Keep excluding custom prompt/Agent Skills for Apple On-Device once tools ship (step 5)

Fifth piece of the Apple Intelligence tool-calling work (see
/Users/rune/.claude/plans/apple-intelligence-tool-calling-plan.md).
Once capabilities.tools flips to true for Apple On-Device (step 7,
not done yet), effectiveSystemPrompt's modelSupportsTools gate would
re-include the user's custom prompt and Agent Skills content — the
same two contributors that blew Phase 1's 4K context budget before
that gate existed (16,377/4,096 tokens on a trivial first message).
Measured the actual tool-usage guidance block instead of assuming it
also needed shortening: ~400 tokens, small enough to keep as-is, so no
separate condensed variant was needed — just keep excluding the two
genuinely large pieces specifically for Apple.

Dormant until step 7: modelSupportsTools is still false for Apple
today, so this has zero behavior change yet. No new tests —
effectiveSystemPrompt has no existing coverage (instance state, not
pure logic) and this change doesn't reduce what's already there.

oAI/ViewModels/ChatViewModel.swift

@@ -255,8 +255,17 @@ Don't narrate future actions ("Let me...") - just use the tools.
         // limit) — skip all three for tool-incapable models.
         let modelSupportsTools = selectedModel?.capabilities.tools ?? true
 
+        // Apple On-Device keeps a 4K context window even with tool calling enabled (unlike every
+        // other tools:true provider, which has a 100K+-token window). Phase 1 measured a trivial
+        // first message hitting 16,377/4,096 tokens before the exclusion below existed — dominated
+        // by the user's own custom prompt and Agent Skills content, not by tool-usage guidance
+        // itself (~400 tokens, measured — small enough to keep as-is). Keep excluding the two actual
+        // big contributors specifically for Apple, even now that modelSupportsTools is true for it.
+        let excludeVerboseCustomization = currentProvider == .appleOnDevice
+
         // Check if user wants to replace the default prompt entirely (BYOP mode)
         if modelSupportsTools,
+           !excludeVerboseCustomization,
            settings.customPromptMode == .replace,
            let customPrompt = settings.systemPrompt,
            !customPrompt.isEmpty {
@@ -282,6 +291,7 @@ Don't narrate future actions ("Let me...") - just use the tools.
 
         // Append custom prompt if in append mode and custom prompt exists
         if modelSupportsTools,
+           !excludeVerboseCustomization,
            settings.customPromptMode == .append,
            let customPrompt = settings.systemPrompt,
            !customPrompt.isEmpty {
@@ -289,7 +299,7 @@ Don't narrate future actions ("Let me...") - just use the tools.
         }
 
         // Append active agent skills (SKILL.md-style behavioral instructions)
-        if modelSupportsTools {
+        if modelSupportsTools && !excludeVerboseCustomization {
             let activeSkills = settings.agentSkills.filter { $0.isActive }
             if !activeSkills.isEmpty {
                 prompt += "\n\n---\n\n## Installed Skills\n\nThe following skills are active. Apply them when relevant:\n\n"