feat: add model management tools to add-ollama-tool skill

Adds four new MCP tools to the existing ollama integration, consolidating
model management (from #1331) into the single add-ollama-tool skill as
requested by @gavrielc:

- ollama_pull_model  — pull a model from the Ollama registry
- ollama_delete_model — delete a local model to free disk space
- ollama_show_model  — inspect modelfile, parameters, and architecture
- ollama_list_running — list models loaded in memory with VRAM/processor info

All four tools follow the existing patterns in this file: OLLAMA_HOST env
var, ollamaFetch() with host.docker.internal fallback, log() and
writeStatus() helpers. No changes to index.ts or container-runner.ts
needed — OLLAMA_HOST is already forwarded via sdkEnv.

Also updates SKILL.md description, tool list, verify steps, and adds a
troubleshooting entry for large-model pull timeouts.

Closes #1331.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.claude/skills/add-ollama-tool/SKILL.md

@@ -1,15 +1,19 @@
 ---
 name: add-ollama-tool
-description: Add Ollama MCP server so the container agent can call local models for cheaper/faster tasks like summarization, translation, or general queries.
+description: Add Ollama MCP server so the container agent can call local models and manage the Ollama model library.
 ---
 
 # Add Ollama Integration
 
-This skill adds a stdio-based MCP server that exposes local Ollama models as tools for the container agent. Claude remains the orchestrator but can offload work to local models.
+This skill adds a stdio-based MCP server that exposes local Ollama models as tools for the container agent. Claude remains the orchestrator but can offload work to local models, and can also manage the model library directly.
 
 Tools added:
-- `ollama_list_models` — lists installed Ollama models
-- `ollama_generate` — sends a prompt to a specified model and returns the response
+- `ollama_list_models` — list installed models with name, size, family, and last modified date
+- `ollama_generate` — send a prompt to a specified model and return the response
+- `ollama_pull_model` — pull (download) a model from the Ollama registry by name
+- `ollama_delete_model` — delete a locally installed model to free disk space
+- `ollama_show_model` — show model details: modelfile, parameters, template, and architecture info
+- `ollama_list_running` — list models currently loaded in memory with memory usage and processor type
 
 ## Phase 1: Pre-flight
 
@@ -106,7 +110,7 @@ launchctl kickstart -k gui/$(id -u)/com.nanoclaw  # macOS
 
 ## Phase 4: Verify
 
-### Test via WhatsApp
+### Test inference
 
 Tell the user:
 
@@ -114,6 +118,12 @@ Tell the user:
 >
 > The agent should use `ollama_list_models` to find available models, then `ollama_generate` to get a response.
 
+### Test model management
+
+> Send a message like: "pull the gemma3:1b model" or "which ollama models are currently loaded in memory?"
+>
+> The agent should call `ollama_pull_model` or `ollama_list_running` respectively.
+
 ### Monitor activity (optional)
 
 Run the watcher script for macOS notifications when Ollama is used:
@@ -129,9 +139,10 @@ tail -f logs/nanoclaw.log | grep -i ollama
 ```
 
 Look for:
-- `Agent output: ... Ollama ...` — agent used Ollama successfully
-- `[OLLAMA] >>> Generating` — generation started (if log surfacing works)
+- `[OLLAMA] >>> Generating` — generation started
 - `[OLLAMA] <<< Done` — generation completed
+- `[OLLAMA] Pulling model:` — pull in progress
+- `[OLLAMA] Deleted:` — model removed
 
 ## Troubleshooting
 
@@ -151,3 +162,7 @@ The agent is trying to run `ollama` CLI inside the container instead of using th
 ### Agent doesn't use Ollama tools
 
 The agent may not know about the tools. Try being explicit: "use the ollama_generate tool with gemma3:1b to answer: ..."
+
+### `ollama_pull_model` times out on large models
+
+Large models (7B+) can take several minutes. The tool uses `stream: false` so it blocks until complete — this is intentional. For very large pulls, use the host CLI directly: `ollama pull <model>`

container/agent-runner/src/ollama-mcp-stdio.ts

@@ -143,5 +143,139 @@ server.tool(
   },
 );
 
+server.tool(
+  'ollama_pull_model',
+  'Pull (download) a model from the Ollama registry by name. Returns the final status once the pull is complete. Use model names like "llama3.2", "mistral", "gemma2:9b".',
+  {
+    model: z.string().describe('Model name to pull, e.g. "llama3.2", "mistral", "gemma2:9b"'),
+  },
+  async (args) => {
+    log(`Pulling model: ${args.model}...`);
+    writeStatus('pulling', `Pulling ${args.model}`);
+    try {
+      const res = await ollamaFetch('/api/pull', {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ model: args.model, stream: false }),
+      });
+      if (!res.ok) {
+        const errorText = await res.text();
+        return {
+          content: [{ type: 'text' as const, text: `Ollama error (${res.status}): ${errorText}` }],
+          isError: true,
+        };
+      }
+      const data = await res.json() as { status: string };
+      log(`Pull complete: ${args.model} — ${data.status}`);
+      writeStatus('done', `Pulled ${args.model}`);
+      return { content: [{ type: 'text' as const, text: `Pull complete: ${args.model} — ${data.status}` }] };
+    } catch (err) {
+      return {
+        content: [{ type: 'text' as const, text: `Failed to pull model: ${err instanceof Error ? err.message : String(err)}` }],
+        isError: true,
+      };
+    }
+  },
+);
+
+server.tool(
+  'ollama_delete_model',
+  'Delete a locally installed Ollama model to free up disk space.',
+  {
+    model: z.string().describe('Model name to delete, e.g. "llama3.2", "mistral:latest"'),
+  },
+  async (args) => {
+    log(`Deleting model: ${args.model}...`);
+    writeStatus('deleting', `Deleting ${args.model}`);
+    try {
+      const res = await ollamaFetch('/api/delete', {
+        method: 'DELETE',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ model: args.model }),
+      });
+      if (!res.ok) {
+        const errorText = await res.text();
+        return {
+          content: [{ type: 'text' as const, text: `Ollama error (${res.status}): ${errorText}` }],
+          isError: true,
+        };
+      }
+      log(`Deleted: ${args.model}`);
+      writeStatus('done', `Deleted ${args.model}`);
+      return { content: [{ type: 'text' as const, text: `Deleted model: ${args.model}` }] };
+    } catch (err) {
+      return {
+        content: [{ type: 'text' as const, text: `Failed to delete model: ${err instanceof Error ? err.message : String(err)}` }],
+        isError: true,
+      };
+    }
+  },
+);
+
+server.tool(
+  'ollama_show_model',
+  'Show details for a locally installed Ollama model: modelfile, parameters, template, system prompt, and architecture info (context length, parameter count, etc.).',
+  {
+    model: z.string().describe('Model name to inspect, e.g. "llama3.2", "mistral:latest"'),
+  },
+  async (args) => {
+    log(`Showing model info: ${args.model}...`);
+    try {
+      const res = await ollamaFetch('/api/show', {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ model: args.model }),
+      });
+      if (!res.ok) {
+        const errorText = await res.text();
+        return {
+          content: [{ type: 'text' as const, text: `Ollama error (${res.status}): ${errorText}` }],
+          isError: true,
+        };
+      }
+      const data = await res.json();
+      return { content: [{ type: 'text' as const, text: JSON.stringify(data, null, 2) }] };
+    } catch (err) {
+      return {
+        content: [{ type: 'text' as const, text: `Failed to show model info: ${err instanceof Error ? err.message : String(err)}` }],
+        isError: true,
+      };
+    }
+  },
+);
+
+server.tool(
+  'ollama_list_running',
+  'List Ollama models currently loaded in memory with their memory usage, processor type (CPU/GPU), and time until they are unloaded.',
+  {},
+  async () => {
+    log('Listing running models...');
+    try {
+      const res = await ollamaFetch('/api/ps');
+      if (!res.ok) {
+        return {
+          content: [{ type: 'text' as const, text: `Ollama API error: ${res.status} ${res.statusText}` }],
+          isError: true,
+        };
+      }
+      const data = await res.json() as { models?: Array<{ name: string; size_vram: number; processor: string; expires_at: string }> };
+      const models = data.models || [];
+      if (models.length === 0) {
+        return { content: [{ type: 'text' as const, text: 'No models currently loaded in memory.' }] };
+      }
+      const list = models
+        .map(m => `- ${m.name} (${(m.size_vram / 1e9).toFixed(1)}GB ${m.processor}, unloads at ${m.expires_at})`)
+        .join('\n');
+      log(`${models.length} model(s) running`);
+      return { content: [{ type: 'text' as const, text: `Models loaded in memory:\n${list}` }] };
+    } catch (err) {
+      return {
+        content: [{ type: 'text' as const, text: `Failed to list running models: ${err instanceof Error ? err.message : String(err)}` }],
+        isError: true,
+      };
+    }
+  },
+);
+
 const transport = new StdioServerTransport();
 await server.connect(transport);