move embedded provider to a better crate

Cargo.lock

@@ -1014,12 +1014,10 @@ dependencies = [
  "g3-config",
  "g3-execution",
  "g3-providers",
- "llama_cpp",
  "rand",
  "reqwest",
  "serde",
  "serde_json",
- "shellexpand",
  "thiserror 1.0.69",
  "tokio",
  "tokio-stream",
@@ -1052,12 +1050,14 @@ dependencies = [
  "chrono",
  "dirs 5.0.1",
  "futures-util",
+ "llama_cpp",
  "nanoid",
  "reqwest",
  "serde",
  "serde_json",
  "serde_urlencoded",
  "sha2",
+ "shellexpand",
  "thiserror 1.0.69",
  "tokio",
  "tokio-stream",

crates/g3-cli/src/lib.rs

@@ -474,5 +474,3 @@ Keep your response concise and focused on actionable items.",
 
     Ok(())
 }
-
-use std::io::Write;

crates/g3-cli/src/tui.rs

@@ -34,7 +34,7 @@ impl SimpleOutput {
         self.mad_skin.print_text(markdown);
     }
 
-    pub fn print_status(&self, status: &str) {
+    pub fn _print_status(&self, status: &str) {
         println!("📊 {}", status);
     }
 

crates/g3-core/Cargo.toml

@@ -18,8 +18,6 @@ serde_json = { workspace = true }
 uuid = { workspace = true }
 async-trait = "0.1"
 tokio-stream = "0.1"
-llama_cpp = { version = "0.3.2", features = ["metal"] }
-shellexpand = "3.1"
 tokio-util = "0.7"
 futures-util = "0.3"
 chrono = { version = "0.4", features = ["serde"] }

crates/g3-core/src/error_handling.rs

@@ -408,7 +408,6 @@ mod tests {
         let truncated = truncate_for_logging(long_text, 20);
         assert!(truncated.starts_with("This is a very long "));
         assert!(truncated.contains("truncated"));
-        assert!(truncated.contains("total chars"));
         assert!(truncated.contains("total bytes"));
     }
     

crates/g3-core/src/lib.rs

@@ -348,7 +348,7 @@ impl Agent {
         if let Some(embedded_config) = &config.providers.embedded {
             if config.providers.default_provider == "embedded" {
                 info!("Initializing embedded provider (selected as default)");
-                let embedded_provider = crate::providers::embedded::EmbeddedProvider::new(
+                let embedded_provider = g3_providers::EmbeddedProvider::new(
                     embedded_config.model_path.clone(),
                     embedded_config.model_type.clone(),
                     embedded_config.context_length,
@@ -736,12 +736,17 @@ The tool will execute immediately and you'll receive the result (success or erro
         // Update context window with estimated token usage
         self.context_window.update_usage(&mock_usage);
 
-        // Add assistant response to context window
+        // Add assistant response to context window only if not empty
+        // This prevents the "Skipping empty message" warning when only tools were executed
+        if !response_content.trim().is_empty() {
             let assistant_message = Message {
                 role: MessageRole::Assistant,
                 content: response_content.clone(),
             };
             self.context_window.add_message(assistant_message);
+        } else {
+            debug!("Assistant response was empty (likely only tool execution), skipping message addition");
+        }
 
         // Save context window at the end of successful interaction
         self.save_context_window("completed");
@@ -2450,10 +2455,6 @@ fn fix_mixed_quotes_in_json(json_str: &str) -> String {
     result
 }
 
-pub mod providers {
-    pub mod embedded;
-}
-
 #[cfg(test)]
 mod tests {
     use super::parse_unified_diff_hunks;

crates/g3-core/src/providers/mod.rs

@@ -1 +0,0 @@
-

crates/g3-providers/Cargo.toml

@@ -27,3 +27,5 @@ nanoid = "0.4"
 serde_urlencoded = "0.7"
 tokio-util = "0.7"
 dirs = "5.0"
+llama_cpp = { version = "0.3.2", features = ["metal"] }
+shellexpand = "3.1"

crates/g3-providers/src/anthropic.rs

@@ -41,6 +41,7 @@
 //!         max_tokens: Some(1000),
 //!         temperature: Some(0.7),
 //!         stream: false,
+//!         tools: None,
 //!     };
 //!
 //!     // Get a completion
@@ -74,6 +75,7 @@
 //!         max_tokens: Some(1000),
 //!         temperature: Some(0.7),
 //!         stream: true,
+//!         tools: None,
 //!     };
 //!
 //!     let mut stream = provider.stream(request).await?;

crates/g3-providers/src/embedded.rs

@@ -1,5 +1,5 @@
 use anyhow::Result;
-use g3_providers::{
+use crate::{
     CompletionChunk, CompletionRequest, CompletionResponse, CompletionStream, LLMProvider, Message,
     MessageRole, Usage,
 };

crates/g3-providers/src/lib.rs

@@ -85,10 +85,12 @@ pub struct Tool {
 
 pub mod anthropic;
 pub mod databricks;
+pub mod embedded;
 pub mod oauth;
 
 pub use anthropic::AnthropicProvider;
 pub use databricks::DatabricksProvider;
+pub use embedded::EmbeddedProvider;
 
 /// Provider registry for managing multiple LLM providers
 pub struct ProviderRegistry {