fix(runtime): SSE行缓冲 — 修复glm tool call参数截断丢失

根因: OpenAI driver的SSE解析直接按TCP chunk分行,
当glm的JSON响应被拆成多个TCP包时,SSE data行被截断,
导致tool call arguments丢失(input={})。

修复:
1. 添加pending_line缓冲区,跨chunk累积不完整的SSE行
2. 只处理完整的行(\n结尾),未完成的保留到下次
3. researcher.infer_action()增加更多字段推断(search/keyword/q等)

验证: 99 tests PASS, 160 hands tests PASS

crates/zclaw-hands/src/hands/researcher.rs

@@ -306,6 +306,33 @@ impl ResearcherHand {
     /// including the enum tag, e.g. sending `{"query": "search terms"}` instead
     /// of `{"action": "search", "query": "search terms"}`.
     fn infer_action(input: &Value) -> Result<ResearcherAction> {
+        // Debug: log all keys in the input
+        let keys: Vec<&str> = input.as_object()
+            .map(|obj| obj.keys().map(|k| k.as_str()).collect())
+            .unwrap_or_default();
+        tracing::warn!(target: "researcher", ?keys, %input, "infer_action examining input");
+
+        // Check for action field with wrong value
+        if let Some(action) = input.get("action").and_then(|v| v.as_str()) {
+            if action == "search" || action == "report" {
+                if let Some(query_val) = input.get("query") {
+                    let query = Self::parse_query(query_val);
+                    if !query.query.trim().is_empty() {
+                        return Ok(if action == "report" {
+                            ResearcherAction::Report { query }
+                        } else {
+                            ResearcherAction::Search { query }
+                        });
+                    }
+                }
+            }
+            if action == "fetch" {
+                if let Some(url) = input.get("url").and_then(|v| v.as_str()) {
+                    return Ok(ResearcherAction::Fetch { url: url.to_string() });
+                }
+            }
+        }
+
         // Has "url" (singular) → fetch
         if let Some(url) = input.get("url").and_then(|v| v.as_str()) {
             if !url.is_empty() && url.starts_with("http") {
@@ -321,10 +348,47 @@ impl ResearcherHand {
                 return Ok(ResearcherAction::Summarize { urls: url_list });
             }
         }
-        // Has "query" (string or object) → search
+        // Has "query" → search
         if let Some(query_val) = input.get("query") {
-            let query: ResearchQuery = if query_val.is_string() {
-                // LLM sent plain string: {"query": "search terms"}
+            let query = Self::parse_query(query_val);
+            if !query.query.trim().is_empty() {
+                return Ok(ResearcherAction::Search { query });
+            }
+        }
+        // Has "search" or "search_query" → search
+        for key in &["search", "search_query", "keyword", "keywords", "q", "text"] {
+            if let Some(val) = input.get(key) {
+                let query = Self::parse_query(val);
+                if !query.query.trim().is_empty() {
+                    return Ok(ResearcherAction::Search { query });
+                }
+            }
+        }
+        // Last resort: if any string field looks like a search query
+        if let Some(obj) = input.as_object() {
+            for (key, val) in obj {
+                if let Some(s) = val.as_str() {
+                    if s.len() > 2 && !s.starts_with("http") && key != "action" && key != "engine" {
+                        tracing::warn!(target: "researcher", key = %key, value = %s, "Using fallback field as query");
+                        return Ok(ResearcherAction::Search { query: ResearchQuery {
+                            query: s.to_string(),
+                            engine: SearchEngine::Auto,
+                            depth: ResearchDepth::Standard,
+                            max_results: 10,
+                            include_related: false,
+                            time_limit_secs: 60,
+                        }});
+                    }
+                }
+            }
+        }
+        Err(zclaw_types::ZclawError::HandError(
+            "无法识别搜索意图：请提供 query（搜索）或 url（获取网页）参数".to_string()
+        ))
+    }
+
+    fn parse_query(query_val: &Value) -> ResearchQuery {
+        if query_val.is_string() {
             ResearchQuery {
                 query: query_val.as_str().unwrap_or("").to_string(),
                 engine: SearchEngine::Auto,
@@ -334,9 +398,11 @@ impl ResearcherHand {
                 time_limit_secs: 60,
             }
         } else {
-                // LLM sent object: {"query": {"query": "...", "engine": "..."}}
             serde_json::from_value(query_val.clone()).unwrap_or_else(|_| ResearchQuery {
                 query: query_val.get("query")
+                    .or_else(|| query_val.get("search"))
+                    .or_else(|| query_val.get("q"))
+                    .or_else(|| query_val.get("keyword"))
                     .and_then(|v| v.as_str())
                     .unwrap_or("")
                     .to_string(),
@@ -346,15 +412,8 @@ impl ResearcherHand {
                 include_related: false,
                 time_limit_secs: 60,
             })
-            };
-            if !query.query.trim().is_empty() {
-                return Ok(ResearcherAction::Search { query });
         }
     }
-        Err(zclaw_types::ZclawError::HandError(
-            "无法识别搜索意图：请提供 query（搜索）或 url（获取网页）参数".to_string()
-        ))
-    }
 
     /// Execute a web search — route to the configured backend
     async fn execute_search(&self, query: &ResearchQuery) -> Result<Vec<SearchResult>> {
@@ -1085,10 +1144,14 @@ impl Hand for ResearcherHand {
     }
 
     async fn execute(&self, _context: &HandContext, input: Value) -> Result<HandResult> {
+        tracing::info!(target: "researcher", input = %input, "Researcher hand received input");
         // Try strict deserialization first, then fall back to inference
         let action: ResearcherAction = match serde_json::from_value(input.clone()) {
             Ok(a) => a,
-            Err(_) => Self::infer_action(&input)?,
+            Err(e) => {
+                tracing::warn!(target: "researcher", error = %e, input = %input, "Strict deserialization failed, trying inference");
+                Self::infer_action(&input)?
+            }
         };
 
         let start = std::time::Instant::now();

crates/zclaw-runtime/src/driver/openai.rs

@@ -163,6 +163,7 @@ impl LlmDriver for OpenAiDriver {
             let mut current_tool_id: Option<String> = None;
             let mut sse_event_count: usize = 0;
             let mut raw_bytes_total: usize = 0;
+            let mut pending_line = String::new(); // Buffer for incomplete SSE lines
 
             while let Some(chunk_result) = byte_stream.next().await {
                 let chunk = match chunk_result {
@@ -180,13 +181,21 @@ impl LlmDriver for OpenAiDriver {
                 if raw_bytes_total <= 600 {
                     tracing::debug!("[OpenAI:stream] RAW chunk ({} bytes): {:?}", text.len(), &text[..text.len().min(500)]);
                 }
-                for line in text.lines() {
+                // Accumulate text and split by lines, handling incomplete last line
+                pending_line.push_str(&text);
+                // Extract complete lines (ending with \n), keep the rest pending
+                let mut complete_lines: Vec<String> = Vec::new();
+                while let Some(pos) = pending_line.find('\n') {
+                    complete_lines.push(pending_line[..pos].to_string());
+                    pending_line = pending_line[pos + 1..].to_string();
+                }
+                for line in complete_lines {
                     let trimmed = line.trim();
                     if trimmed.is_empty() || trimmed.starts_with(':') {
                         continue; // Skip empty lines and SSE comments
                     }
                     // Handle both "data: " (standard) and "data:" (no space)
-                    let data = if let Some(d) = trimmed.strip_prefix("data: ") {
+                    let data: Option<&str> = if let Some(d) = trimmed.strip_prefix("data: ") {
                         Some(d)
                     } else if let Some(d) = trimmed.strip_prefix("data:") {
                         Some(d.trim_start())