chore: Bump version to 0.1.8

This release includes two critical fixes:

1. fix: Accept '*** Create File:' as alias for '*** Add File:' in patch parser
   - Claude sometimes uses 'Create File' syntax instead of 'Add File'
   - Parser now accepts both markers to prevent validation failures
   - Updated error message to include both valid syntaxes

2. fix: Increase default max_tokens from 8192 to 20480
   - Claude Sonnet 4.5 was getting cut off mid-task
   - New default is 5 * 4096 = 20480 tokens
   - Claude Sonnet 4.5 supports up to 64K tokens
   - Gives Claude enough space to complete comprehensive tasks

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

llmx-rs/Cargo.lock

@@ -178,7 +178,7 @@ checksum = "a23eb6b1614318a8071c9b2521f36b424b2c83db5eb3a0fead4a6c0809af6e61"
 
 [[package]]
 name = "app_test_support"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -765,9 +765,9 @@ dependencies = [
 
 [[package]]
 name = "clap"
-version = "4.5.51"
+version = "4.5.52"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4c26d721170e0295f191a69bd9a1f93efcdb0aff38684b61ab5750468972e5f5"
+checksum = "aa8120877db0e5c011242f96806ce3c94e0737ab8108532a76a3300a01db2ab8"
 dependencies = [
  "clap_builder",
  "clap_derive",
@@ -775,9 +775,9 @@ dependencies = [
 
 [[package]]
 name = "clap_builder"
-version = "4.5.51"
+version = "4.5.52"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75835f0c7bf681bfd05abe44e965760fea999a5286c6eb2d59883634fd02011a"
+checksum = "02576b399397b659c26064fbc92a75fede9d18ffd5f80ca1cd74ddab167016e1"
 dependencies = [
  "anstream",
  "anstyle",
@@ -952,7 +952,7 @@ checksum = "773648b94d0e5d620f64f280777445740e61fe701025087ec8b57f45c791888b"
 
 [[package]]
 name = "core_test_support"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -2828,7 +2828,7 @@ checksum = "6373607a59f0be73a39b6fe456b8192fcc3585f602af20751600e974dd455e77"
 
 [[package]]
 name = "llmx-ansi-escape"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "ansi-to-tui",
  "ratatui",
@@ -2837,7 +2837,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-app-server"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "app_test_support",
@@ -2872,7 +2872,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-app-server-protocol"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "clap",
@@ -2890,7 +2890,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-apply-patch"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -2905,7 +2905,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-arg0"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "dotenvy",
@@ -2918,7 +2918,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-async-utils"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "async-trait",
  "pretty_assertions",
@@ -2942,7 +2942,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-backend-openapi-models"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "serde",
  "serde_json",
@@ -2951,7 +2951,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-chatgpt"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "clap",
@@ -2966,7 +2966,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cli"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3006,7 +3006,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cloud-tasks"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -3032,7 +3032,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cloud-tasks-client"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -3047,7 +3047,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-common"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "clap",
  "llmx-app-server-protocol",
@@ -3059,7 +3059,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-core"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "askama",
@@ -3140,7 +3140,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-exec"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3173,7 +3173,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-execpolicy"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "allocative",
  "anyhow",
@@ -3193,7 +3193,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-feedback"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "llmx-protocol",
@@ -3204,7 +3204,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-file-search"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "clap",
@@ -3217,7 +3217,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-git"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "assert_matches",
  "once_cell",
@@ -3233,7 +3233,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-keyring-store"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "keyring",
  "tracing",
@@ -3241,7 +3241,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-linux-sandbox"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "clap",
  "landlock",
@@ -3254,7 +3254,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-login"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "base64",
@@ -3278,7 +3278,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-mcp-server"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3305,7 +3305,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-ollama"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "assert_matches",
  "async-stream",
@@ -3321,7 +3321,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-otel"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "chrono",
  "eventsource-stream",
@@ -3342,14 +3342,14 @@ dependencies = [
 
 [[package]]
 name = "llmx-process-hardening"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "libc",
 ]
 
 [[package]]
 name = "llmx-protocol"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "base64",
@@ -3375,7 +3375,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-responses-api-proxy"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "clap",
@@ -3391,7 +3391,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-rmcp-client"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "axum",
@@ -3420,7 +3420,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-stdio-to-uds"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3431,7 +3431,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-tui"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "arboard",
@@ -3496,7 +3496,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-cache"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "lru",
  "sha1",
@@ -3505,7 +3505,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-image"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "base64",
  "image",
@@ -3517,7 +3517,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-json-to-toml"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "pretty_assertions",
  "serde_json",
@@ -3526,7 +3526,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-pty"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "portable-pty",
@@ -3535,7 +3535,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-readiness"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "assert_matches",
  "async-trait",
@@ -3546,11 +3546,11 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-string"
-version = "0.1.7"
+version = "0.1.8"
 
 [[package]]
 name = "llmx-utils-tokenizer"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "pretty_assertions",
@@ -3660,7 +3660,7 @@ checksum = "47e1ffaa40ddd1f3ed91f717a33c8c0ee23fff369e3aa8772b9605cc1d22f4c3"
 
 [[package]]
 name = "mcp-types"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "schemars 0.8.22",
  "serde",
@@ -3670,7 +3670,7 @@ dependencies = [
 
 [[package]]
 name = "mcp_test_support"
-version = "0.1.7"
+version = "0.1.8"
 dependencies = [
  "anyhow",
  "assert_cmd",

llmx-rs/Cargo.toml

@@ -43,7 +43,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.1.7"
+version = "0.1.8"
 # Track the edition for all workspace crates in one place. Individual
 # crates can still override this value, but keeping it here means new
 # crates created with `cargo new -w ...` automatically inherit the 2024

llmx-rs/apply-patch/src/parser.rs

@@ -31,6 +31,7 @@ use thiserror::Error;
 const BEGIN_PATCH_MARKER: &str = "*** Begin Patch";
 const END_PATCH_MARKER: &str = "*** End Patch";
 const ADD_FILE_MARKER: &str = "*** Add File: ";
+const CREATE_FILE_MARKER: &str = "*** Create File: "; // Alias for Add File
 const DELETE_FILE_MARKER: &str = "*** Delete File: ";
 const UPDATE_FILE_MARKER: &str = "*** Update File: ";
 const MOVE_TO_MARKER: &str = "*** Move to: ";
@@ -245,8 +246,8 @@ fn check_start_and_end_lines_strict(
 fn parse_one_hunk(lines: &[&str], line_number: usize) -> Result<(Hunk, usize), ParseError> {
     // Be tolerant of case mismatches and extra padding around marker strings.
     let first_line = lines[0].trim();
-    if let Some(path) = first_line.strip_prefix(ADD_FILE_MARKER) {
-        // Add File
+    if let Some(path) = first_line.strip_prefix(ADD_FILE_MARKER).or_else(|| first_line.strip_prefix(CREATE_FILE_MARKER)) {
+        // Add File (also accepts Create File as alias)
         let mut contents = String::new();
         let mut parsed_lines = 1;
         for add_line in &lines[1..] {
@@ -331,7 +332,7 @@ fn parse_one_hunk(lines: &[&str], line_number: usize) -> Result<(Hunk, usize), P
 
     Err(InvalidHunkError {
         message: format!(
-            "'{first_line}' is not a valid hunk header. Valid hunk headers: '*** Add File: {{path}}', '*** Delete File: {{path}}', '*** Update File: {{path}}'"
+            "'{first_line}' is not a valid hunk header. Valid hunk headers: '*** Add File: {{path}}', '*** Create File: {{path}}', '*** Delete File: {{path}}', '*** Update File: {{path}}'"
         ),
         line_number,
     })

llmx-rs/core/src/chat_completions.rs

@@ -443,8 +443,8 @@ pub(crate) async fn stream_chat_completions(
     });
 
     // Add max_tokens - required by Anthropic Messages API
-    // Use provider config value or default to 8192
-    let max_tokens = provider.max_tokens.unwrap_or(8192);
+    // Use provider config value or default to 20480 (5 * 4096, Claude Sonnet 4.5 supports up to 64K)
+    let max_tokens = provider.max_tokens.unwrap_or(20480);
     if let Some(obj) = payload.as_object_mut() {
         obj.insert("max_tokens".to_string(), json!(max_tokens));
     }
@@ -610,7 +610,9 @@ async fn process_chat_sse<S>(
 ) where
     S: Stream<Item = Result<Bytes>> + Unpin,
 {
+    debug!("process_chat_sse started, idle_timeout={:?}", idle_timeout);
     let mut stream = stream.eventsource();
+    debug!("SSE stream initialized, waiting for first event");
 
     // State to accumulate a function call across streaming chunks.
     // OpenAI may split the `arguments` string over multiple `delta` events
@@ -645,7 +647,14 @@ async fn process_chat_sse<S>(
                 return;
             }
             Ok(None) => {
-                // Stream closed gracefully – emit Completed with dummy id.
+                // Stream closed gracefully – emit any pending items first, then Completed
+                debug!("Stream closed gracefully (Ok(None)), emitting pending items");
+                if let Some(item) = assistant_item.take() {
+                    let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+                }
+                if let Some(item) = reasoning_item.take() {
+                    let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+                }
                 let _ = tx_event
                     .send(Ok(ResponseEvent::Completed {
                         response_id: String::new(),
@@ -841,6 +850,7 @@ async fn process_chat_sse<S>(
 
             // Emit end-of-turn when finish_reason signals completion.
             if let Some(finish_reason) = choice.get("finish_reason").and_then(|v| v.as_str()) {
+                debug!("Received finish_reason: {}", finish_reason);
                 match finish_reason {
                     "tool_calls" if fn_call_state.active => {
                         // First, flush the terminal raw reasoning so UIs can finalize
@@ -859,27 +869,46 @@ async fn process_chat_sse<S>(
 
                         let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
                     }
-                    "stop" => {
-                        // Regular turn without tool-call. Emit the final assistant message
-                        // as a single OutputItemDone so non-delta consumers see the result.
+                    "stop" | "length" => {
+                        // Regular turn without tool-call, or hit max_tokens limit.
+                        debug!("Processing finish_reason={}, assistant_item.is_some()={}, reasoning_item.is_some()={}",
+                            finish_reason, assistant_item.is_some(), reasoning_item.is_some());
+                        // Emit the final assistant message as a single OutputItemDone so non-delta consumers see the result.
+                        if let Some(item) = assistant_item.take() {
+                            debug!("Emitting assistant_item: {:?}", item);
+                            let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+                        } else {
+                            debug!("No assistant_item to emit");
+                        }
+                        // Also emit a terminal Reasoning item so UIs can finalize raw reasoning.
+                        if let Some(item) = reasoning_item.take() {
+                            debug!("Emitting reasoning_item");
+                            let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+                        } else {
+                            debug!("No reasoning_item to emit");
+                        }
+                    }
+                    _ => {
+                        // Unknown finish_reason - still emit pending items to avoid hanging
+                        debug!("Unknown finish_reason: {}, emitting pending items", finish_reason);
                         if let Some(item) = assistant_item.take() {
                             let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
                         }
-                        // Also emit a terminal Reasoning item so UIs can finalize raw reasoning.
                         if let Some(item) = reasoning_item.take() {
                             let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
                         }
                     }
-                    _ => {}
                 }
 
                 // Emit Completed regardless of reason so the agent can advance.
+                debug!("Sending Completed event after finish_reason={}", finish_reason);
                 let _ = tx_event
                     .send(Ok(ResponseEvent::Completed {
                         response_id: String::new(),
                         token_usage: token_usage.clone(),
                     }))
                     .await;
+                debug!("Completed event sent, returning from SSE processor");
 
                 // Prepare for potential next turn (should not happen in same stream).
                 // fn_call_state = FunctionCallState::default();
@@ -888,6 +917,22 @@ async fn process_chat_sse<S>(
             }
         }
     }
+
+    // Stream ended without finish_reason - this can happen when the stream closes abruptly
+    debug!("Stream ended without finish_reason, emitting final items and Completed event");
+    if let Some(item) = assistant_item.take() {
+        let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+    }
+    if let Some(item) = reasoning_item.take() {
+        let _ = tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
+    }
+    // Send Completed event so llmx knows the turn is done
+    let _ = tx_event
+        .send(Ok(ResponseEvent::Completed {
+            response_id: String::new(),
+            token_usage: token_usage.clone(),
+        }))
+        .await;
 }
 
 /// Optional client-side aggregation helper

llmx-rs/core/src/event_mapping.rs

@@ -54,7 +54,7 @@ fn parse_user_message(message: &[ContentItem]) -> Option<UserMessageItem> {
     Some(UserMessageItem::new(&content))
 }
 
-fn parse_agent_message(id: Option<&String>, message: &[ContentItem]) -> AgentMessageItem {
+fn parse_agent_message(id: Option<&String>, message: &[ContentItem]) -> Option<AgentMessageItem> {
     let mut content: Vec<AgentMessageContent> = Vec::new();
     for content_item in message.iter() {
         match content_item {
@@ -69,18 +69,23 @@ fn parse_agent_message(id: Option<&String>, message: &[ContentItem]) -> AgentMes
             }
         }
     }
+
+    // If the message has no content, return None to signal turn completion
+    // This happens when the API ends a turn with an empty assistant message (e.g., after tool calls)
+    if content.is_empty() {
+        return None;
+    }
+
     let id = id.cloned().unwrap_or_else(|| Uuid::new_v4().to_string());
-    AgentMessageItem { id, content }
+    Some(AgentMessageItem { id, content })
 }
 
 pub fn parse_turn_item(item: &ResponseItem) -> Option<TurnItem> {
     match item {
         ResponseItem::Message { role, content, id } => match role.as_str() {
             "user" => parse_user_message(content).map(TurnItem::UserMessage),
-            "assistant" => Some(TurnItem::AgentMessage(parse_agent_message(
-                id.as_ref(),
-                content,
-            ))),
+            "assistant" => parse_agent_message(id.as_ref(), content)
+                .map(TurnItem::AgentMessage),
             "system" => None,
             _ => None,
         },

llmx-rs/core/src/model_provider_info.rs

@@ -294,7 +294,7 @@ pub fn built_in_model_providers() -> HashMap<String, ModelProviderInfo> {
                 request_max_retries: None,
                 stream_max_retries: None,
                 stream_idle_timeout_ms: None,
-                max_tokens: None,
+            max_tokens: None,
                 requires_openai_auth: false,
             },
         ),
@@ -335,7 +335,7 @@ pub fn built_in_model_providers() -> HashMap<String, ModelProviderInfo> {
                 request_max_retries: None,
                 stream_max_retries: None,
                 stream_idle_timeout_ms: None,
-                max_tokens: None,
+            max_tokens: None,
                 requires_openai_auth: true,
             },
         ),
@@ -381,7 +381,7 @@ pub fn create_oss_provider_with_base_url(base_url: &str) -> ModelProviderInfo {
         request_max_retries: None,
         stream_max_retries: None,
         stream_idle_timeout_ms: None,
-        max_tokens: None,
+            max_tokens: None,
         requires_openai_auth: false,
     }
 }
@@ -422,6 +422,7 @@ base_url = "http://localhost:11434/v1"
             request_max_retries: None,
             stream_max_retries: None,
             stream_idle_timeout_ms: None,
+            max_tokens: None,
             requires_openai_auth: false,
         };
 
@@ -452,6 +453,7 @@ query_params = { api-version = "2025-04-01-preview" }
             request_max_retries: None,
             stream_max_retries: None,
             stream_idle_timeout_ms: None,
+            max_tokens: None,
             requires_openai_auth: false,
         };
 
@@ -485,6 +487,7 @@ env_http_headers = { "X-Example-Env-Header" = "EXAMPLE_ENV_VAR" }
             request_max_retries: None,
             stream_max_retries: None,
             stream_idle_timeout_ms: None,
+            max_tokens: None,
             requires_openai_auth: false,
         };
 
@@ -508,6 +511,7 @@ env_http_headers = { "X-Example-Env-Header" = "EXAMPLE_ENV_VAR" }
                 request_max_retries: None,
                 stream_max_retries: None,
                 stream_idle_timeout_ms: None,
+            max_tokens: None,
                 requires_openai_auth: false,
             }
         }
@@ -541,6 +545,7 @@ env_http_headers = { "X-Example-Env-Header" = "EXAMPLE_ENV_VAR" }
             request_max_retries: None,
             stream_max_retries: None,
             stream_idle_timeout_ms: None,
+            max_tokens: None,
             requires_openai_auth: false,
         };
         assert!(named_provider.is_azure_responses_endpoint());