Enhance skill generator documentation and templates

- Updated Phase 1 and Phase 2 documentation to include next phase links and data flow details.
- Expanded Phase 5 documentation to include comprehensive validation and README generation steps, along with validation report structure.
- Added purpose and usage context sections to various action and script templates (e.g., autonomous-action, llm-action, script-bash).
- Improved commands management by simplifying the command scanning logic and enabling/disabling commands through renaming files.
- Enhanced dashboard command manager to format group names and display nested groups with appropriate icons and colors.
- Updated LiteLLM executor to allow model overrides during execution.
- Added action reference guide and template reference sections to the skill-tuning SKILL.md for better navigation and understanding.

ccw/src/tools/claude-cli-tools.ts

@@ -802,13 +802,15 @@ export function getNativeResume(projectDir: string): boolean {
  */
 export function addClaudeApiEndpoint(
   projectDir: string,
-  endpoint: { id: string; name: string; enabled: boolean }
+  endpoint: { id: string; name: string; enabled: boolean; model?: string }
 ): ClaudeCliToolsConfig {
   const config = loadClaudeCliTools(projectDir);
 
   // Add as a tool with type: 'api-endpoint'
   config.tools[endpoint.name] = {
     enabled: endpoint.enabled,
+    primaryModel: endpoint.model,  // Use endpoint.model as primaryModel (can be overridden via --model)
+    secondaryModel: endpoint.model,  // Same as primary for fallback
     tags: [],
     type: 'api-endpoint',
     id: endpoint.id  // Store endpoint ID for settings lookup

ccw/src/tools/cli-executor-core.ts

@@ -557,6 +557,11 @@ async function executeCliTool(
       // id field is the LiteLLM endpoint ID (e.g., "g25")
       const litellmEndpointId = toolConfig.id || toolName;
 
+      // Use configured primary model if no explicit model provided
+      // This allows --model parameter to override the tool's primaryModel
+      // Use undefined if primaryModel is empty string (endpoint.model will be used as fallback)
+      const apiEndpointEffectiveModel = model || (toolConfig.primaryModel || undefined);
+
       // Find LiteLLM endpoint configuration
       const litellmEndpoint = findEndpointById(workingDir, litellmEndpointId);
       if (litellmEndpoint) {
@@ -568,13 +573,14 @@ async function executeCliTool(
           });
         }
 
-        // Execute via LiteLLM
+        // Execute via LiteLLM with model override
         const result = await executeLiteLLMEndpoint({
           prompt,
           endpointId: litellmEndpointId,
           baseDir: workingDir,
           cwd: cd || workingDir,
           includeDirs: includeDirs ? includeDirs.split(',').map(d => d.trim()) : undefined,
+          model: apiEndpointEffectiveModel, // Pass effective model (--model or primaryModel)
           onOutput: onOutput || undefined,
         });
 
@@ -587,7 +593,7 @@ async function executeCliTool(
           id: customId || `${Date.now()}-litellm`,
           timestamp: new Date(startTime).toISOString(),
           tool: toolName,
-          model: litellmEndpoint.model,
+          model: result.model, // Use effective model from result (reflects any override)
           mode,
           prompt,
           status: result.success ? 'success' : 'error',

ccw/src/tools/litellm-executor.ts

@@ -19,6 +19,7 @@ export interface LiteLLMExecutionOptions {
   cwd?: string; // Working directory for file resolution
   includeDirs?: string[]; // Additional directories for @patterns
   enableCache?: boolean; // Override endpoint cache setting
+  model?: string; // Override model for this execution (if not specified, uses endpoint.model)
   onOutput?: (unit: CliOutputUnit) => void;
   /** Number of retries after the initial attempt (default: 0) */
   maxRetries?: number;
@@ -56,7 +57,7 @@ export function extractPatterns(prompt: string): string[] {
 export async function executeLiteLLMEndpoint(
   options: LiteLLMExecutionOptions
 ): Promise<LiteLLMExecutionResult> {
-  const { prompt, endpointId, baseDir, cwd, includeDirs, enableCache, onOutput } = options;
+  const { prompt, endpointId, baseDir, cwd, includeDirs, enableCache, model: modelOverride, onOutput } = options;
 
   // 1. Find endpoint configuration
   const endpoint = findEndpointById(baseDir, endpointId);
@@ -96,7 +97,10 @@ export async function executeLiteLLMEndpoint(
     };
   }
 
-  // 3. Process context cache if enabled
+  // 3. Determine effective model: use override if provided, otherwise use endpoint.model
+  const effectiveModel = modelOverride || endpoint.model;
+
+  // 4. Process context cache if enabled
   let finalPrompt = prompt;
   let cacheUsed = false;
   let cachedFiles: string[] = [];
@@ -168,12 +172,12 @@ export async function executeLiteLLMEndpoint(
     }
   }
 
-  // 4. Call LiteLLM
+  // 5. Call LiteLLM
   try {
     if (onOutput) {
       onOutput({
         type: 'stderr',
-        content: `[LiteLLM: Calling ${provider.type}/${endpoint.model}]\n`,
+        content: `[LiteLLM: Calling ${provider.type}/${effectiveModel}]\n`,
         timestamp: new Date().toISOString()
       });
     }
@@ -206,14 +210,14 @@ export async function executeLiteLLMEndpoint(
       delete process.env['CCW_LITELLM_EXTRA_HEADERS'];
     }
 
-    // Use litellm-client to call chat
+    // Use litellm-client to call chat with effective model
     const response = await callWithRetries(
-      () => client.chat(finalPrompt, endpoint.model),
+      () => client.chat(finalPrompt, effectiveModel),
       {
         maxRetries: options.maxRetries ?? 0,
         baseDelayMs: options.retryBaseDelayMs ?? 1000,
         onOutput,
-        rateLimitKey: `${provider.type}:${endpoint.model}`,
+        rateLimitKey: `${provider.type}:${effectiveModel}`,
       },
     );
 
@@ -228,7 +232,7 @@ export async function executeLiteLLMEndpoint(
     return {
       success: true,
       output: response,
-      model: endpoint.model,
+      model: effectiveModel,
       provider: provider.type,
       cacheUsed,
       cachedFiles,
@@ -246,7 +250,7 @@ export async function executeLiteLLMEndpoint(
     return {
       success: false,
       output: '',
-      model: endpoint.model,
+      model: effectiveModel,
       provider: provider.type,
       cacheUsed,
       error: errorMsg,