[+] Add new preset")
+        choices.append("<- Back")
+        return choices
+
+    last_preset_name: str | None = None
+    while True:
+        try:
+            console.clear()
+            _show_section_header(
+                "Model Presets",
+                "Create, edit or delete named model presets for quick switching",
+            )
+            choices = get_preset_choices()
+            default_choice = None
+            if last_preset_name:
+                for c in choices:
+                    if c.startswith(last_preset_name + " ("):
+                        default_choice = c
+                        break
+            answer = _select_with_back(
+                "Select preset:", choices, default=default_choice
+            )
+
+            if answer is _BACK_PRESSED or answer is None or answer == "<- Back":
+                break
+
+            assert isinstance(answer, str)
+
+            if answer == "[+] Add new preset":
+                name_input = _get_questionary().text(
+                    "Preset name:",
+                    validate=lambda t: True if t and t.strip() else "Name cannot be empty",
+                ).ask()
+                if not name_input:
+                    continue
+                name = name_input.strip()
+                if name in config.model_presets:
+                    console.print(f"[yellow]! Preset '{name}' already exists[/yellow]")
+                    _pause()
+                    continue
+                new_preset = ModelPresetConfig(model="")
+                updated = _configure_pydantic_model(new_preset, f"New Preset: {name}")
+                if updated is not None:
+                    config.model_presets[name] = updated
+                    _sync_preset_cache(config)
+                    last_preset_name = name
+                continue
+
+            # Editing / deleting an existing preset
+            # Extract preset name from "name (model)" format
+            preset_name = answer.split(" (", 1)[0]
+            preset = config.model_presets.get(preset_name)
+            if preset is None:
+                continue
+
+            last_preset_name = preset_name
+
+            choices = ["Edit", "Cancel"]
+            if preset_name != "default":
+                choices.insert(1, "Delete")
+            action = _select_with_back(
+                f"Preset: {preset_name}",
+                choices,
+                default="Edit",
+            )
+            if action is _BACK_PRESSED or action == "Cancel" or action is None:
+                continue
+
+            if action == "Delete":
+                confirm = _get_questionary().confirm(
+                    f"Delete preset '{preset_name}'?",
+                    default=False,
+                ).ask()
+                if confirm:
+                    del config.model_presets[preset_name]
+                    _sync_preset_cache(config)
+                    last_preset_name = None
+                continue
+
+            if action == "Edit":
+                updated = _configure_pydantic_model(preset, f"Edit Preset: {preset_name}")
+                if updated is not None:
+                    config.model_presets[preset_name] = updated
+                    _sync_preset_cache(config)
+
+        except KeyboardInterrupt:
+            console.print("\n[dim]Returning to main menu...[/dim]")
+            break
+
+
 # --- Provider Configuration ---
 
 
@@ -795,12 +1027,23 @@ def _configure_providers(config: Config) -> None:
                 choices.append(display)
         return choices + ["<- Back"]
 
+    last_provider_key: str | None = None
     while True:
         try:
             console.clear()
             _show_section_header("LLM Providers", "Select a provider to configure API key and endpoint")
             choices = get_provider_choices()
-            answer = _select_with_back("Select provider:", choices)
+            default_choice = None
+            if last_provider_key:
+                display = _get_provider_names().get(last_provider_key)
+                if display:
+                    for c in choices:
+                        if c.replace(" *", "") == display:
+                            default_choice = c
+                            break
+            answer = _select_with_back(
+                "Select provider:", choices, default=default_choice
+            )
 
             if answer is _BACK_PRESSED or answer is None or answer == "<- Back":
                 break
@@ -812,6 +1055,7 @@ def _configure_providers(config: Config) -> None:
             # Find the actual provider key from display names
             for name, display in _get_provider_names().items():
                 if display == provider_name:
+                    last_provider_key = name
                     _configure_provider(config, name)
                     break
 
@@ -885,17 +1129,21 @@ def _configure_channels(config: Config) -> None:
     channel_names = list(_get_channel_names().keys())
     choices = channel_names + ["<- Back"]
 
+    last_choice: str | None = None
     while True:
         try:
             console.clear()
             _show_section_header("Chat Channels", "Select a channel to configure connection settings")
-            answer = _select_with_back("Select channel:", choices)
+            answer = _select_with_back(
+                "Select channel:", choices, default=last_choice
+            )
 
             if answer is _BACK_PRESSED or answer is None or answer == "<- Back":
                 break
 
             # Type guard: answer is now guaranteed to be a string
             assert isinstance(answer, str)
+            last_choice = answer
             _configure_channel(config, answer)
         except KeyboardInterrupt:
             console.print("\n[dim]Returning to main menu...[/dim]")
@@ -1003,6 +1251,12 @@ def _show_summary(config: Config) -> None:
         channel_rows.append((display, status))
     _print_summary_panel(channel_rows, "Chat Channels")
 
+    # Model Presets
+    preset_rows = []
+    for name, preset in config.model_presets.items():
+        preset_rows.append((name, f"{preset.model} (ctx={preset.context_window_tokens})"))
+    _print_summary_panel(preset_rows, "Model Presets")
+
     # Settings sections
     for title, model in [
         ("Agent Settings", config.agents.defaults),
@@ -1072,7 +1326,9 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
 
     original_config = base_config.model_copy(deep=True)
     config = base_config.model_copy(deep=True)
+    _sync_preset_cache(config)
 
+    last_main_choice: str | None = None
     while True:
         console.clear()
         _show_main_menu_header()
@@ -1082,6 +1338,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
                 "What would you like to configure?",
                 choices=[
                     "[P] LLM Provider",
+                    "[M] Model Presets",
                     "[C] Chat Channel",
                     "[H] Channel Common",
                     "[A] Agent Settings",
@@ -1092,6 +1349,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
                     "[S] Save and Exit",
                     "[X] Exit Without Saving",
                 ],
+                default=last_main_choice,
                 qmark=">",
             ).ask()
         except KeyboardInterrupt:
@@ -1105,8 +1363,9 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
                 return OnboardResult(config=original_config, should_save=False)
             continue
 
-        _MENU_DISPATCH = {
+        _menu_dispatch = {
             "[P] LLM Provider": lambda: _configure_providers(config),
+            "[M] Model Presets": lambda: _configure_model_presets(config),
             "[C] Chat Channel": lambda: _configure_channels(config),
             "[H] Channel Common": lambda: _configure_general_settings(config, "Channel Common"),
             "[A] Agent Settings": lambda: _configure_general_settings(config, "Agent Settings"),
@@ -1121,6 +1380,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
         if answer == "[X] Exit Without Saving":
             return OnboardResult(config=original_config, should_save=False)
 
-        action_fn = _MENU_DISPATCH.get(answer)
+        action_fn = _menu_dispatch.get(answer)
         if action_fn:
+            last_main_choice = answer
             action_fn()

nanobot/config/schema.py

@@ -3,7 +3,7 @@
 from pathlib import Path
 from typing import Any, Literal
 
-from pydantic import AliasChoices, BaseModel, ConfigDict, Field
+from pydantic import AliasChoices, BaseModel, ConfigDict, Field, model_validator
 from pydantic.alias_generators import to_camel
 from pydantic_settings import BaseSettings
 
@@ -65,18 +65,34 @@ class DreamConfig(Base):
         return f"every {hours}h"
 
 
+class ModelPresetConfig(Base):
+    """A named set of model + generation parameters for quick switching."""
+
+    model: str
+    provider: str = "auto"
+    max_tokens: int = 8192
+    context_window_tokens: int = 65_536
+    temperature: float = 0.1
+    reasoning_effort: str | None = None
+
+
 class AgentDefaults(Base):
     """Default agent configuration."""
 
     workspace: str = "~/.nanobot/workspace"
+    model_preset: str | None = None  # Active preset name — takes precedence over fields below
+    # Fallback fields (used when model_preset is not set):
     model: str = "anthropic/claude-opus-4-5"
     provider: str = (
         "auto"  # Provider name (e.g. "anthropic", "openrouter") or "auto" for auto-detection
     )
     max_tokens: int = 8192
     context_window_tokens: int = 65_536
-    context_block_limit: int | None = None
     temperature: float = 0.1
+    reasoning_effort: str | None = None  # low / medium / high / adaptive - enables LLM thinking mode
+    # End fallback fields
+
+    context_block_limit: int | None = None
     max_tool_iterations: int = 200
     max_concurrent_subagents: int = Field(default=1, ge=1)
     max_tool_result_chars: int = 16_000
@@ -88,7 +104,9 @@ class AgentDefaults(Base):
         validation_alias=AliasChoices("toolHintMaxLength"),
         serialization_alias="toolHintMaxLength",
     )  # Max characters for tool hint display (e.g. "$ cd …/project && npm test")
-    reasoning_effort: str | None = None  # low / medium / high / adaptive - enables LLM thinking mode
+    fallback_presets: list[str] = Field(
+        default_factory=list
+    )  # Ordered fallback chain. Each item must be a preset name defined in model_presets.
     timezone: str = "UTC"  # IANA timezone, e.g. "Asia/Shanghai", "America/New_York"
     unified_session: bool = False  # Share one session across all channels (single-user multi-device)
     disabled_skills: list[str] = Field(default_factory=list)  # Skill names to exclude from loading (e.g. ["summarize", "skill-creator"])
@@ -273,6 +291,54 @@ class Config(BaseSettings):
     api: ApiConfig = Field(default_factory=ApiConfig)
     gateway: GatewayConfig = Field(default_factory=GatewayConfig)
     tools: ToolsConfig = Field(default_factory=ToolsConfig)
+    model_presets: dict[str, ModelPresetConfig] = Field(default_factory=dict)
+
+    @model_validator(mode="after")
+    def _sync_and_validate_preset(self) -> "Config":
+        """Expose agents.defaults model fields as the implicit 'default' preset
+        and validate the active preset reference.
+
+        This guarantees that ``model_presets`` is never empty and that legacy
+        configs (which only set ``agents.defaults.model`` etc.) continue to work
+        without explicitly declaring a preset.
+        """
+        self._refresh_default_preset()
+        defaults = self.agents.defaults
+        if defaults.model_preset is None:
+            defaults.model_preset = "default"
+        if defaults.model_preset not in self.model_presets:
+            raise ValueError(f"model_preset {defaults.model_preset!r} not found in model_presets")
+        for fb in defaults.fallback_presets:
+            if fb not in self.model_presets:
+                raise ValueError(f"fallback_presets entry {fb!r} not found in model_presets")
+        return self
+
+    def _refresh_default_preset(self) -> None:
+        """Rebuild the implicit 'default' preset from current agents.defaults.
+
+        Called inside ``_sync_and_validate_preset`` (model validator) and
+        ``resolve_preset()`` so that runtime mutations (e.g. tests directly
+        setting ``defaults.model``) are reflected.
+        """
+        d = self.agents.defaults
+        self.model_presets["default"] = ModelPresetConfig(
+            model=d.model,
+            provider=d.provider,
+            max_tokens=d.max_tokens,
+            context_window_tokens=d.context_window_tokens,
+            temperature=d.temperature,
+            reasoning_effort=d.reasoning_effort,
+        )
+
+    def resolve_preset(self) -> ModelPresetConfig:
+        """Return the active preset.
+
+        The implicit ``"default"`` preset is rebuilt from current defaults every
+        time so that runtime mutations (e.g. tests setting ``defaults.model``)
+        are always reflected.
+        """
+        self._refresh_default_preset()
+        return self.model_presets[self.agents.defaults.model_preset]
 
     @property
     def workspace_path(self) -> Path:
@@ -285,15 +351,16 @@ class Config(BaseSettings):
         """Match provider config and its registry name. Returns (config, spec_name)."""
         from nanobot.providers.registry import PROVIDERS, find_by_name
 
-        forced = self.agents.defaults.provider
+        resolved = self.resolve_preset()
+        forced = resolved.provider
         if forced != "auto":
             spec = find_by_name(forced)
             if spec:
-                p = getattr(self.providers, spec.name, None)
+                provider_cfg = getattr(self.providers, spec.name, None)
-                return (p, spec.name) if p else (None, None)
+                return (provider_cfg, spec.name) if provider_cfg else (None, None)
             return None, None
 
-        model_lower = (model or self.agents.defaults.model).lower()
+        model_lower = (model or resolved.model).lower()
         model_normalized = model_lower.replace("-", "_")
         model_prefix = model_lower.split("/", 1)[0] if "/" in model_lower else ""
         normalized_prefix = model_prefix.replace("-", "_")

nanobot/nanobot.py

@@ -8,7 +8,6 @@ from typing import Any
 
 from nanobot.agent.hook import AgentHook, SDKCaptureHook
 from nanobot.agent.loop import AgentLoop
-from nanobot.bus.queue import MessageBus
 
 
 @dataclass(slots=True)
@@ -62,32 +61,7 @@ class Nanobot:
                 Path(workspace).expanduser().resolve()
             )
 
-        provider = _make_provider(config)
+        loop = AgentLoop.from_config(config)
-        bus = MessageBus()
-        defaults = config.agents.defaults
-
-        loop = AgentLoop(
-            bus=bus,
-            provider=provider,
-            workspace=config.workspace_path,
-            model=defaults.model,
-            max_iterations=defaults.max_tool_iterations,
-            context_window_tokens=defaults.context_window_tokens,
-            context_block_limit=defaults.context_block_limit,
-            max_tool_result_chars=defaults.max_tool_result_chars,
-            provider_retry_mode=defaults.provider_retry_mode,
-            tool_hint_max_length=defaults.tool_hint_max_length,
-            web_config=config.tools.web,
-            exec_config=config.tools.exec,
-            restrict_to_workspace=config.tools.restrict_to_workspace,
-            mcp_servers=config.tools.mcp_servers,
-            timezone=defaults.timezone,
-            unified_session=defaults.unified_session,
-            disabled_skills=defaults.disabled_skills,
-            session_ttl_minutes=defaults.session_ttl_minutes,
-            consolidation_ratio=defaults.consolidation_ratio,
-            tools_config=config.tools,
-        )
         return cls(loop)
 
     async def run(
@@ -124,8 +98,3 @@ class Nanobot:
         )
 
 
-def _make_provider(config: Any) -> Any:
-    """Create the LLM provider from config (extracted from CLI)."""
-    from nanobot.providers.factory import make_provider
-
-    return make_provider(config)

nanobot/providers/base.py

@@ -137,7 +137,9 @@ class LLMProvider(ABC):
         "insufficient_quota",
         "insufficient quota",
         "quota exceeded",
+        "quota_exceeded",
         "quota exhausted",
+        "quota_exhausted",
         "billing hard limit",
         "billing_hard_limit_reached",
         "billing not active",

nanobot/providers/factory.py

@@ -4,11 +4,16 @@ from __future__ import annotations
 
 from dataclasses import dataclass
 from pathlib import Path
+from typing import TYPE_CHECKING
 
 from nanobot.config.schema import Config
 from nanobot.providers.base import GenerationSettings, LLMProvider
 from nanobot.providers.registry import find_by_name
 
+if TYPE_CHECKING:
+    from nanobot.config.schema import ModelPresetConfig, ProviderConfig
+    from nanobot.providers.registry import ProviderSpec
+
 
 @dataclass(frozen=True)
 class ProviderSnapshot:
@@ -18,22 +23,62 @@ class ProviderSnapshot:
     signature: tuple[object, ...]
 
 
-def make_provider(config: Config) -> LLMProvider:
+@dataclass(frozen=True)
-    """Create the LLM provider implied by config."""
+class _ProviderInfo:
-    model = config.agents.defaults.model
+    """Resolved metadata needed to build and validate an LLM provider."""
-    provider_name = config.get_provider_name(model)
+
-    p = config.get_provider(model)
+    name: str | None
-    spec = find_by_name(provider_name) if provider_name else None
+    cfg: ProviderConfig | None
+    spec: ProviderSpec | None
+    api_base: str | None
+    backend: str
+
+
+def _resolve_provider_info(
+    config: Config,
+    model: str,
+    preset: ModelPresetConfig,
+) -> _ProviderInfo:
+    """Derive provider name, config, spec and api_base from preset or auto-detection."""
+    if preset.provider != "auto":
+        name = preset.provider
+        cfg = getattr(config.providers, name, None)
+        spec = find_by_name(name)
+        api_base = (
+            cfg.api_base
+            if cfg and cfg.api_base
+            else (spec.default_api_base if spec and spec.default_api_base else None)
+        )
+    else:
+        name = config.get_provider_name(model)
+        cfg = config.get_provider(model)
+        spec = find_by_name(name) if name else None
+        api_base = config.get_api_base(model)
+
     backend = spec.backend if spec else "openai_compat"
+    return _ProviderInfo(name=name, cfg=cfg, spec=spec, api_base=api_base, backend=backend)
+
+
+def _validate_provider(info: _ProviderInfo, model: str) -> None:
+    """Ensure credentials / endpoints are present before instantiation."""
+    cfg = info.cfg
+    backend = info.backend
+    name = info.name
 
     if backend == "azure_openai":
-        if not p or not p.api_key or not p.api_base:
+        if not cfg or not cfg.api_key or not cfg.api_base:
             raise ValueError("Azure OpenAI requires api_key and api_base in config.")
     elif backend == "openai_compat" and not model.startswith("bedrock/"):
-        needs_key = not (p and p.api_key)
+        needs_key = not (cfg and cfg.api_key)
-        exempt = spec and (spec.is_oauth or spec.is_local or spec.is_direct)
+        exempt = info.spec and (info.spec.is_oauth or info.spec.is_local or info.spec.is_direct)
         if needs_key and not exempt:
-            raise ValueError(f"No API key configured for provider '{provider_name}'.")
+            raise ValueError(f"No API key configured for provider '{name}'.")
+
+
+def _create_provider(model: str, info: _ProviderInfo) -> LLMProvider:
+    """Instantiate the concrete provider class for *backend*."""
+    cfg = info.cfg
+    backend = info.backend
 
     if backend == "openai_codex":
         from nanobot.providers.openai_codex_provider import OpenAICodexProvider
@@ -43,8 +88,8 @@ def make_provider(config: Config) -> LLMProvider:
         from nanobot.providers.azure_openai_provider import AzureOpenAIProvider
 
         provider = AzureOpenAIProvider(
-            api_key=p.api_key,
+            api_key=cfg.api_key if cfg else None,
-            api_base=p.api_base,
+            api_base=info.api_base,
             default_model=model,
         )
     elif backend == "github_copilot":
@@ -55,70 +100,103 @@ def make_provider(config: Config) -> LLMProvider:
         from nanobot.providers.anthropic_provider import AnthropicProvider
 
         provider = AnthropicProvider(
-            api_key=p.api_key if p else None,
+            api_key=cfg.api_key if cfg else None,
-            api_base=config.get_api_base(model),
+            api_base=info.api_base,
             default_model=model,
-            extra_headers=p.extra_headers if p else None,
+            extra_headers=cfg.extra_headers if cfg else None,
         )
     elif backend == "bedrock":
         from nanobot.providers.bedrock_provider import BedrockProvider
 
         provider = BedrockProvider(
-            api_key=p.api_key if p else None,
+            api_key=cfg.api_key if cfg else None,
-            api_base=p.api_base if p else None,
+            api_base=info.api_base if cfg else None,
             default_model=model,
-            region=getattr(p, "region", None) if p else None,
+            region=getattr(cfg, "region", None) if cfg else None,
-            profile=getattr(p, "profile", None) if p else None,
+            profile=getattr(cfg, "profile", None) if cfg else None,
-            extra_body=p.extra_body if p else None,
+            extra_body=cfg.extra_body if cfg else None,
         )
     else:
         from nanobot.providers.openai_compat_provider import OpenAICompatProvider
 
         provider = OpenAICompatProvider(
-            api_key=p.api_key if p else None,
+            api_key=cfg.api_key if cfg else None,
-            api_base=config.get_api_base(model),
+            api_base=info.api_base,
             default_model=model,
-            extra_headers=p.extra_headers if p else None,
+            extra_headers=cfg.extra_headers if cfg else None,
-            spec=spec,
+            spec=info.spec,
-            extra_body=p.extra_body if p else None,
+            extra_body=cfg.extra_body if cfg else None,
         )
-
-    defaults = config.agents.defaults
-    provider.generation = GenerationSettings(
-        temperature=defaults.temperature,
-        max_tokens=defaults.max_tokens,
-        reasoning_effort=defaults.reasoning_effort,
-    )
     return provider
 
 
+def _apply_generation(provider: LLMProvider, preset: ModelPresetConfig) -> None:
+    provider.generation = GenerationSettings(
+        temperature=preset.temperature,
+        max_tokens=preset.max_tokens,
+        reasoning_effort=preset.reasoning_effort,
+    )
+
+
+def build_provider_for_preset(config: Config, preset: ModelPresetConfig) -> LLMProvider:
+    """Create an LLM provider from a full *preset* (model + provider + generation)."""
+    info = _resolve_provider_info(config, preset.model, preset)
+    _validate_provider(info, preset.model)
+    provider = _create_provider(preset.model, info)
+    _apply_generation(provider, preset)
+    return provider
+
+
+def make_provider(config: Config) -> LLMProvider:
+    """Create the LLM provider implied by config (legacy entrypoint)."""
+    resolved = config.resolve_preset()
+    return build_provider_for_preset(config, resolved)
+
+
+def make_provider_factory(config: Config):
+    """Build a cached factory that creates providers for preset names.
+
+    The factory looks up *preset_name* in ``config.model_presets`` and builds
+    the provider from the preset's full configuration.
+    """
+    cache: dict[str, LLMProvider] = {}
+    presets = config.model_presets
+
+    def factory(preset_name: str) -> LLMProvider:
+        preset = presets.get(preset_name)
+        if preset is None:
+            raise ValueError(f"Preset {preset_name!r} not found in model_presets")
+        if preset_name not in cache:
+            cache[preset_name] = build_provider_for_preset(config, preset)
+        return cache[preset_name]
+
+    return factory
+
+
 def provider_signature(config: Config) -> tuple[object, ...]:
     """Return the config fields that affect the primary LLM provider."""
-    model = config.agents.defaults.model
+    resolved = config.resolve_preset()
     defaults = config.agents.defaults
-    p = config.get_provider(model)
     return (
-        model,
+        resolved.model,
-        defaults.provider,
+        resolved.provider,
-        config.get_provider_name(model),
+        config.get_provider_name(resolved.model),
-        config.get_api_key(model),
+        config.get_api_key(resolved.model),
-        config.get_api_base(model),
+        config.get_api_base(resolved.model),
-        p.extra_headers if p else None,
+        resolved.max_tokens,
-        p.extra_body if p else None,
+        resolved.temperature,
-        getattr(p, "region", None) if p else None,
+        resolved.reasoning_effort,
-        getattr(p, "profile", None) if p else None,
+        resolved.context_window_tokens,
-        defaults.max_tokens,
+        tuple(defaults.fallback_presets),
-        defaults.temperature,
-        defaults.reasoning_effort,
-        defaults.context_window_tokens,
     )
 
 
 def build_provider_snapshot(config: Config) -> ProviderSnapshot:
+    resolved = config.resolve_preset()
     return ProviderSnapshot(
         provider=make_provider(config),
-        model=config.agents.defaults.model,
+        model=resolved.model,
-        context_window_tokens=config.agents.defaults.context_window_tokens,
+        context_window_tokens=resolved.context_window_tokens,
         signature=provider_signature(config),
     )
 

nanobot/providers/failover.py

@@ -0,0 +1,183 @@
+"""Provider-like failover router used after provider-local retry is exhausted."""
+
+from __future__ import annotations
+
+import asyncio
+from collections.abc import Awaitable, Callable
+from typing import Any
+
+from loguru import logger
+
+from nanobot.providers.base import GenerationSettings, LLMProvider, LLMResponse
+
+
+class ModelRouter(LLMProvider):
+    """Try fallback model candidates for eligible transient final errors."""
+
+    def __init__(
+        self,
+        *,
+        primary_provider: LLMProvider,
+        primary_model: str,
+        fallback_presets: list[str],
+        provider_factory: Callable[[str], LLMProvider] | None = None,
+        per_candidate_timeout_s: float | None = None,
+    ) -> None:
+        super().__init__(
+            api_key=getattr(primary_provider, "api_key", None),
+            api_base=getattr(primary_provider, "api_base", None),
+        )
+        self.primary_provider = primary_provider
+        self.primary_model = primary_model
+        self.fallback_presets = list(fallback_presets)
+        self._provider_factory = provider_factory
+        self._provider_cache: dict[str, LLMProvider] = {}
+        self.per_candidate_timeout_s = per_candidate_timeout_s
+        self.generation = getattr(primary_provider, "generation", GenerationSettings())
+
+    def get_default_model(self) -> str:
+        return self.primary_model
+
+    async def chat(self, **kwargs: Any) -> LLMResponse:
+        async def call(provider: LLMProvider, candidate_model: str, _unused_delta: Any) -> LLMResponse:
+            return await provider.chat(**{**kwargs, "model": candidate_model})
+        return await self._route(call)
+
+    async def chat_stream(self, **kwargs: Any) -> LLMResponse:
+        async def call(provider: LLMProvider, candidate_model: str, content_delta: Any) -> LLMResponse:
+            return await provider.chat_stream(
+                **{**kwargs, "model": candidate_model, "on_content_delta": content_delta}
+            )
+        return await self._route(call, on_content_delta=kwargs.get("on_content_delta"))
+
+    @property
+    def supports_progress_deltas(self) -> bool:  # type: ignore[override]
+        return getattr(self.primary_provider, "supports_progress_deltas", False)
+
+    @classmethod
+    def _should_failover(cls, response: LLMResponse) -> bool:
+        if response.finish_reason != "error":
+            return False
+        if response.error_should_retry is False:
+            return False
+        if response.error_kind == "configuration":
+            return False
+        return True
+
+    def _resolve(self, model: str) -> tuple[LLMProvider, str]:
+        """Return (provider, actual_model_name) for a preset name.
+
+        Caches results so factory is only invoked once per unique name.
+        """
+        if model in self._provider_cache:
+            cached_provider = self._provider_cache[model]
+            return cached_provider, cached_provider.get_default_model()
+        if self._provider_factory is None:
+            raise ValueError(
+                f"Cannot resolve fallback model {model!r}: no provider_factory configured"
+            )
+        provider = self._provider_factory(model)
+        self._provider_cache[model] = provider
+        return provider, provider.get_default_model()
+
+    async def _with_timeout(self, coro: Awaitable[LLMResponse]) -> LLMResponse:
+        timeout_s = self.per_candidate_timeout_s
+        if timeout_s is None:
+            return await coro
+        try:
+            return await asyncio.wait_for(coro, timeout=timeout_s)
+        except asyncio.TimeoutError:
+            return LLMResponse(
+                content=f"Error calling LLM: timed out after {timeout_s:g}s",
+                finish_reason="error",
+                error_kind="timeout",
+            )
+
+    @staticmethod
+    def _resolver_error(label: str, exc: Exception) -> LLMResponse:
+        logger.warning("Failed to resolve fallback model {}: {}", label, exc)
+        return LLMResponse(
+            content=f"Error configuring fallback model {label}: {exc}",
+            finish_reason="error",
+            error_kind="configuration",
+            error_should_retry=False,
+        )
+
+    async def _route(
+        self,
+        call: Callable[[LLMProvider, str, Callable[[str], Awaitable[None]] | None], Awaitable[LLMResponse]],
+        *,
+        on_content_delta: Callable[[str], Awaitable[None]] | None = None,
+    ) -> LLMResponse:
+        """Try primary then each fallback candidate, lazily resolving providers."""
+
+        async def _try_one(label: str, provider: LLMProvider, model: str) -> LLMResponse:
+            try:
+                return await self._with_timeout(call(provider, model, on_content_delta))
+            except asyncio.CancelledError:
+                raise
+            except Exception as exc:
+                return self._resolver_error(label, exc)
+
+        # Primary
+        response = await _try_one("primary", self.primary_provider, self.primary_model)
+        if response.finish_reason != "error":
+            return response
+        if not self._should_failover(response):
+            return response
+
+        # Fallbacks
+        for name in self.fallback_presets:
+            try:
+                provider, model = self._resolve(name)
+            except Exception as exc:
+                logger.warning("Failed to resolve fallback model {}: {}", name, exc)
+                return self._resolver_error(name, exc)
+
+            response = await _try_one(name, provider, model)
+            if response.finish_reason != "error":
+                logger.info("LLM failover selected model={}", name)
+                return response
+            if not self._should_failover(response):
+                return response
+
+        logger.warning("LLM failover exhausted after all candidates")
+        return response
+
+    async def chat_with_retry(self, **kwargs: Any) -> LLMResponse:
+        async def call(
+            provider: LLMProvider, candidate_model: str, _unused_delta: Any
+        ) -> LLMResponse:
+            return await provider.chat_with_retry(
+                **{**kwargs, "model": candidate_model}
+            )
+        return await self._route(call)
+
+    async def chat_stream_with_retry(self, **kwargs: Any) -> LLMResponse:
+        on_content_delta = kwargs.pop("on_content_delta", None)
+
+        async def call(
+            provider: LLMProvider,
+            candidate_model: str,
+            content_delta: Callable[[str], Awaitable[None]] | None,
+        ) -> LLMResponse:
+            buffered: list[str] = []
+
+            async def buffer_delta(delta: str) -> None:
+                buffered.append(delta)
+
+            kwargs["on_content_delta"] = buffer_delta if content_delta else None
+            response = await provider.chat_stream_with_retry(
+                **{**kwargs, "model": candidate_model}
+            )
+            if response.finish_reason != "error" and content_delta:
+                try:
+                    for delta in buffered:
+                        await content_delta(delta)
+                except asyncio.CancelledError:
+                    raise
+                except Exception:
+                    logger.exception("Failover delta callback failed for model={}", candidate_model)
+            return response
+
+        return await self._route(call, on_content_delta=on_content_delta)

nanobot/skills/create-instance/SKILL.md

@@ -0,0 +1,64 @@
+---
+name: create-instance
+description: "Create a new nanobot instance with separate config and workspace. Use when the user wants to set up a new bot, create a new instance for a different channel, persona, or purpose. Triggers on: create instance, new bot, set up bot, add bot, create telegram/discord/feishu/slack/wechat/wecom/dingtalk/qq/email/matrix/msteams/whatsapp bot, multi-instance setup."
+---
+
+# Create Instance
+
+Set up a new nanobot instance with its own config and workspace.
+
+## Steps
+
+1. **Collect information** (ask one at a time if not already provided):
+   - **Instance name** (required): short identifier, e.g. `telegram-bot`, `work-slack`
+   - **Channel type** (required): see table below
+   - **Model** (optional): LLM model, defaults to current instance
+
+2. **Do NOT collect secrets** in the chat (API keys, bot tokens). API keys are automatically inherited from the current instance via `--inherit-config`. Channel-specific tokens must be filled in manually after creation.
+
+3. **Run the creation script**:
+
+```bash
+python <skill-dir>/scripts/create_instance.py --name <name> --channel <channel> --inherit-config <current-config>
+```
+
+- `<skill-dir>` — the directory containing this SKILL.md
+- `<current-config>` — current instance's config path, typically `~/.nanobot/config.json`
+- Optional: `--model <model>`, `--config-dir <path>`
+
+**Exec tool constraints:**
+- Use forward-slash paths (works on all platforms)
+- Do not wrap paths in quotes
+- Do not use `cd`; pass the full script path directly
+
+4. **Report results** to the user:
+   - Config and workspace paths (script outputs them)
+   - Required fields to fill in (script lists them)
+   - Start command: `nanobot gateway --config <config-path>`
+
+## Available Channels
+
+| Channel | Key | Required Fields |
+|---------|-----|-----------------|
+| Telegram | `telegram` | token |
+| Discord | `discord` | token |
+| Feishu / Lark | `feishu` | app_id, app_secret |
+| DingTalk | `dingtalk` | client_id, client_secret |
+| Slack | `slack` | bot_token, app_token |
+| WeCom | `wecom` | bot_id, secret |
+| WeChat OA | `weixin` | token |
+| WhatsApp | `whatsapp` | bridge_token |
+| QQ | `qq` | app_id, secret |
+| Email | `email` | imap_host, imap_username, imap_password, smtp_host, smtp_username, smtp_password, from_address |
+| Matrix | `matrix` | user_id, password or access_token |
+| MS Teams | `msteams` | app_id, app_password, tenant_id |
+| MoChat | `mochat` | claw_token |
+| WebSocket | `websocket` | token |
+
+For detailed channel configuration including optional fields, see `references/channels.md`.
+
+## Troubleshooting
+
+- **"Unknown channel"**: Channel name must match the Key column exactly. Run the script without arguments to see usage.
+- **"Config already exists"**: Use a different `--name` or `--config-dir` to create in a new location.
+- **Port conflicts**: The script auto-assigns free ports for gateway and API if defaults are in use.

nanobot/skills/create-instance/references/channels.md

@@ -0,0 +1,194 @@
+# Channel Configuration Reference
+
+Detailed configuration for each supported channel.
+
+## Field Types
+
+- **Required**: defaults to empty string `""`, must be filled in before the instance can start
+- **Optional**: has a sensible default, can be customized
+
+---
+
+## telegram
+
+**Required:**
+- `token` — Bot token from @BotFather
+
+**Notable optional:**
+- `proxy` — HTTP proxy URL
+- `group_policy` — `"open"` (all messages) or `"mention"` (default, only when @mentioned)
+- `streaming` — Enable streaming responses (default: true)
+- `reply_to_message` — Reply to the triggering message (default: false)
+- `react_emoji` — Emoji for "thinking" reaction (default: `"eyes"`)
+- `inline_keyboards` — Enable inline keyboard buttons (default: false)
+
+## discord
+
+**Required:**
+- `token` — Bot token from Discord Developer Portal
+
+**Notable optional:**
+- `allow_channels` — Restrict to specific channel IDs
+- `group_policy` — `"mention"` (default) or `"open"`
+- `streaming` — Enable streaming (default: true)
+- `proxy` — HTTP proxy URL
+- `intents` — Discord gateway intents (default: 37377)
+- `read_receipt_emoji` — Emoji for read receipt
+- `working_emoji` — Emoji for "working" indicator
+
+## feishu
+
+**Required:**
+- `app_id` — Feishu app ID
+- `app_secret` — Feishu app secret
+
+**Notable optional:**
+- `encrypt_key` — Event encryption key
+- `verification_token` — Event verification token
+- `domain` — `"feishu"` (default) or `"lark"`
+- `group_policy` — `"mention"` (default) or `"open"`
+- `streaming` — Enable streaming (default: true)
+
+## dingtalk
+
+**Required:**
+- `client_id` — DingTalk app client ID
+- `client_secret` — DingTalk app client secret
+
+**Notable optional:**
+- `allow_from` — Allowed user IDs
+
+## slack
+
+**Required:**
+- `bot_token` — Bot OAuth token (`xoxb-...`)
+- `app_token` — App-level token (`xapp-...`)
+
+**Notable optional:**
+- `mode` — `"socket"` (default, Socket Mode) or `"webhook"`
+- `reply_in_thread` — Reply in thread (default: true)
+- `react_emoji` — "thinking" emoji (default: `"eyes"`)
+- `done_emoji` — "done" emoji (default: `"white_check_mark"`)
+- `group_policy` — `"mention"` (default) or `"open"`
+- `dm.enabled` — Enable DM support
+- `dm.policy` — DM policy
+- `dm.allow_from` — Allowed DM users
+
+## wecom
+
+**Required:**
+- `bot_id` — WeCom bot ID
+- `secret` — WeCom bot secret
+
+**Notable optional:**
+- `allow_from` — Allowed users
+- `welcome_message` — Welcome message for new chats
+
+## weixin
+
+**Required:**
+- `token` — WeChat Official Account token
+
+**Notable optional:**
+- `base_url` — API base URL
+- `cdn_base_url` — CDN base URL
+- `state_dir` — State persistence directory
+- `poll_timeout` — Long polling timeout
+
+## whatsapp
+
+**Required:**
+- `bridge_token` — WhatsApp bridge token (auto-generated if absent)
+
+**Notable optional:**
+- `bridge_url` — Bridge WebSocket URL (default: `"ws://localhost:3001"`)
+- `group_policy` — `"open"` (default) or `"mention"`
+
+## qq
+
+**Required:**
+- `app_id` — QQ bot app ID
+- `secret` — QQ bot secret
+
+**Notable optional:**
+- `msg_format` — `"plain"` or `"markdown"`
+- `ack_message` — Acknowledgment message text
+- `media_dir` — Media file directory
+
+## email
+
+**Required:**
+- `imap_host` — IMAP server hostname
+- `imap_username` — IMAP login username
+- `imap_password` — IMAP login password
+- `smtp_host` — SMTP server hostname
+- `smtp_username` — SMTP login username
+- `smtp_password` — SMTP login password
+- `from_address` — Sender email address
+
+**Notable optional:**
+- `imap_port` — IMAP port (default: 993)
+- `smtp_port` — SMTP port (default: 587)
+- `imap_use_ssl` — Use SSL for IMAP (default: true)
+- `smtp_use_tls` — Use TLS for SMTP (default: true)
+- `poll_interval_seconds` — Polling interval (default: 30)
+- `mark_seen` — Mark emails as read (default: true)
+- `max_body_chars` — Max email body length (default: 12000)
+- `subject_prefix` — Reply subject prefix (default: `"Re: "`)
+- `verify_dkim` — Verify DKIM signatures (default: true)
+- `verify_spf` — Verify SPF records (default: true)
+- `allowed_attachment_types` — Allowed file extensions
+- `max_attachment_size` — Max attachment size in bytes
+- `consent_granted` — Must be set to `true` for the channel to start (default: false)
+- `auto_reply_enabled` — Enable auto-reply (default: true)
+
+## matrix
+
+**Required:**
+- `user_id` — Matrix user ID (e.g. `@bot:matrix.org`)
+- `password` or `access_token` — Login password OR access token
+
+**Notable optional:**
+- `homeserver` — Homeserver URL (default: `"https://matrix.org"`)
+- `device_id` — Device ID
+- `e2eeEnabled` — Enable end-to-end encryption (default: true)
+- `group_policy` — `"open"`, `"mention"`, or `"allowlist"`
+- `streaming` — Enable streaming (default: false)
+- `max_media_bytes` — Max media file size (default: 20MB)
+
+## msteams
+
+**Required:**
+- `app_id` — Azure AD app ID
+- `app_password` — Azure AD app password/secret
+- `tenant_id` — Azure AD tenant ID
+
+**Notable optional:**
+- `host` — Listen host (default: `"0.0.0.0"`)
+- `port` — Listen port (default: 3978)
+- `reply_in_thread` — Reply in thread (default: true)
+- `validate_inbound_auth` — Validate incoming auth (default: true)
+
+## mochat
+
+**Required:**
+- `claw_token` — MoChat Claw token
+
+**Notable optional:**
+- `base_url` — API base URL
+- `socket_url` — WebSocket URL
+- `refresh_interval_ms` — Refresh interval in ms
+- `watch_timeout_ms` — Watch timeout in ms
+
+## websocket
+
+Built-in WebSocket channel for programmatic access.
+
+**Required:**
+- `token` — Authentication token (enabled by default; set `websocket_requires_token: false` to disable)
+
+**Notable optional:**
+- `host` — Listen host (default: `"127.0.0.1"`)
+- `port` — Listen port (default: 8765)
+- `allow_from` — Allowed origins (default: `["*"]`)
+- `streaming` — Enable streaming (default: true)

nanobot/skills/create-instance/scripts/create_instance.py

@@ -0,0 +1,250 @@
+#!/usr/bin/env python3
+"""Create a new nanobot instance with a dedicated config and workspace.
+
+Usage:
+    create_instance.py --name <name> --channel <channel> [--model <model>] [--config-dir <dir>]
+
+Examples:
+    create_instance.py --name telegram-bot --channel telegram
+    create_instance.py --name discord-bot --channel discord --model deepseek/deepseek-chat
+    create_instance.py --name my-bot --channel telegram --config-dir ~/.nanobot-custom
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+import socket
+import sys
+from pathlib import Path
+
+
+def _validate_name(name: str) -> str:
+    """Normalize and validate instance name."""
+    name = name.strip().lower()
+    name = re.sub(r"[^a-z0-9-]", "-", name)
+    name = re.sub(r"-{2,}", "-", name)
+    name = name.strip("-")
+    if not name:
+        print("[ERROR] Instance name must contain at least one letter or digit.", file=sys.stderr)
+        sys.exit(1)
+    if len(name) > 64:
+        print(f"[ERROR] Instance name too long ({len(name)} chars, max 64).", file=sys.stderr)
+        sys.exit(1)
+    return name
+
+
+def _get_available_channels() -> list[str]:
+    """Get list of available channel names without importing channel classes."""
+    from nanobot.channels.registry import discover_channel_names
+
+    return discover_channel_names()
+
+
+def _run_onboard(config_path: Path, workspace: Path) -> None:
+    """Create skeleton config + workspace using nanobot's programmatic API."""
+    from nanobot.cli.commands import _onboard_plugins
+    from nanobot.config.loader import save_config, set_config_path
+    from nanobot.config.paths import get_workspace_path
+    from nanobot.config.schema import Config
+    from nanobot.utils.helpers import sync_workspace_templates
+
+    config = Config()
+    config.agents.defaults.workspace = str(workspace)
+    set_config_path(config_path)
+    save_config(config, config_path)
+    _onboard_plugins(config_path)
+
+    workspace_path = get_workspace_path(config.workspace_path)
+    if not workspace_path.exists():
+        workspace_path.mkdir(parents=True, exist_ok=True)
+    sync_workspace_templates(workspace_path)
+
+
+def _patch_config(
+    config_path: Path,
+    *,
+    channel: str,
+    workspace: Path,
+    model: str | None,
+    inherit_config_path: Path | None = None,
+) -> dict:
+    """Patch the generated config: enable channel, set workspace, optionally set model."""
+    data = json.loads(config_path.read_text(encoding="utf-8"))
+
+    # Inherit providers and model from current instance
+    if inherit_config_path and inherit_config_path.exists():
+        try:
+            src = json.loads(inherit_config_path.read_text(encoding="utf-8"))
+
+            # Inherit providers (API keys, api_base, etc.)
+            src_providers = src.get("providers", {})
+            if src_providers:
+                data.setdefault("providers", {})
+                for key, val in src_providers.items():
+                    if isinstance(val, dict) and val.get("apiKey"):
+                        data["providers"][key] = val
+
+            # Inherit model if not explicitly overridden
+            if not model:
+                parent_model = src.get("agents", {}).get("defaults", {}).get("model")
+                if parent_model:
+                    model = parent_model
+
+        except Exception as exc:
+            print(f"[WARN] Could not inherit from {inherit_config_path}: {exc}", file=sys.stderr)
+
+    # Set workspace and model
+    data.setdefault("agents", {}).setdefault("defaults", {})
+    data["agents"]["defaults"]["workspace"] = str(workspace)
+    if model:
+        data["agents"]["defaults"]["model"] = model
+
+    # Enable the target channel
+    channels = data.setdefault("channels", {})
+    if channel in channels and isinstance(channels[channel], dict):
+        channels[channel]["enabled"] = True
+    else:
+        channels[channel] = {"enabled": True}
+
+    # Auto-assign ports if defaults are already in use
+    _assign_free_ports(data)
+
+    # Validate with Pydantic, then save
+    from nanobot.config.schema import Config
+
+    Config.model_validate(data)
+    config_path.write_text(json.dumps(data, indent=2, ensure_ascii=False), encoding="utf-8")
+    return data
+
+
+def _is_port_in_use(port: int, host: str = "127.0.0.1") -> bool:
+    """Check if a port is already in use."""
+    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
+        try:
+            s.bind((host, port))
+            return False
+        except OSError:
+            return True
+
+
+def _find_free_port(start: int, host: str = "127.0.0.1", max_tries: int = 100) -> int:
+    """Find the first free port starting from `start`."""
+    for port in range(start, start + max_tries):
+        if not _is_port_in_use(port, host):
+            return port
+    # OS-level fallback: ask the kernel for an ephemeral port
+    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
+        s.bind((host, 0))
+        return s.getsockname()[1]
+
+
+def _assign_free_ports(data: dict) -> None:
+    """If default gateway or API ports are in use, assign free ones."""
+    from nanobot.config.schema import ApiConfig, GatewayConfig
+
+    defaults = [
+        ("gateway", GatewayConfig()),
+        ("api", ApiConfig()),
+    ]
+    for key, default_cfg in defaults:
+        section = data.setdefault(key, {})
+        port = section.get("port", default_cfg.port)
+        host = section.get("host", default_cfg.host)
+        if _is_port_in_use(port, host):
+            section["port"] = _find_free_port(port + 1, host)
+
+
+def _get_channel_required_fields(channel: str) -> list[str]:
+    """Inspect a channel's default config and list fields that are empty strings."""
+    try:
+        from nanobot.channels.registry import load_channel_class
+
+        cls = load_channel_class(channel)
+        default = cls.default_config()
+        return sorted(k for k, v in default.items() if isinstance(v, str) and v == "" and k != "enabled")
+    except Exception as exc:
+        print(f"[WARN] Could not inspect channel '{channel}' defaults: {exc}", file=sys.stderr)
+        return []
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        description="Create a new nanobot instance.",
+    )
+    parser.add_argument("--name", required=True, help="Instance name (e.g. telegram-bot)")
+    parser.add_argument("--channel", required=True, help="Channel type (e.g. telegram, discord)")
+    parser.add_argument("--model", default=None, help="LLM model (default: same as current instance)")
+    parser.add_argument(
+        "--config-dir",
+        default=None,
+        help="Config directory (default: ~/.nanobot-{name})",
+    )
+    parser.add_argument(
+        "--inherit-config",
+        default=None,
+        help="Path to current instance's config.json to copy API keys from",
+    )
+    args = parser.parse_args()
+
+    # Validate name
+    name = _validate_name(args.name)
+
+    # Validate channel
+    available = _get_available_channels()
+    if args.channel not in available:
+        print(f"[ERROR] Unknown channel: {args.channel}", file=sys.stderr)
+        print(f"Available channels: {', '.join(sorted(available))}", file=sys.stderr)
+        sys.exit(1)
+
+    # Resolve paths
+    home = Path.home()
+    config_dir = Path(args.config_dir).expanduser().resolve() if args.config_dir else home / f".nanobot-{name}"
+    config_path = config_dir / "config.json"
+    workspace = config_dir / "workspace"
+
+    # Check for duplicate
+    if config_path.exists():
+        print(f"[ERROR] Config already exists at {config_path}", file=sys.stderr)
+        print("Delete it first or use a different --config-dir.", file=sys.stderr)
+        sys.exit(1)
+
+    print(f"Creating instance '{name}'...")
+    print(f"  Config dir: {config_dir}")
+    print(f"  Workspace:  {workspace}")
+    print(f"  Channel:    {args.channel}")
+    if args.model:
+        print(f"  Model:      {args.model}")
+
+    # Run onboard
+    _run_onboard(config_path, workspace)
+
+    # Patch config
+    inherit_path = Path(args.inherit_config).expanduser().resolve() if args.inherit_config else None
+    _patch_config(
+        config_path,
+        channel=args.channel,
+        workspace=workspace,
+        model=args.model,
+        inherit_config_path=inherit_path,
+    )
+
+    # Report
+    print(f"\n[OK] Instance '{name}' created successfully.")
+    print(f"  Config: {config_path}")
+    print(f"  Workspace: {workspace}")
+
+    # List fields the user needs to fill in
+    required_fields = _get_channel_required_fields(args.channel)
+    if required_fields:
+        print(f"\n[IMPORTANT] Edit {config_path} and fill in these fields:")
+        for field in required_fields:
+            print(f"  - channels.{args.channel}.{field}")
+
+    print(f"\nTo start the instance:")
+    print(f"  nanobot gateway --config {config_path}")
+
+
+if __name__ == "__main__":
+    main()

tests/agent/test_onboard_logic.py

@@ -4,27 +4,22 @@ These tests focus on the business logic behind the onboard wizard,
 without testing the interactive UI components.
 """
 
-import json
 from pathlib import Path
 from types import SimpleNamespace
 from typing import Any, cast
 
-import pytest
 from pydantic import BaseModel, Field
 
 from nanobot.cli import onboard as onboard_wizard
-
-# Import functions to test
 from nanobot.cli.commands import _merge_missing_defaults
 from nanobot.cli.onboard import (
     _BACK_PRESSED,
     _configure_pydantic_model,
     _format_value,
+    _get_constraint_hint,
     _get_field_display_name,
     _get_field_type_info,
-    _get_constraint_hint,
     _input_text,
-    _validate_field_constraint,
     run_onboard,
 )
 from nanobot.config.schema import Config
@@ -640,8 +635,8 @@ class TestValidateFieldConstraint:
 
     def test_real_send_max_retries_field(self):
         """Validate against the actual ChannelsConfig.send_max_retries field."""
-        from nanobot.config.schema import ChannelsConfig
         from nanobot.cli.onboard import _validate_field_constraint
+        from nanobot.config.schema import ChannelsConfig
 
         field_info = ChannelsConfig.model_fields["send_max_retries"]
         assert _validate_field_constraint(3, field_info) is None
@@ -833,12 +828,11 @@ class TestMainMenuUpdate:
 
     def test_main_menu_dispatch_includes_channel_common(self):
         """Main menu dispatch should route [H] to Channel Common."""
-        from nanobot.cli.onboard import run_onboard
 
         # We verify by checking the dispatch table is set up correctly
         # The menu items are defined inline in run_onboard, so we test
         # that _configure_general_settings handles the new sections.
-        from nanobot.cli.onboard import _SETTINGS_SECTIONS, _SETTINGS_GETTER, _SETTINGS_SETTER
+        from nanobot.cli.onboard import _SETTINGS_GETTER, _SETTINGS_SECTIONS, _SETTINGS_SETTER
 
         assert "Channel Common" in _SETTINGS_SECTIONS
         assert "Channel Common" in _SETTINGS_GETTER
@@ -846,7 +840,7 @@ class TestMainMenuUpdate:
 
     def test_main_menu_dispatch_includes_api_server(self):
         """Main menu dispatch should route [I] to API Server."""
-        from nanobot.cli.onboard import _SETTINGS_SECTIONS, _SETTINGS_GETTER, _SETTINGS_SETTER
+        from nanobot.cli.onboard import _SETTINGS_GETTER, _SETTINGS_SECTIONS, _SETTINGS_SETTER
 
         assert "API Server" in _SETTINGS_SECTIONS
         assert "API Server" in _SETTINGS_GETTER
@@ -960,3 +954,464 @@ class TestMainMenuUpdate:
 
         assert result.should_save is True
         assert pause_called["n"] == 1
+
+
+class TestInputTextEmptyString:
+    """Tests for _input_text empty-string handling bug fix."""
+
+    def test_empty_string_returned_not_none(self, monkeypatch):
+        """_input_text should return empty string, not None, when user enters ''."""
+        monkeypatch.setattr(
+            onboard_wizard,
+            "_get_questionary",
+            lambda: SimpleNamespace(text=lambda *a, **kw: SimpleNamespace(ask=lambda: "")),
+        )
+
+        result = _input_text("Name", "old", "str")
+        assert result == ""
+
+    def test_none_still_returns_none(self, monkeypatch):
+        """_input_text should return None when questionary returns None."""
+        monkeypatch.setattr(
+            onboard_wizard,
+            "_get_questionary",
+            lambda: SimpleNamespace(text=lambda *a, **kw: SimpleNamespace(ask=lambda: None)),
+        )
+
+        result = _input_text("Name", "old", "str")
+        assert result is None
+
+
+class TestIsStrOrNone:
+    """Tests for _is_str_or_none helper."""
+
+    def test_str_or_none_true(self):
+        from nanobot.cli.onboard import _is_str_or_none
+
+        assert _is_str_or_none(str | None) is True
+
+    def test_optional_str_true(self):
+        from typing import Optional
+        from nanobot.cli.onboard import _is_str_or_none
+
+        assert _is_str_or_none(Optional[str]) is True
+
+    def test_str_only_false(self):
+        from nanobot.cli.onboard import _is_str_or_none
+
+        assert _is_str_or_none(str) is False
+
+    def test_int_or_none_false(self):
+        from nanobot.cli.onboard import _is_str_or_none
+
+        assert _is_str_or_none(int | None) is False
+
+
+class TestConfigurePydanticModelEmptyString:
+    """Tests that optional string fields are cleared when empty string is entered."""
+
+    def test_optional_str_empty_string_becomes_none(self, monkeypatch):
+        """Entering '' for an optional str field should set it to None."""
+        from pydantic import BaseModel
+        from nanobot.cli.onboard import _is_str_or_none
+
+        class M(BaseModel):
+            api_key: str | None = None
+
+        model = M(api_key="secret")
+
+        call_count = {"select": 0}
+
+        def fake_select(_prompt, choices, default=None):
+            call_count["select"] += 1
+            # First call: select the api_key field, then Done
+            if call_count["select"] == 1:
+                for c in choices:
+                    if "Api Key" in c:
+                        return c
+                return choices[0]
+            return "[Done]"
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select)
+        monkeypatch.setattr(onboard_wizard, "_show_config_panel", lambda *a, **kw: None)
+        # Simulate user entering empty string
+        monkeypatch.setattr(
+            onboard_wizard, "_input_with_existing", lambda *a, **kw: ""
+        )
+
+        result = _configure_pydantic_model(model, "Test")
+        assert result is not None
+        assert result.api_key is None
+
+    def test_required_str_empty_string_kept(self, monkeypatch):
+        """Entering '' for a required str field should keep the empty string."""
+        from pydantic import BaseModel
+
+        class M(BaseModel):
+            api_key: str = ""
+
+        model = M(api_key="secret")
+
+        call_count = {"select": 0}
+
+        def fake_select(_prompt, choices, default=None):
+            call_count["select"] += 1
+            if call_count["select"] == 1:
+                for c in choices:
+                    if "Api Key" in c:
+                        return c
+                return choices[0]
+            return "[Done]"
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select)
+        monkeypatch.setattr(onboard_wizard, "_show_config_panel", lambda *a, **kw: None)
+        monkeypatch.setattr(
+            onboard_wizard, "_input_with_existing", lambda *a, **kw: ""
+        )
+
+        result = _configure_pydantic_model(model, "Test")
+        assert result is not None
+        assert result.api_key == ""
+
+
+class TestModelPresetWizard:
+    """Tests for model preset CRUD in the onboard wizard."""
+
+    def test_sync_preset_cache(self):
+        """_sync_preset_cache should populate the module-level cache."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _sync_preset_cache
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        config.model_presets = {
+            "fast": ModelPresetConfig(model="gpt-4.1-mini"),
+            "power": ModelPresetConfig(model="gpt-4.1"),
+        }
+        _sync_preset_cache(config)
+        assert _MODEL_PRESET_CACHE == {"fast", "power"}
+
+    def test_model_preset_add(self, monkeypatch):
+        """_configure_model_presets should add a new preset."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _configure_model_presets
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        _MODEL_PRESET_CACHE.clear()
+
+        responses = iter([
+            "[+] Add new preset",
+            "my-preset",
+            "<- Back",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_text(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_configure(*_model, **_kwargs):
+            return ModelPresetConfig(model="gpt-test", temperature=0.5)
+
+        # _select_with_back returns a string/sentinel directly (not a prompt object)
+        def fake_select_with_back(*_args, **_kwargs):
+            return next(responses)
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select_with_back)
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select, text=fake_text))
+        monkeypatch.setattr(onboard_wizard, "_configure_pydantic_model", fake_configure)
+        monkeypatch.setattr(onboard_wizard, "_show_section_header", lambda *a, **kw: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None))
+
+        _configure_model_presets(config)
+
+        assert "my-preset" in config.model_presets
+        assert config.model_presets["my-preset"].model == "gpt-test"
+        assert config.model_presets["my-preset"].temperature == 0.5
+
+    def test_model_preset_delete(self, monkeypatch):
+        """_configure_model_presets should delete an existing preset."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _configure_model_presets
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        config.model_presets = {"old": ModelPresetConfig(model="x")}
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.add("old")
+
+        responses = iter([
+            "old (x)",
+            "Delete",
+            True,
+            "<- Back",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_confirm(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_select_with_back(*_args, **_kwargs):
+            return next(responses)
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select_with_back)
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select, confirm=fake_confirm))
+        monkeypatch.setattr(onboard_wizard, "_show_section_header", lambda *a, **kw: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None))
+
+        _configure_model_presets(config)
+
+        assert "old" not in config.model_presets
+        assert "old" not in _MODEL_PRESET_CACHE
+
+    def test_model_preset_field_handler(self, monkeypatch):
+        """_handle_model_preset_field should set a preset name from choices."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_model_preset_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.update({"fast", "power"})
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "fast")
+
+        defaults = AgentDefaults()
+        _handle_model_preset_field(defaults, "model_preset", "Model Preset", None)
+        assert defaults.model_preset == "fast"
+
+    def test_model_preset_field_handler_clear(self, monkeypatch):
+        """_handle_model_preset_field should clear preset when (clear/unset) chosen."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_model_preset_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.add("fast")
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "(clear/unset)")
+
+        defaults = AgentDefaults(model_preset="fast")
+        _handle_model_preset_field(defaults, "model_preset", "Model Preset", "fast")
+        assert defaults.model_preset is None
+
+    def test_main_menu_dispatch_includes_model_presets(self):
+        """run_onboard dispatch should route [M] to Model Presets."""
+        from nanobot.cli.onboard import _configure_model_presets
+
+        # The function should be importable and callable
+        assert callable(_configure_model_presets)
+
+    def test_run_onboard_model_presets_edit(self, monkeypatch):
+        """run_onboard should handle [M] Model Presets correctly."""
+        initial_config = Config()
+
+        responses = iter([
+            "[M] Model Presets",
+            KeyboardInterrupt(),
+            "[S] Save and Exit",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        preset_mutated = {"n": 0}
+
+        def fake_configure_model_presets(config):
+            preset_mutated["n"] += 1
+            # Mutate config so unsaved changes are detected
+            from nanobot.config.schema import ModelPresetConfig
+            config.model_presets["test"] = ModelPresetConfig(model="x")
+
+        monkeypatch.setattr(onboard_wizard, "_show_main_menu_header", lambda: None)
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select))
+        monkeypatch.setattr(onboard_wizard, "_configure_model_presets", fake_configure_model_presets)
+
+        result = run_onboard(initial_config=initial_config)
+
+        assert result.should_save is True
+        assert preset_mutated["n"] == 1
+
+    def test_summary_shows_model_presets(self, monkeypatch):
+        """_show_summary should include model presets panel."""
+        from nanobot.cli.onboard import _show_summary
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        config.model_presets = {
+            "fast": ModelPresetConfig(model="gpt-4.1-mini"),
+        }
+
+        panels = []
+
+        def fake_print_summary(rows, title):
+            panels.append(title)
+
+        monkeypatch.setattr(onboard_wizard, "_print_summary_panel", fake_print_summary)
+        monkeypatch.setattr(onboard_wizard, "_get_provider_names", lambda: {})
+        monkeypatch.setattr(onboard_wizard, "_get_channel_names", lambda: {})
+        monkeypatch.setattr(onboard_wizard, "_pause", lambda: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(print=lambda *a, **kw: None))
+
+        _show_summary(config)
+
+        assert "Model Presets" in panels
+
+    def test_provider_field_handler(self, monkeypatch):
+        """_handle_provider_field should set a provider from the registry list."""
+        from nanobot.cli.onboard import _handle_provider_field
+        from nanobot.config.schema import ModelPresetConfig
+
+        monkeypatch.setattr(
+            onboard_wizard, "_get_provider_names", lambda: {"moonshot": "Moonshot", "openai": "OpenAI"}
+        )
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "moonshot")
+
+        preset = ModelPresetConfig(model="x")
+        _handle_provider_field(preset, "provider", "Provider", "auto")
+        assert preset.provider == "moonshot"
+
+    def test_provider_field_handler_back_pressed(self, monkeypatch):
+        """_handle_provider_field should not modify value when back is pressed."""
+        from nanobot.cli.onboard import _BACK_PRESSED, _handle_provider_field
+        from nanobot.config.schema import ModelPresetConfig
+
+        monkeypatch.setattr(
+            onboard_wizard, "_get_provider_names", lambda: {"moonshot": "Moonshot"}
+        )
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: _BACK_PRESSED)
+
+        preset = ModelPresetConfig(model="x", provider="auto")
+        _handle_provider_field(preset, "provider", "Provider", "auto")
+        assert preset.provider == "auto"
+
+    def test_fallback_presets_add_preset_and_done(self, monkeypatch):
+        """_handle_fallback_presets_field should add a preset and save on Done."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_fallback_presets_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.update({"fast", "power"})
+
+        responses = iter(["[+] Add preset", "[Done]"])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "fast")
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select))
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None, print=lambda *a, **kw: None))
+
+        defaults = AgentDefaults()
+        _handle_fallback_presets_field(defaults, "fallback_presets", "Fallback Presets", [])
+        assert defaults.fallback_presets == ["fast"]
+
+    def test_fallback_presets_back_preserves_existing(self, monkeypatch):
+        """_handle_fallback_presets_field should not modify value on Back."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_fallback_presets_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.add("fast")
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt("<- Back")
+
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select))
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None, print=lambda *a, **kw: None))
+
+        defaults = AgentDefaults(fallback_presets=["existing"])
+        _handle_fallback_presets_field(defaults, "fallback_presets", "Fallback Presets", ["existing"])
+        assert defaults.fallback_presets == ["existing"]
+
+    def test_fallback_presets_remove_last(self, monkeypatch):
+        """_handle_fallback_presets_field should remove last item."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_fallback_presets_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+
+        responses = iter(["[-] Remove last", "[Done]"])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select))
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None, print=lambda *a, **kw: None))
+
+        defaults = AgentDefaults(fallback_presets=["a", "b"])
+        _handle_fallback_presets_field(defaults, "fallback_presets", "Fallback Presets", ["a", "b"])
+        assert defaults.fallback_presets == ["a"]
+
+    def test_fallback_presets_no_presets_shows_warning(self, monkeypatch):
+        """_handle_fallback_presets_field should warn when no presets exist."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_fallback_presets_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+
+        responses = iter(["[+] Add preset", "[Done]"])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select, press_any_key_to_continue=lambda: FakePrompt(None)))
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None, print=lambda *a, **kw: None))
+
+        defaults = AgentDefaults()
+        _handle_fallback_presets_field(defaults, "fallback_presets", "Fallback Presets", [])
+        assert defaults.fallback_presets == []

tests/agent/test_self_model_preset.py

@@ -0,0 +1,89 @@
+# tests/agent/test_self_model_preset.py
+from pathlib import Path
+from typing import Any
+from unittest.mock import MagicMock
+
+from nanobot.agent.loop import AgentLoop
+from nanobot.config.schema import ModelPresetConfig, MyToolConfig, ToolsConfig
+from nanobot.providers.base import GenerationSettings
+
+
+def _make_loop(presets: dict | None = None) -> tuple[AgentLoop, Any]:
+    provider = MagicMock()
+    provider.get_default_model.return_value = "test-model"
+    provider.generation = GenerationSettings(temperature=0.1, max_tokens=8192)
+
+    def _factory(name: str):
+        preset = (presets or {}).get(name)
+        if preset:
+            new_provider = MagicMock()
+            new_provider.generation = GenerationSettings(
+                temperature=preset.temperature,
+                max_tokens=preset.max_tokens,
+                reasoning_effort=preset.reasoning_effort,
+            )
+            return new_provider
+        return provider
+
+    loop = AgentLoop(
+        bus=MagicMock(),
+        provider=provider,
+        workspace=Path("/tmp/test"),
+        model="test-model",
+        context_window_tokens=65536,
+        model_presets=presets or {},
+        provider_factory=_factory,
+        tools_config=ToolsConfig(my=MyToolConfig(allow_set=True)),
+    )
+    tool = loop.tools.get("my")
+    return loop, tool
+
+
+async def test_set_model_preset_updates_all_fields() -> None:
+    presets = {
+        "gpt5": ModelPresetConfig(
+            model="gpt-5",
+            provider="openai",
+            max_tokens=16384,
+            context_window_tokens=128000,
+            temperature=0.2,
+        ),
+    }
+    loop, tool = _make_loop(presets)
+    await tool.execute(action="set", key="model_preset", value="gpt5")
+
+    assert loop.model == "gpt-5"
+    assert loop.context_window_tokens == 128000
+    assert loop.provider.generation.temperature == 0.2
+    assert loop.provider.generation.max_tokens == 16384
+    assert loop._active_preset == "gpt5"
+
+
+async def test_set_model_preset_unknown_returns_error() -> None:
+    loop, tool = _make_loop({})
+    result = await tool.execute(action="set", key="model_preset", value="nope")
+
+    assert "Error" in result or "not found" in result
+
+
+async def test_check_model_preset_shows_current() -> None:
+    presets = {"gpt5": ModelPresetConfig(model="gpt-5", provider="openai")}
+    loop, tool = _make_loop(presets)
+    await tool.execute(action="set", key="model_preset", value="gpt5")
+    result = await tool.execute(action="check", key="model_preset")
+
+    assert "gpt5" in result
+
+
+async def test_check_model_presets_shows_available() -> None:
+    presets = {
+        "gpt5": ModelPresetConfig(model="gpt-5", provider="openai"),
+        "ds": ModelPresetConfig(model="deepseek-chat", provider="deepseek"),
+    }
+    loop, tool = _make_loop(presets)
+    result = await tool.execute(action="check", key="model_presets")
+
+    assert "gpt5" in result
+    assert "ds" in result
+
+

tests/agent/tools/test_self_tool.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 import time
 from pathlib import Path
-from unittest.mock import AsyncMock, MagicMock
+from unittest.mock import MagicMock
 
 import pytest
 from pydantic import BaseModel
@@ -35,6 +35,7 @@ def _make_mock_loop(**overrides):
     loop._concurrency_gate = None
     loop._unified_session = False
     loop._extra_hooks = []
+    loop.model_preset = None
 
     # web_config mock — needed for check tests
     loop.web_config = MagicMock()
@@ -76,7 +77,7 @@ class TestInspectSummary:
         tool = _make_tool()
         result = await tool.execute(action="check")
         assert "max_iterations: 40" in result
-        assert "context_window_tokens: 65536" in result
+        assert "model_preset" in result
 
     @pytest.mark.asyncio
     async def test_inspect_includes_runtime_vars(self):
@@ -92,8 +93,7 @@ class TestInspectSummary:
         tool = _make_tool()
         result = await tool.execute(action="check")
         assert "max_iterations" in result
-        assert "context_window_tokens" in result
+        assert "model_preset" in result
-        assert "model" in result
         assert "workspace" in result
         assert "provider_retry_mode" in result
         assert "max_tool_result_chars" in result
@@ -231,13 +231,13 @@ class TestModifyRestricted:
     @pytest.mark.asyncio
     async def test_modify_string_int_coerced(self):
         tool = _make_tool()
-        result = await tool.execute(action="set", key="max_iterations", value="80")
+        await tool.execute(action="set", key="max_iterations", value="80")
         assert tool._loop.max_iterations == 80
 
     @pytest.mark.asyncio
     async def test_modify_context_window_valid(self):
         tool = _make_tool()
-        result = await tool.execute(action="set", key="context_window_tokens", value=131072)
+        await tool.execute(action="set", key="context_window_tokens", value=131072)
         assert tool._loop.context_window_tokens == 131072
 
     @pytest.mark.asyncio
@@ -337,13 +337,13 @@ class TestModifyFree:
     @pytest.mark.asyncio
     async def test_modify_allows_list(self):
         tool = _make_tool()
-        result = await tool.execute(action="set", key="items", value=[1, 2, 3])
+        await tool.execute(action="set", key="items", value=[1, 2, 3])
         assert tool._loop._runtime_vars["items"] == [1, 2, 3]
 
     @pytest.mark.asyncio
     async def test_modify_allows_dict(self):
         tool = _make_tool()
-        result = await tool.execute(action="set", key="data", value={"a": 1})
+        await tool.execute(action="set", key="data", value={"a": 1})
         assert tool._loop._runtime_vars["data"] == {"a": 1}
 
     @pytest.mark.asyncio
@@ -392,6 +392,26 @@ class TestModifyFree:
         assert "Error" in result
         assert tool._loop.max_tool_result_chars == 16000
 
+    @pytest.mark.asyncio
+    async def test_modify_model_clears_active_preset(self):
+        """Directly modifying model must clear _active_preset so state stays consistent."""
+        tool = _make_tool()
+        tool._loop._active_preset = "gpt5"
+        result = await tool.execute(action="set", key="model", value="other-model")
+        assert "Set model" in result
+        assert tool._loop.model == "other-model"
+        assert tool._loop._active_preset is None
+
+    @pytest.mark.asyncio
+    async def test_modify_context_window_tokens_clears_active_preset(self):
+        """Directly modifying context_window_tokens must clear _active_preset."""
+        tool = _make_tool()
+        tool._loop._active_preset = "gpt5"
+        result = await tool.execute(action="set", key="context_window_tokens", value=32768)
+        assert "Set context_window_tokens" in result
+        assert tool._loop.context_window_tokens == 32768
+        assert tool._loop._active_preset is None
+
 
 # ---------------------------------------------------------------------------
 # set — previously BLOCKED/READONLY now open
@@ -689,8 +709,8 @@ class TestSubagentHookStatus:
     @pytest.mark.asyncio
     async def test_after_iteration_updates_status(self):
         """after_iteration should copy iteration, tool_events, usage to status."""
-        from nanobot.agent.subagent import SubagentStatus, _SubagentHook
         from nanobot.agent.hook import AgentHookContext
+        from nanobot.agent.subagent import SubagentStatus, _SubagentHook
 
         status = SubagentStatus(
             task_id="test",
@@ -716,8 +736,8 @@ class TestSubagentHookStatus:
     @pytest.mark.asyncio
     async def test_after_iteration_with_error(self):
         """after_iteration should set status.error when context has an error."""
-        from nanobot.agent.subagent import SubagentStatus, _SubagentHook
         from nanobot.agent.hook import AgentHookContext
+        from nanobot.agent.subagent import SubagentStatus, _SubagentHook
 
         status = SubagentStatus(
             task_id="test",
@@ -739,8 +759,8 @@ class TestSubagentHookStatus:
     @pytest.mark.asyncio
     async def test_after_iteration_no_status_is_noop(self):
         """after_iteration with no status should be a no-op."""
-        from nanobot.agent.subagent import _SubagentHook
         from nanobot.agent.hook import AgentHookContext
+        from nanobot.agent.subagent import _SubagentHook
 
         hook = _SubagentHook("test")
         context = AgentHookContext(iteration=1, messages=[])
@@ -757,7 +777,6 @@ class TestCheckpointCallback:
     async def test_checkpoint_updates_phase_and_iteration(self):
         """The _on_checkpoint callback should update status.phase and iteration."""
         from nanobot.agent.subagent import SubagentStatus
-        import asyncio
 
         status = SubagentStatus(
             task_id="cp",

tests/channels/test_weixin_channel.py

@@ -48,11 +48,11 @@ def test_make_headers_includes_route_tag_when_configured() -> None:
     assert headers["Authorization"] == "Bearer token"
     assert headers["SKRouteTag"] == "123"
     assert headers["iLink-App-Id"] == "bot"
-    assert headers["iLink-App-ClientVersion"] == str((2 << 16) | (1 << 8) | 1)
+    assert headers["iLink-App-ClientVersion"] == str((2 << 16) | (1 << 8) | 7)
 
 
 def test_channel_version_matches_reference_plugin_version() -> None:
-    assert WEIXIN_CHANNEL_VERSION == "2.1.1"
+    assert WEIXIN_CHANNEL_VERSION == "2.1.7"
 
 
 def test_save_and_load_state_persists_context_tokens(tmp_path) -> None:
@@ -1250,3 +1250,222 @@ async def test_send_text_succeeds_on_zero_errcode() -> None:
     await channel._send_text("wx-user", "hello", "ctx-ok")
 
     channel._api_post.assert_awaited_once()
+
+
+@pytest.mark.asyncio
+async def test_send_text_raises_on_nonzero_ret_even_when_errcode_zero() -> None:
+    """_send_text must raise when the API returns ret != 0, even if errcode is 0.
+
+    The iLink API signals failure through either field.  Checking only errcode
+    caused silent message drops (responses generated but never delivered).
+    """
+    channel, _bus = _make_channel()
+    channel._client = object()
+    channel._token = "token"
+    channel._api_post = AsyncMock(
+        return_value={"ret": -100, "errcode": 0, "errmsg": "internal error"}
+    )
+
+    with pytest.raises(RuntimeError, match="WeChat send text error.*ret=-100.*errcode=0"):
+        await channel._send_text("wx-user", "hello", "ctx-ok")
+
+    channel._api_post.assert_awaited_once()
+
+
+# ---------------------------------------------------------------------------
+# Tests for _poll_once not silently dropping messages on processing errors
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_poll_once_logs_exception_on_process_message_failure(monkeypatch) -> None:
+    """When _process_message raises, _poll_once must log the error and continue
+    processing remaining messages instead of silently swallowing the exception."""
+    channel, _bus = _make_channel()
+    channel._client = SimpleNamespace(timeout=None)
+    channel._token = "token"
+    channel._get_updates_buf = "old-buf"
+
+    calls = []
+    logged_messages: list[str] = []
+
+    async def _failing_process(msg: dict) -> None:
+        calls.append(msg.get("message_id"))
+        if msg.get("message_id") == "msg-1":
+            raise RuntimeError("processing failed")
+
+    channel._process_message = _failing_process  # type: ignore[method-assign]
+
+    monkeypatch.setattr(
+        channel.logger,
+        "exception",
+        lambda message, *args, **kwargs: logged_messages.append(str(message)),
+    )
+
+    channel._api_post = AsyncMock(  # type: ignore[method-assign]
+        return_value={
+            "ret": 0,
+            "errcode": 0,
+            "get_updates_buf": "new-buf",
+            "msgs": [
+                {"message_id": "msg-1", "message_type": 1},
+                {"message_id": "msg-2", "message_type": 1},
+            ],
+        }
+    )
+
+    await channel._poll_once()
+
+    # Both messages should have been attempted
+    assert calls == ["msg-1", "msg-2"]
+    # Buffer should still advance (already updated before processing)
+    assert channel._get_updates_buf == "new-buf"
+    # Error should be logged
+    assert any("Failed to process WeChat message" in m for m in logged_messages)
+
+
+@pytest.mark.asyncio
+async def test_poll_loop_logs_exception_and_continues_on_poll_failure(monkeypatch) -> None:
+    """When _poll_once raises a non-timeout exception, the start() loop must log
+    the error and continue polling instead of exiting silently."""
+    channel, _bus = _make_channel()
+    channel._client = object()
+    channel._token = "token"
+    channel.config.token = "token"  # skip QR login in start()
+    channel._running = True
+
+    call_count = 0
+    logged_messages: list[str] = []
+
+    async def _failing_poll() -> None:
+        nonlocal call_count
+        call_count += 1
+        if call_count == 1:
+            raise RuntimeError("poll exploded")
+        channel._running = False  # Stop after second call
+
+    channel._poll_once = _failing_poll  # type: ignore[method-assign]
+
+    monkeypatch.setattr(
+        channel.logger,
+        "exception",
+        lambda message, *args, **kwargs: logged_messages.append(str(message)),
+    )
+
+    # Use a tiny retry delay so the test finishes quickly
+    original_retry = weixin_mod.RETRY_DELAY_S
+    weixin_mod.RETRY_DELAY_S = 0.01
+    try:
+        await channel.start()
+    finally:
+        weixin_mod.RETRY_DELAY_S = original_retry
+
+    assert call_count == 2
+    assert any("WeChat poll loop error" in m for m in logged_messages)
+
+
+@pytest.mark.asyncio
+async def test_send_text_retries_without_context_token_on_ret_minus_two() -> None:
+    """If sendmessage returns ret=-2 with a context_token, retry without it."""
+    channel, _bus = _make_channel()
+    channel._client = object()
+    channel._token = "token"
+    channel._context_tokens["wx-user"] = "expired-token"
+
+    channel._api_post = AsyncMock(
+        side_effect=[
+            {"ret": -2},  # first attempt with token fails
+            {"ret": 0},   # retry without token succeeds
+        ]
+    )
+
+    await channel._send_text("wx-user", "hello", "expired-token")
+
+    # Should have called API twice
+    assert channel._api_post.await_count == 2
+    # First call includes context_token
+    first_body = channel._api_post.await_args_list[0].args[1]
+    assert first_body["msg"]["context_token"] == "expired-token"
+    # Second call does NOT include context_token
+    second_body = channel._api_post.await_args_list[1].args[1]
+    assert "context_token" not in second_body["msg"]
+    # Expired token should be cleared from cache
+    assert "wx-user" not in channel._context_tokens
+
+
+@pytest.mark.asyncio
+async def test_send_text_raises_when_retry_also_fails_with_stale_session() -> None:
+    """If both attempts return stale-session ret=-2, raise so ChannelManager retries."""
+    channel, _bus = _make_channel()
+    channel._client = object()
+    channel._token = "token"
+    channel._context_tokens["wx-user"] = "bad-token"
+
+    channel._api_post = AsyncMock(
+        side_effect=[
+            {"ret": -2},  # with token
+            {"ret": -2},  # without token
+        ]
+    )
+
+    with pytest.raises(RuntimeError, match="WeChat send text error"):
+        await channel._send_text("wx-user", "hello", "bad-token")
+
+    assert channel._api_post.await_count == 2
+    # Token is NOT cleared because retry also failed
+    assert channel._context_tokens.get("wx-user") == "bad-token"
+
+
+@pytest.mark.asyncio
+async def test_send_text_raises_on_ret_minus_two_when_no_context_token() -> None:
+    """If no context_token was provided, ret=-2 stale session is raised."""
+    channel, _bus = _make_channel()
+    channel._client = object()
+    channel._token = "token"
+
+    channel._api_post = AsyncMock(return_value={"ret": -2})
+
+    with pytest.raises(RuntimeError, match="WeChat send text error"):
+        await channel._send_text("wx-user", "hello", "")
+
+    # Only one API call (no retry possible without token)
+    channel._api_post.assert_awaited_once()
+
+
+# ---------------------------------------------------------------------------
+# Tests for _is_stale_session_ret (hermes-agent#17228 / #18105)
+# ---------------------------------------------------------------------------
+
+
+class TestIsStaleSessionRet:
+    """Verify stale-session detection for iLink ret=-2 / errcode=-2 responses."""
+
+    def test_ret_minus_2_with_empty_errmsg_is_stale(self):
+        assert weixin_mod._is_stale_session_ret(-2, 0, "") is True
+        assert weixin_mod._is_stale_session_ret(-2, 0, None) is True
+
+    def test_errcode_minus_2_with_empty_errmsg_is_stale(self):
+        assert weixin_mod._is_stale_session_ret(0, -2, "") is True
+        assert weixin_mod._is_stale_session_ret(0, -2, None) is True
+
+    def test_ret_minus_2_with_unknown_error_is_stale(self):
+        assert weixin_mod._is_stale_session_ret(-2, 0, "unknown error") is True
+        assert weixin_mod._is_stale_session_ret(-2, 0, "UNKNOWN ERROR") is True
+
+    def test_errcode_minus_2_with_unknown_error_is_stale(self):
+        assert weixin_mod._is_stale_session_ret(0, -2, "unknown error") is True
+
+    def test_ret_minus_2_with_frequency_limit_is_not_stale(self):
+        assert weixin_mod._is_stale_session_ret(-2, 0, "frequency limit") is False
+        assert weixin_mod._is_stale_session_ret(-2, 0, "too frequently") is False
+
+    def test_errcode_minus_2_with_frequency_limit_is_not_stale(self):
+        assert weixin_mod._is_stale_session_ret(0, -2, "freq limit") is False
+
+    def test_success_codes_are_not_stale(self):
+        assert weixin_mod._is_stale_session_ret(0, 0, "") is False
+        assert weixin_mod._is_stale_session_ret(0, 0, None) is False
+
+    def test_other_errors_are_not_stale(self):
+        assert weixin_mod._is_stale_session_ret(-14, -14, "session timeout") is False
+        assert weixin_mod._is_stale_session_ret(-100, 0, "internal error") is False

tests/cli/test_commands.py

@@ -9,7 +9,7 @@ import pytest
 from typer.testing import CliRunner
 
 from nanobot.bus.events import OutboundMessage
-from nanobot.cli.commands import _make_provider, app
+from nanobot.cli.commands import app
 from nanobot.config.schema import Config
 from nanobot.cron.types import CronJob, CronPayload
 from nanobot.providers.factory import ProviderSnapshot
@@ -488,8 +488,8 @@ def test_openai_compat_provider_passes_model_through():
 
 
 def test_make_provider_uses_github_copilot_backend():
-    from nanobot.cli.commands import _make_provider
     from nanobot.config.schema import Config
+    from nanobot.providers.factory import build_provider_for_preset
 
     config = Config.model_validate(
         {
@@ -503,7 +503,7 @@ def test_make_provider_uses_github_copilot_backend():
     )
 
     with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
-        provider = _make_provider(config)
+        provider = build_provider_for_preset(config, config.resolve_preset())
 
     assert provider.__class__.__name__ == "GitHubCopilotProvider"
 
@@ -562,6 +562,8 @@ def test_openai_codex_strip_prefix_supports_hyphen_and_underscore():
 
 
 def test_make_provider_passes_extra_headers_to_custom_provider():
+    from nanobot.providers.factory import build_provider_for_preset
+
     config = Config.model_validate(
         {
             "agents": {"defaults": {"provider": "custom", "model": "gpt-4o-mini"}},
@@ -579,7 +581,7 @@ def test_make_provider_passes_extra_headers_to_custom_provider():
     )
 
     with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI") as mock_async_openai:
-        _make_provider(config)
+        build_provider_for_preset(config, config.resolve_preset())
 
     kwargs = mock_async_openai.call_args.kwargs
     assert kwargs["api_key"] == "test-key"
@@ -597,11 +599,11 @@ def mock_agent_runtime(tmp_path):
     with patch("nanobot.config.loader.load_config", return_value=config) as mock_load_config, \
          patch("nanobot.config.loader.resolve_config_env_vars", side_effect=lambda c: c), \
          patch("nanobot.cli.commands.sync_workspace_templates") as mock_sync_templates, \
-         patch("nanobot.cli.commands._make_provider", return_value=object()), \
+         patch("nanobot.providers.factory.build_provider_for_preset", return_value=MagicMock(generation=MagicMock(max_tokens=8192))), \
          patch("nanobot.cli.commands._print_agent_response") as mock_print_response, \
          patch("nanobot.bus.queue.MessageBus"), \
          patch("nanobot.cron.service.CronService"), \
-         patch("nanobot.agent.loop.AgentLoop") as mock_agent_loop_cls:
+         patch("nanobot.cli.commands.AgentLoop") as mock_agent_loop_cls:
         agent_loop = MagicMock()
         agent_loop.channels_config = None
         agent_loop.process_direct = AsyncMock(
@@ -609,6 +611,7 @@ def mock_agent_runtime(tmp_path):
         )
         agent_loop.close_mcp = AsyncMock(return_value=None)
         mock_agent_loop_cls.return_value = agent_loop
+        mock_agent_loop_cls.from_config.return_value = agent_loop
 
         yield {
             "config": config,
@@ -639,7 +642,7 @@ def test_agent_uses_default_config_when_no_workspace_or_config_flags(mock_agent_
     assert mock_agent_runtime["sync_templates"].call_args.args == (
         mock_agent_runtime["config"].workspace_path,
     )
-    assert mock_agent_runtime["agent_loop_cls"].call_args.kwargs["workspace"] == (
+    assert mock_agent_runtime["agent_loop_cls"].from_config.call_args.args[0].workspace_path == (
         mock_agent_runtime["config"].workspace_path
     )
     mock_agent_runtime["agent_loop"].process_direct.assert_awaited_once()
@@ -672,7 +675,7 @@ def test_agent_config_sets_active_path(monkeypatch, tmp_path: Path) -> None:
     )
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: object())
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *a, **k: MagicMock(generation=MagicMock(max_tokens=8192)))
     monkeypatch.setattr("nanobot.bus.queue.MessageBus", lambda: object())
     monkeypatch.setattr("nanobot.cron.service.CronService", lambda _store: object())
 
@@ -680,13 +683,17 @@ def test_agent_config_sets_active_path(monkeypatch, tmp_path: Path) -> None:
         def __init__(self, *args, **kwargs) -> None:
             pass
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(channel="cli", chat_id="direct", content="ok")
 
         async def close_mcp(self) -> None:
             return None
 
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.cli.commands._print_agent_response", lambda *_args, **_kwargs: None)
 
     result = runner.invoke(app, ["agent", "-m", "hello", "-c", str(config_file)])
@@ -707,7 +714,7 @@ def test_agent_uses_workspace_directory_for_cron_store(monkeypatch, tmp_path: Pa
     monkeypatch.setattr("nanobot.config.loader.set_config_path", lambda _path: None)
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: object())
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *a, **k: MagicMock(generation=MagicMock(max_tokens=8192)))
     monkeypatch.setattr("nanobot.bus.queue.MessageBus", lambda: object())
 
     class _FakeCron:
@@ -718,6 +725,10 @@ def test_agent_uses_workspace_directory_for_cron_store(monkeypatch, tmp_path: Pa
         def __init__(self, *args, **kwargs) -> None:
             pass
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(channel="cli", chat_id="direct", content="ok")
 
@@ -725,7 +736,7 @@ def test_agent_uses_workspace_directory_for_cron_store(monkeypatch, tmp_path: Pa
             return None
 
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCron)
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.cli.commands._print_agent_response", lambda *_args, **_kwargs: None)
 
     result = runner.invoke(app, ["agent", "-m", "hello", "-c", str(config_file)])
@@ -753,7 +764,7 @@ def test_agent_workspace_override_does_not_migrate_legacy_cron(
     monkeypatch.setattr("nanobot.config.loader.set_config_path", lambda _path: None)
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: object())
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *a, **k: MagicMock(generation=MagicMock(max_tokens=8192)))
     monkeypatch.setattr("nanobot.bus.queue.MessageBus", lambda: object())
     monkeypatch.setattr("nanobot.config.paths.get_cron_dir", lambda: legacy_dir)
 
@@ -765,6 +776,10 @@ def test_agent_workspace_override_does_not_migrate_legacy_cron(
         def __init__(self, *args, **kwargs) -> None:
             pass
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(channel="cli", chat_id="direct", content="ok")
 
@@ -772,7 +787,7 @@ def test_agent_workspace_override_does_not_migrate_legacy_cron(
             return None
 
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCron)
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.cli.commands._print_agent_response", lambda *_args, **_kwargs: None)
 
     result = runner.invoke(
@@ -806,7 +821,7 @@ def test_agent_custom_config_workspace_does_not_migrate_legacy_cron(
     monkeypatch.setattr("nanobot.config.loader.set_config_path", lambda _path: None)
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: object())
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *a, **k: MagicMock(generation=MagicMock(max_tokens=8192)))
     monkeypatch.setattr("nanobot.bus.queue.MessageBus", lambda: object())
     monkeypatch.setattr("nanobot.config.paths.get_cron_dir", lambda: legacy_dir)
 
@@ -818,6 +833,10 @@ def test_agent_custom_config_workspace_does_not_migrate_legacy_cron(
         def __init__(self, *args, **kwargs) -> None:
             pass
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(channel="cli", chat_id="direct", content="ok")
 
@@ -825,7 +844,7 @@ def test_agent_custom_config_workspace_does_not_migrate_legacy_cron(
             return None
 
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCron)
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr(
         "nanobot.cli.commands._print_agent_response", lambda *_args, **_kwargs: None
     )
@@ -846,7 +865,7 @@ def test_agent_overrides_workspace_path(mock_agent_runtime):
     assert result.exit_code == 0
     assert mock_agent_runtime["config"].agents.defaults.workspace == str(workspace_path)
     assert mock_agent_runtime["sync_templates"].call_args.args == (workspace_path,)
-    assert mock_agent_runtime["agent_loop_cls"].call_args.kwargs["workspace"] == workspace_path
+    assert mock_agent_runtime["agent_loop_cls"].from_config.call_args.args[0].workspace_path == workspace_path
 
 
 def test_agent_workspace_override_wins_over_config_workspace(mock_agent_runtime, tmp_path: Path):
@@ -863,7 +882,7 @@ def test_agent_workspace_override_wins_over_config_workspace(mock_agent_runtime,
     assert mock_agent_runtime["load_config"].call_args.args == (config_path.resolve(),)
     assert mock_agent_runtime["config"].agents.defaults.workspace == str(workspace_path)
     assert mock_agent_runtime["sync_templates"].call_args.args == (workspace_path,)
-    assert mock_agent_runtime["agent_loop_cls"].call_args.kwargs["workspace"] == workspace_path
+    assert mock_agent_runtime["agent_loop_cls"].from_config.call_args.args[0].workspace_path == workspace_path
 
 
 def test_agent_hints_about_deprecated_memory_window(mock_agent_runtime, tmp_path):
@@ -928,8 +947,8 @@ def _patch_cli_command_runtime(
         sync_templates or (lambda _path: None),
     )
     monkeypatch.setattr(
-        "nanobot.cli.commands._make_provider",
+        "nanobot.providers.factory.build_provider_for_preset",
-        provider_factory,
+        lambda *_a, **_k: provider_factory(Config()),
     )
     monkeypatch.setattr(
         "nanobot.providers.factory.build_provider_snapshot",
@@ -962,6 +981,10 @@ def _patch_serve_runtime(monkeypatch, config: Config, seen: dict[str, object]) -
         def __init__(self, **kwargs) -> None:
             seen["workspace"] = kwargs["workspace"]
 
+        @classmethod
+        def from_config(cls, config, bus=None, **kwargs):
+            return cls(workspace=config.workspace_path, **kwargs)
+
         async def _connect_mcp(self) -> None:
             return None
 
@@ -985,7 +1008,7 @@ def _patch_serve_runtime(monkeypatch, config: Config, seen: dict[str, object]) -
         message_bus=lambda: object(),
         session_manager=lambda _workspace: object(),
     )
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.api.server.create_app", _fake_create_app)
     monkeypatch.setattr("aiohttp.web.run_app", _fake_run_app)
 
@@ -1077,7 +1100,7 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
     monkeypatch.setattr("nanobot.config.loader.set_config_path", lambda _path: None)
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: provider)
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *_a, **_k: provider)
     monkeypatch.setattr(
         "nanobot.providers.factory.build_provider_snapshot",
         lambda _config: _test_provider_snapshot(provider, _config),
@@ -1117,8 +1140,13 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
     class _FakeAgentLoop:
         def __init__(self, *args, **kwargs) -> None:
             self.model = "test-model"
+            self.provider = object()
             self.tools = {}
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(
                 channel="telegram",
@@ -1152,7 +1180,7 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
         return True
 
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCron)
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.channels.manager.ChannelManager", _StopAfterCronSetup)
     monkeypatch.setattr(
         "nanobot.utils.evaluator.evaluate_response",
@@ -1181,7 +1209,7 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
 
     assert response == "Time to stretch."
     assert seen["response"] == "Time to stretch."
-    assert seen["provider"] is provider
+    assert seen["provider"] is not None  # provider resolved inside AgentLoop
     assert seen["model"] == "test-model"
     assert seen["task_context"] == (
         "The scheduled time has arrived. Deliver this reminder to the user now, "
@@ -1228,7 +1256,7 @@ def test_gateway_cron_job_suppresses_intermediate_progress(
     monkeypatch.setattr("nanobot.config.loader.set_config_path", lambda _path: None)
     monkeypatch.setattr("nanobot.config.loader.load_config", lambda _path=None: config)
     monkeypatch.setattr("nanobot.cli.commands.sync_workspace_templates", lambda _path: None)
-    monkeypatch.setattr("nanobot.cli.commands._make_provider", lambda _config: object())
+    monkeypatch.setattr("nanobot.providers.factory.build_provider_for_preset", lambda *a, **k: MagicMock(generation=MagicMock(max_tokens=8192)))
     monkeypatch.setattr(
         "nanobot.providers.factory.build_provider_snapshot",
         lambda _config: _test_provider_snapshot(object(), _config),
@@ -1248,8 +1276,13 @@ def test_gateway_cron_job_suppresses_intermediate_progress(
     class _FakeAgentLoop:
         def __init__(self, *args, **kwargs) -> None:
             self.model = "test-model"
+            self.provider = object()
             self.tools = {}
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(*args, **kwargs)
+
         async def process_direct(self, *_args, on_progress=None, **_kwargs):
             seen["on_progress"] = on_progress
             return OutboundMessage(
@@ -1275,7 +1308,7 @@ def test_gateway_cron_job_suppresses_intermediate_progress(
         return False
 
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCron)
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.channels.manager.ChannelManager", _StopAfterCronSetup)
     monkeypatch.setattr(
         "nanobot.utils.evaluator.evaluate_response",
@@ -1480,9 +1513,14 @@ def test_gateway_health_endpoint_binds_and_serves_expected_responses(
     class _FakeAgentLoop:
         def __init__(self, **_kwargs) -> None:
             self.model = "test-model"
+            self.provider = object()
             self.dream = _FakeDream()
             self.sessions = _FakeSessionManager()
 
+        @classmethod
+        def from_config(cls, *args, **kwargs):
+            return cls(**kwargs)
+
         async def run(self) -> None:
             await asyncio.Event().wait()
 
@@ -1571,7 +1609,7 @@ def test_gateway_health_endpoint_binds_and_serves_expected_responses(
         message_bus=lambda: object(),
         session_manager=lambda _workspace: object(),
     )
-    monkeypatch.setattr("nanobot.agent.loop.AgentLoop", _FakeAgentLoop)
+    monkeypatch.setattr("nanobot.cli.commands.AgentLoop", _FakeAgentLoop)
     monkeypatch.setattr("nanobot.channels.manager.ChannelManager", _FakeChannelManager)
     monkeypatch.setattr("nanobot.cron.service.CronService", _FakeCronService)
     monkeypatch.setattr("nanobot.heartbeat.service.HeartbeatService", _FakeHeartbeatService)

tests/config/test_model_presets.py

@@ -0,0 +1,264 @@
+from nanobot.config.schema import Config, ModelPresetConfig
+
+
+def test_model_preset_config_accepts_model_and_provider_separately() -> None:
+    preset = ModelPresetConfig(model="gpt-5", provider="openai")
+    assert preset.model == "gpt-5"
+    assert preset.provider == "openai"
+
+
+def test_model_preset_config_defaults() -> None:
+    preset = ModelPresetConfig(model="test-model")
+    assert preset.provider == "auto"
+    assert preset.max_tokens == 8192
+    assert preset.context_window_tokens == 65_536
+    assert preset.temperature == 0.1
+    assert preset.reasoning_effort is None
+
+
+def test_model_preset_config_all_fields() -> None:
+    preset = ModelPresetConfig(
+        model="deepseek-r1",
+        provider="deepseek",
+        max_tokens=16384,
+        context_window_tokens=131072,
+        temperature=0.2,
+        reasoning_effort="high",
+    )
+    assert preset.model == "deepseek-r1"
+    assert preset.provider == "deepseek"
+    assert preset.max_tokens == 16384
+    assert preset.context_window_tokens == 131072
+    assert preset.temperature == 0.2
+    assert preset.reasoning_effort == "high"
+
+
+def test_config_accepts_model_presets_dict() -> None:
+    cfg = Config(model_presets={
+        "gpt5": ModelPresetConfig(model="gpt-5", provider="openai", max_tokens=16384),
+        "ds": ModelPresetConfig(model="deepseek-chat", provider="deepseek"),
+    })
+    assert "gpt5" in cfg.model_presets
+    assert cfg.model_presets["gpt5"].max_tokens == 16384
+    assert cfg.model_presets["ds"].model == "deepseek-chat"
+
+
+def test_resolve_preset_returns_preset_values() -> None:
+    cfg = Config.model_validate({
+        "model_presets": {
+            "gpt5": {
+                "model": "gpt-5",
+                "provider": "openai",
+                "max_tokens": 16384,
+                "context_window_tokens": 128000,
+                "temperature": 0.2,
+            },
+        },
+        "agents": {"defaults": {"model_preset": "gpt5"}},
+    })
+    r = cfg.resolve_preset()
+    assert r.model == "gpt-5"
+    assert r.provider == "openai"
+    assert r.max_tokens == 16384
+    assert r.context_window_tokens == 128000
+    assert r.temperature == 0.2
+
+
+def test_resolve_preset_ignores_old_config_fields() -> None:
+    """Preset wins completely — old config remnants are ignored."""
+    cfg = Config.model_validate({
+        "model_presets": {
+            "gpt5": {
+                "model": "gpt-5",
+                "provider": "openai",
+                "max_tokens": 16384,
+                "context_window_tokens": 128000,
+                "temperature": 0.2,
+            },
+        },
+        "agents": {
+            "defaults": {
+                "model_preset": "gpt5",
+                "model": "old-model",
+                "temperature": 0.5,
+            },
+        },
+    })
+    r = cfg.resolve_preset()
+    assert r.model == "gpt-5"
+    assert r.temperature == 0.2
+    assert r.max_tokens == 16384
+
+
+def test_preset_not_found_raises_error() -> None:
+    import pytest
+    with pytest.raises(Exception, match="model_preset.*not found"):
+        Config.model_validate({
+            "model_presets": {},
+            "agents": {"defaults": {"model_preset": "nonexistent"}},
+        })
+
+
+def test_fallback_presets_invalid_preset_raises_error() -> None:
+    import pytest
+    with pytest.raises(Exception, match="fallback_presets.*not found"):
+        Config.model_validate({
+            "model_presets": {
+                "valid": {"model": "gpt-4"},
+            },
+            "agents": {"defaults": {"fallback_presets": ["invalid_preset"]}},
+        })
+
+
+def test_resolve_preset_without_preset_returns_defaults() -> None:
+    """Backward compat: no explicit preset → resolve_preset returns the auto-created 'default' preset."""
+    cfg = Config.model_validate({
+        "agents": {"defaults": {"model": "deepseek-chat"}},
+    })
+    assert cfg.agents.defaults.model_preset == "default"
+    r = cfg.resolve_preset()
+    assert r.model == "deepseek-chat"
+    assert r.max_tokens == 8192
+
+
+def test_agent_loop_stores_model_presets() -> None:
+    from pathlib import Path
+    from unittest.mock import MagicMock
+
+    from nanobot.agent.loop import AgentLoop
+
+    presets = {
+        "gpt5": ModelPresetConfig(model="gpt-5", provider="openai"),
+    }
+    provider = MagicMock()
+    provider.get_default_model.return_value = "test"
+
+    loop = AgentLoop(
+        bus=MagicMock(),
+        provider=provider,
+        workspace=Path("/tmp/test"),
+        model_presets=presets,
+    )
+    assert loop.model_presets == presets
+
+
+def test_resolve_preset_with_reasoning_effort() -> None:
+    cfg = Config.model_validate({
+        "model_presets": {
+            "ds-r1": {
+                "model": "deepseek-r1",
+                "provider": "deepseek",
+                "reasoning_effort": "high",
+            },
+        },
+        "agents": {"defaults": {"model_preset": "ds-r1"}},
+    })
+    assert cfg.resolve_preset().reasoning_effort == "high"
+
+
+def test_preset_routes_to_correct_provider() -> None:
+    """resolve_preset + _match_provider uses the preset's model+provider."""
+    cfg = Config.model_validate({
+        "model_presets": {
+            "ds": {"model": "deepseek-chat", "provider": "deepseek"},
+        },
+        "providers": {"deepseek": {"api_key": "test-key"}},
+        "agents": {"defaults": {"model_preset": "ds"}},
+    })
+    provider_name = cfg.get_provider_name()
+    assert provider_name == "deepseek"
+
+
+def test_preset_with_auto_provider_uses_keyword_matching() -> None:
+    cfg = Config.model_validate({
+        "model_presets": {
+            "auto-ds": {"model": "deepseek-chat", "provider": "auto"},
+        },
+        "providers": {"deepseek": {"api_key": "test-key"}},
+        "agents": {"defaults": {"model_preset": "auto-ds"}},
+    })
+    provider_name = cfg.get_provider_name()
+    assert provider_name == "deepseek"
+
+
+def test_backward_compat_no_preset() -> None:
+    """Existing configs without model_presets are automatically promoted to the 'default' preset."""
+    cfg = Config.model_validate({
+        "providers": {"anthropic": {"api_key": "test-key"}},
+        "agents": {"defaults": {"model": "anthropic/claude-opus-4-5"}},
+    })
+    assert cfg.resolve_preset().model == "anthropic/claude-opus-4-5"
+    assert cfg.agents.defaults.model_preset == "default"
+    assert "default" in cfg.model_presets
+    assert cfg.get_provider_name() == "anthropic"
+
+
+def test_resolve_preset_overrides_all_model_fields() -> None:
+    """When model_preset is set, resolve_preset returns preset values, not individual fields."""
+    cfg = Config.model_validate({
+        "model_presets": {
+            "gpt5": {"model": "gpt-5", "provider": "openai", "max_tokens": 16384},
+        },
+        "providers": {"openai": {"api_key": "test-key"}},
+        "agents": {
+            "defaults": {
+                "model_preset": "gpt5",
+                "model": "legacy-model",
+                "max_tokens": 4096,
+            },
+        },
+    })
+    r = cfg.resolve_preset()
+    assert r.model == "gpt-5"
+    assert r.provider == "openai"
+    assert r.max_tokens == 16384
+
+
+def test_empty_model_presets_dict_is_harmless() -> None:
+    cfg = Config.model_validate({"model_presets": {}})
+    assert cfg.resolve_preset().model == "anthropic/claude-opus-4-5"
+
+
+def test_factory_uses_preset_provider_not_defaults() -> None:
+    """When creating a provider for a non-active preset, the preset's own provider must be used."""
+    from nanobot.providers.factory import make_provider_factory
+
+    cfg = Config.model_validate({
+        "model_presets": {
+            "kimi": {"model": "kimi-k2.6", "provider": "moonshot"},
+            "zhipu": {"model": "glm-5.1", "provider": "zhipu"},
+        },
+        "providers": {
+            "moonshot": {"api_key": "moonshot-key", "api_base": "https://api.moonshot.ai/v1"},
+            "zhipu": {"api_key": "zhipu-key", "api_base": "https://open.bigmodel.cn/api/paas/v4"},
+        },
+        "agents": {"defaults": {"model_preset": "kimi"}},
+    })
+
+    factory = make_provider_factory(cfg)
+    zhipu_provider = factory("zhipu")
+
+    assert zhipu_provider.api_base == "https://open.bigmodel.cn/api/paas/v4"
+    assert getattr(zhipu_provider, "api_key", None) == "zhipu-key"
+
+    # Also verify the active preset provider is still correct
+    moonshot_provider = factory("kimi")
+    assert moonshot_provider.api_base == "https://api.moonshot.ai/v1"
+
+
+def test_factory_rejects_unknown_preset_name() -> None:
+    """Factory must raise ValueError when asked for a preset not in model_presets."""
+    import pytest
+
+    from nanobot.providers.factory import make_provider_factory
+
+    cfg = Config.model_validate({
+        "model_presets": {
+            "known": {"model": "gpt-4", "provider": "openai"},
+        },
+        "providers": {"openai": {"api_key": "test-key"}},
+    })
+
+    factory = make_provider_factory(cfg)
+    with pytest.raises(ValueError, match="Preset 'unknown' not found"):
+        factory("unknown")

tests/skills/test_create_instance.py

@@ -0,0 +1,168 @@
+"""Tests for nanobot/skills/create-instance/scripts/create_instance.py."""
+
+from __future__ import annotations
+
+import json
+import socket
+import subprocess
+import sys
+from pathlib import Path
+
+import pytest
+
+SCRIPT = Path(__file__).parent.parent.parent / "nanobot" / "skills" / "create-instance" / "scripts" / "create_instance.py"
+
+
+@pytest.fixture
+def tmp_home(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> Path:
+    """Point HOME at a temp dir so nanobot writes configs there."""
+    monkeypatch.setenv("HOME", str(tmp_path))
+    monkeypatch.delenv("NANOBOT_CONFIG", raising=False)
+    return tmp_path
+
+
+def _run_script(*args: str, cwd: Path | None = None) -> subprocess.CompletedProcess:
+    """Run create_instance.py as a subprocess."""
+    return subprocess.run(
+        [sys.executable, str(SCRIPT), *args],
+        capture_output=True,
+        text=True,
+        encoding="utf-8",
+        errors="replace",
+        cwd=cwd,
+    )
+
+
+class TestValidation:
+    """Argument validation tests."""
+
+    def test_missing_required_args_exits_with_error(self) -> None:
+        result = _run_script()
+        assert result.returncode != 0
+
+    def test_invalid_channel_exits_with_error(self, tmp_home: Path) -> None:
+        result = _run_script("--name", "test", "--channel", "nonexistent_channel")
+        assert result.returncode != 0
+        assert "nonexistent_channel" in result.stderr or "nonexistent_channel" in result.stdout
+
+
+class TestCreateInstance:
+    """End-to-end instance creation tests."""
+
+    def test_creates_config_and_workspace(self, tmp_home: Path) -> None:
+        config_dir = tmp_home / ".nanobot-test"
+        result = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+        )
+        assert result.returncode == 0, result.stderr
+
+        config_path = config_dir / "config.json"
+        assert config_path.exists(), f"Config not created at {config_path}"
+
+        workspace = config_dir / "workspace"
+        assert workspace.exists(), f"Workspace not created at {workspace}"
+
+    def test_config_has_channel_enabled(self, tmp_home: Path) -> None:
+        config_dir = tmp_home / ".nanobot-test"
+        result = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+        )
+        assert result.returncode == 0, result.stderr
+
+        data = json.loads((config_dir / "config.json").read_text(encoding="utf-8"))
+        assert data["channels"]["telegram"]["enabled"] is True
+
+    def test_config_workspace_path_set(self, tmp_home: Path) -> None:
+        config_dir = tmp_home / ".nanobot-test"
+        result = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+        )
+        assert result.returncode == 0, result.stderr
+
+        data = json.loads((config_dir / "config.json").read_text(encoding="utf-8"))
+        ws = data["agents"]["defaults"]["workspace"]
+        assert str(config_dir / "workspace") in ws or "workspace" in ws
+
+    def test_model_override(self, tmp_home: Path) -> None:
+        config_dir = tmp_home / ".nanobot-test"
+        result = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--model", "deepseek/deepseek-chat",
+            "--config-dir", str(config_dir),
+        )
+        assert result.returncode == 0, result.stderr
+
+        data = json.loads((config_dir / "config.json").read_text(encoding="utf-8"))
+        assert data["agents"]["defaults"]["model"] == "deepseek/deepseek-chat"
+
+    def test_rejects_duplicate_instance(self, tmp_home: Path) -> None:
+        config_dir = tmp_home / ".nanobot-test"
+        result1 = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+        )
+        assert result1.returncode == 0
+
+        result2 = _run_script(
+            "--name", "test-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+        )
+        assert result2.returncode != 0
+
+    def test_port_reassigned_when_default_in_use(self, tmp_home: Path) -> None:
+        """When default gateway port is occupied, script should pick a different one."""
+        config_dir = tmp_home / ".nanobot-test"
+
+        # Bind to the default gateway port to simulate a running instance
+        with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as blocker:
+            blocker.bind(("127.0.0.1", 18790))
+            blocker.listen(1)
+
+            result = _run_script(
+                "--name", "test-bot",
+                "--channel", "telegram",
+                "--config-dir", str(config_dir),
+            )
+            assert result.returncode == 0, result.stderr
+
+        data = json.loads((config_dir / "config.json").read_text(encoding="utf-8"))
+        assert data["gateway"]["port"] != 18790
+
+    def test_inherits_api_key_from_current_instance(self, tmp_home: Path) -> None:
+        """API keys from --inherit-config should be copied to new instance."""
+        # Create a fake "current instance" config with an API key
+        src_dir = tmp_home / ".nanobot-current"
+        src_dir.mkdir()
+        src_config = src_dir / "config.json"
+        src_config.write_text(json.dumps({
+            "providers": {
+                "anthropic": {"apiKey": "sk-test-key-12345"},
+                "deepseek": {"apiKey": "dsk-another-key"},
+                "openai": {},  # no key, should not be copied
+            },
+        }), encoding="utf-8")
+
+        config_dir = tmp_home / ".nanobot-new"
+        result = _run_script(
+            "--name", "new-bot",
+            "--channel", "telegram",
+            "--config-dir", str(config_dir),
+            "--inherit-config", str(src_config),
+        )
+        assert result.returncode == 0, result.stderr
+
+        data = json.loads((config_dir / "config.json").read_text(encoding="utf-8"))
+        providers = data.get("providers", {})
+        assert providers.get("anthropic", {}).get("apiKey") == "sk-test-key-12345"
+        assert providers.get("deepseek", {}).get("apiKey") == "dsk-another-key"
+        # openai had no key, so it should not be in the new config's providers
+        assert providers.get("openai", {}).get("apiKey") is None

tests/test_nanobot_facade.py

@@ -39,7 +39,7 @@ def test_from_config_default_path():
     from nanobot.config.schema import Config
 
     with patch("nanobot.config.loader.load_config") as mock_load, \
-         patch("nanobot.nanobot._make_provider") as mock_prov:
+         patch("nanobot.providers.factory.build_provider_for_preset") as mock_prov:
         mock_load.return_value = Config()
         mock_prov.return_value = MagicMock()
         mock_prov.return_value.get_default_model.return_value = "test"
@@ -127,7 +127,7 @@ def test_workspace_override(tmp_path):
 
 def test_sdk_make_provider_uses_github_copilot_backend():
     from nanobot.config.schema import Config
-    from nanobot.nanobot import _make_provider
+    from nanobot.providers.factory import make_provider
 
     config = Config.model_validate(
         {
@@ -141,7 +141,7 @@ def test_sdk_make_provider_uses_github_copilot_backend():
     )
 
     with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
-        provider = _make_provider(config)
+        provider = make_provider(config)
 
     assert provider.__class__.__name__ == "GitHubCopilotProvider"
 

tests/test_preset_failover_smoke.py

@@ -0,0 +1,467 @@
+"""End-to-end smoke tests for model presets + failover.
+
+Uses a local aiohttp fake OpenAI server so requests are real HTTP,
+not mocked at the provider level.
+"""
+
+from __future__ import annotations
+
+import json
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from nanobot.nanobot import Nanobot
+from nanobot.providers.base import GenerationSettings, LLMProvider
+from nanobot.providers.failover import ModelRouter
+from nanobot.providers.openai_compat_provider import OpenAICompatProvider
+
+try:
+    from aiohttp import web
+    from aiohttp.test_utils import TestServer
+
+    HAS_AIOHTTP = True
+except ImportError:
+    HAS_AIOHTTP = False
+
+
+@pytest.fixture(autouse=True)
+def _disable_proxy_for_localhost_tests(monkeypatch):
+    """Prevent httpx from routing localhost requests through a system proxy."""
+    monkeypatch.delenv("ALL_PROXY", raising=False)
+    monkeypatch.delenv("HTTP_PROXY", raising=False)
+    monkeypatch.delenv("HTTPS_PROXY", raising=False)
+    monkeypatch.setenv("NO_PROXY", "127.0.0.1,localhost")
+
+
+# ---------------------------------------------------------------------------
+# Helpers (mock-level preset tests)
+# ---------------------------------------------------------------------------
+
+def _write_config(tmp_path: Path, **overrides) -> Path:
+    data = {
+        "providers": {
+            "openrouter": {"apiKey": "sk-test-key"},
+            "openai": {"apiKey": "sk-openai-test"},
+        },
+        "agents": {"defaults": {"model": "openai/gpt-4.1"}},
+        "tools": {"my": {"allowSet": True}},
+    }
+    data.update(overrides)
+    config_path = tmp_path / "config.json"
+    config_path.write_text(json.dumps(data))
+    return config_path
+
+
+
+# ---------------------------------------------------------------------------
+# 1. Model Preset Mock Tests
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_preset_loaded_at_startup(tmp_path: Path) -> None:
+    config_path = _write_config(
+        tmp_path,
+        model_presets={
+            "fast": {
+                "model": "gpt-4.1-mini",
+                "provider": "openai",
+                "max_tokens": 4096,
+                "context_window_tokens": 128000,
+                "temperature": 0.3,
+            }
+        },
+        agents={"defaults": {"model_preset": "fast", "model": "ignored-model"}},
+    )
+
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        bot = Nanobot.from_config(config_path, workspace=tmp_path)
+
+    loop = bot._loop
+    assert loop.model == "gpt-4.1-mini"
+    assert loop.context_window_tokens == 128000
+    assert loop.provider.generation.temperature == 0.3
+    assert loop.provider.generation.max_tokens == 4096
+    assert loop.model_preset == "fast"
+
+
+@pytest.mark.asyncio
+async def test_preset_runtime_switch_updates_all_fields(tmp_path: Path) -> None:
+    config_path = _write_config(
+        tmp_path,
+        model_presets={
+            "cheap": {
+                "model": "gpt-4.1-mini",
+                "provider": "openai",
+                "max_tokens": 2048,
+                "context_window_tokens": 64000,
+                "temperature": 0.5,
+            },
+            "power": {
+                "model": "gpt-4.1",
+                "provider": "openai",
+                "max_tokens": 8192,
+                "context_window_tokens": 256000,
+                "temperature": 0.1,
+            },
+        },
+        agents={"defaults": {"model_preset": "cheap"}},
+    )
+
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        bot = Nanobot.from_config(config_path, workspace=tmp_path)
+
+    loop = bot._loop
+    assert loop.model == "gpt-4.1-mini"
+
+    my_tool = loop.tools.get("my")
+    result = await my_tool.execute(action="set", key="model_preset", value="power")
+    assert "Error" not in result
+
+    assert loop.model == "gpt-4.1"
+    assert loop.context_window_tokens == 256000
+    assert loop.provider.generation.temperature == 0.1
+    assert loop.provider.generation.max_tokens == 8192
+    assert loop.model_preset == "power"
+
+
+@pytest.mark.asyncio
+async def test_preset_switch_unknown_returns_error(tmp_path: Path) -> None:
+    config_path = _write_config(
+        tmp_path,
+        model_presets={"a": {"model": "model-a"}},
+        agents={"defaults": {"model_preset": "a"}},
+    )
+
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        bot = Nanobot.from_config(config_path, workspace=tmp_path)
+
+    loop = bot._loop
+    original_model = loop.model
+
+    my_tool = loop.tools.get("my")
+    result = await my_tool.execute(action="set", key="model_preset", value="nonexistent")
+    assert "not found" in result.lower()
+
+    assert loop.model == original_model
+    assert loop.model_preset == "a"
+
+
+@pytest.mark.asyncio
+async def test_preset_model_with_fallback_presets_in_config(tmp_path: Path) -> None:
+    config_path = _write_config(
+        tmp_path,
+        model_presets={
+            "prod": {
+                "model": "gpt-4.1",
+                "provider": "openai",
+                "max_tokens": 8192,
+                "temperature": 0.1,
+            },
+            "fallback": {
+                "model": "gpt-4.1-mini",
+                "provider": "openai",
+                "max_tokens": 4096,
+                "temperature": 0.2,
+            },
+        },
+        agents={
+            "defaults": {
+                "model_preset": "prod",
+                "fallback_presets": ["fallback"],
+            }
+        },
+    )
+
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        bot = Nanobot.from_config(config_path, workspace=tmp_path)
+
+    loop = bot._loop
+    assert loop.model == "gpt-4.1"
+    assert isinstance(loop.provider, ModelRouter)
+    assert loop.provider.fallback_presets == ["fallback"]
+
+
+@pytest.mark.asyncio
+async def test_fallback_presets_wired_to_all_subsystems(tmp_path: Path) -> None:
+    """When fallback_presets is configured, every subsystem that calls the LLM
+    must use the same ModelRouter instance, not the raw primary provider."""
+    config_path = _write_config(
+        tmp_path,
+        model_presets={
+            "prod": {
+                "model": "gpt-4.1",
+                "provider": "openai",
+                "max_tokens": 8192,
+                "temperature": 0.1,
+            },
+            "fallback": {
+                "model": "gpt-4.1-mini",
+                "provider": "openai",
+                "max_tokens": 4096,
+                "temperature": 0.2,
+            },
+        },
+        agents={
+            "defaults": {
+                "model_preset": "prod",
+                "fallback_presets": ["fallback"],
+            }
+        },
+    )
+
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        bot = Nanobot.from_config(config_path, workspace=tmp_path)
+
+    loop = bot._loop
+    router = loop.provider
+    assert isinstance(router, ModelRouter)
+
+    # Every LLM-consuming subsystem must share the same router
+    assert loop.runner.provider is router, "AgentRunner must use ModelRouter"
+    assert loop.subagents.provider is router, "SubagentManager must use ModelRouter"
+    assert loop.consolidator.provider is router, "Consolidator must use ModelRouter"
+    assert loop.dream.provider is router, "Dream must use ModelRouter"
+
+
+# ---------------------------------------------------------------------------
+# 2. Real HTTP Smoke Tests (aiohttp fake OpenAI server)
+# ---------------------------------------------------------------------------
+
+@pytest.mark.skipif(not HAS_AIOHTTP, reason="aiohttp not installed")
+@pytest.mark.asyncio
+async def test_preset_generation_params_reach_http_request() -> None:
+    """Provider.generation settings must appear in the actual HTTP request body."""
+    requests_log: list[dict] = []
+
+    async def handler(request: web.Request) -> web.Response:
+        body = await request.json()
+        requests_log.append(body)
+        return web.json_response({
+            "id": "chatcmpl-test",
+            "object": "chat.completion",
+            "model": body.get("model"),
+            "choices": [{
+                "index": 0,
+                "message": {"role": "assistant", "content": "pong"},
+                "finish_reason": "stop",
+            }],
+        })
+
+    app = web.Application()
+    app.router.add_post("/chat/completions", handler)
+    server = TestServer(app)
+    await server.start_server()
+    try:
+        base_url = str(server.make_url("/"))
+        provider = OpenAICompatProvider(
+            api_key="test",
+            api_base=base_url,
+            default_model="test-model",
+        )
+        provider.generation = GenerationSettings(temperature=0.42, max_tokens=1024)
+
+        with patch.object(LLMProvider, "_CHAT_RETRY_DELAYS", (0,)):
+            response = await provider.chat_with_retry(
+                messages=[{"role": "user", "content": "ping"}],
+            )
+
+        assert response.finish_reason != "error"
+        assert len(requests_log) >= 1
+        req = requests_log[0]
+        assert req["model"] == "test-model"
+        assert req["temperature"] == 0.42
+        assert req["max_tokens"] == 1024
+    finally:
+        await server.close()
+
+
+@pytest.mark.skipif(not HAS_AIOHTTP, reason="aiohttp not installed")
+@pytest.mark.asyncio
+async def test_failover_sends_second_request_to_fallback_model() -> None:
+    """Primary returns 503; after retry exhaustion ModelRouter hits fallback."""
+    requests_log: list[dict] = []
+
+    async def handler(request: web.Request) -> web.Response:
+        body = await request.json()
+        requests_log.append(body)
+        model = body.get("model")
+
+        if model == "primary-model":
+            return web.Response(
+                status=503,
+                body=json.dumps({"error": {"message": "overloaded", "type": "server_error"}}),
+                content_type="application/json",
+            )
+
+        return web.json_response({
+            "id": "chatcmpl-test",
+            "object": "chat.completion",
+            "model": model,
+            "choices": [{
+                "index": 0,
+                "message": {"role": "assistant", "content": "fallback-ok"},
+                "finish_reason": "stop",
+            }],
+        })
+
+    app = web.Application()
+    app.router.add_post("/chat/completions", handler)
+    server = TestServer(app)
+    await server.start_server()
+    try:
+        base_url = str(server.make_url("/"))
+        primary = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="primary-model"
+        )
+        fallback = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="fallback-model"
+        )
+
+        factory = MagicMock(return_value=fallback)
+
+        router = ModelRouter(
+            primary_provider=primary,
+            primary_model="primary-model",
+            fallback_presets=["fallback-model"],
+            provider_factory=factory,
+        )
+
+        with patch.object(LLMProvider, "_CHAT_RETRY_DELAYS", (0,)):
+            response = await router.chat_with_retry(
+                messages=[{"role": "user", "content": "hi"}],
+            )
+
+        assert response.finish_reason != "error"
+        assert response.content == "fallback-ok"
+
+        models_requested = [r["model"] for r in requests_log]
+        assert "primary-model" in models_requested
+        assert "fallback-model" in models_requested
+        factory.assert_called_once_with("fallback-model")
+    finally:
+        await server.close()
+
+
+@pytest.mark.skipif(not HAS_AIOHTTP, reason="aiohttp not installed")
+@pytest.mark.asyncio
+async def test_failover_on_quota_429() -> None:
+    """Quota 429 on one provider may still work on a different provider."""
+    requests_log: list[dict] = []
+
+    async def handler(request: web.Request) -> web.Response:
+        body = await request.json()
+        requests_log.append(body)
+        return web.Response(
+            status=429,
+            body=json.dumps({
+                "error": {
+                    "message": "insufficient quota",
+                    "type": "insufficient_quota",
+                    "code": "insufficient_quota",
+                }
+            }),
+            content_type="application/json",
+        )
+
+    app = web.Application()
+    app.router.add_post("/chat/completions", handler)
+    server = TestServer(app)
+    await server.start_server()
+    try:
+        base_url = str(server.make_url("/"))
+        primary = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="primary-model"
+        )
+        fallback = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="fallback-model"
+        )
+
+        factory = MagicMock(return_value=fallback)
+
+        router = ModelRouter(
+            primary_provider=primary,
+            primary_model="primary-model",
+            fallback_presets=["fallback-model"],
+            provider_factory=factory,
+        )
+
+        with patch.object(LLMProvider, "_CHAT_RETRY_DELAYS", (0,)):
+            response = await router.chat_with_retry(
+                messages=[{"role": "user", "content": "hi"}],
+            )
+
+        # Quota 429 SHOULD trigger failover — another provider may still work.
+        factory.assert_called_once_with("fallback-model")
+        assert response.finish_reason == "error"
+        # Both primary and fallback should have been requested.
+        assert len(requests_log) == 2
+    finally:
+        await server.close()
+
+
+@pytest.mark.skipif(not HAS_AIOHTTP, reason="aiohttp not installed")
+@pytest.mark.asyncio
+async def test_model_router_failover_integration() -> None:
+    """ModelRouter -> real HTTP failover chain (primary 503, fallback 200)."""
+    requests_log: list[dict] = []
+
+    async def handler(request: web.Request) -> web.Response:
+        body = await request.json()
+        requests_log.append(body)
+        model = body.get("model")
+
+        if model == "primary-model":
+            return web.Response(
+                status=503,
+                body=json.dumps({"error": {"message": "overloaded", "type": "server_error"}}),
+                content_type="application/json",
+            )
+
+        return web.json_response({
+            "id": "chatcmpl-test",
+            "object": "chat.completion",
+            "model": model,
+            "choices": [{
+                "index": 0,
+                "message": {"role": "assistant", "content": "fallback-ok"},
+                "finish_reason": "stop",
+            }],
+        })
+
+    app = web.Application()
+    app.router.add_post("/chat/completions", handler)
+    server = TestServer(app)
+    await server.start_server()
+    try:
+        base_url = str(server.make_url("/"))
+        primary = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="primary-model"
+        )
+        fallback = OpenAICompatProvider(
+            api_key="test", api_base=base_url, default_model="fallback-model"
+        )
+
+        factory = MagicMock(return_value=fallback)
+
+        router = ModelRouter(
+            primary_provider=primary,
+            primary_model="primary-model",
+            fallback_presets=["fallback-model"],
+            provider_factory=factory,
+        )
+
+        with patch.object(LLMProvider, "_CHAT_RETRY_DELAYS", (0,)):
+            response = await router.chat_with_retry(
+                messages=[{"role": "user", "content": "hello"}],
+            )
+
+        assert response.finish_reason != "error"
+        assert response.content == "fallback-ok"
+        models_requested = [r["model"] for r in requests_log]
+        assert "primary-model" in models_requested
+        assert "fallback-model" in models_requested
+        factory.assert_called_once_with("fallback-model")
+    finally:
+        await server.close()