From a04553013d81610f73b7fe3bdb98a0df53451942 Mon Sep 17 00:00:00 2001
From: Afonso Coutinho <afonso@omelhorsite.pt>
Date: Tue, 2 Jun 2026 16:57:20 +0100
Subject: [PATCH] fix: Anthropic responses with multiple text blocks lose all
 but the first (#1255)

* fix: concatenate all Anthropic text blocks, not just the first

* test: Anthropic response parsing concatenates text blocks
---
 src/llm_core.py                        | 17 +++++++++++-----
 tests/test_anthropic_response_parse.py | 27 ++++++++++++++++++++++++++
 2 files changed, 39 insertions(+), 5 deletions(-)
 create mode 100644 tests/test_anthropic_response_parse.py

diff --git a/src/llm_core.py b/src/llm_core.py
index 77afbd1..1e3d12a 100644
--- a/src/llm_core.py
+++ b/src/llm_core.py
@@ -562,11 +562,18 @@ def _build_anthropic_headers(headers):
     return h
 
 def _parse_anthropic_response(data: dict) -> str:
-    """Extract text from Anthropic response."""
-    for block in data.get("content", []):
-        if block.get("type") == "text":
-            return block.get("text", "")
-    return ""
+    """Extract text from an Anthropic response.
+
+    The Messages API `content` is an array that can hold more than one text
+    block (e.g. text split around a tool_use block, or citation-segmented
+    text). Concatenate them all instead of returning only the first, which
+    silently dropped the rest of the reply.
+    """
+    return "".join(
+        block.get("text", "")
+        for block in data.get("content", [])
+        if isinstance(block, dict) and block.get("type") == "text"
+    )
 
 
 def _sanitize_llm_messages(messages: List[Dict]) -> List[Dict]:
diff --git a/tests/test_anthropic_response_parse.py b/tests/test_anthropic_response_parse.py
new file mode 100644
index 0000000..e41c9bb
--- /dev/null
+++ b/tests/test_anthropic_response_parse.py
@@ -0,0 +1,27 @@
+"""Tests for _parse_anthropic_response (src/llm_core.py)."""
+
+from src.llm_core import _parse_anthropic_response
+
+
+def test_concatenates_multiple_text_blocks():
+    # Regression: only the first text block was returned, dropping the rest.
+    data = {"content": [
+        {"type": "text", "text": "Part A "},
+        {"type": "tool_use", "id": "t1", "name": "x", "input": {}},
+        {"type": "text", "text": "Part B"},
+    ]}
+    assert _parse_anthropic_response(data) == "Part A Part B"
+
+
+def test_skips_non_text_blocks():
+    data = {"content": [
+        {"type": "thinking", "thinking": "..."},
+        {"type": "text", "text": "answer"},
+    ]}
+    assert _parse_anthropic_response(data) == "answer"
+
+
+def test_single_block_and_empty():
+    assert _parse_anthropic_response({"content": [{"type": "text", "text": "hi"}]}) == "hi"
+    assert _parse_anthropic_response({"content": []}) == ""
+    assert _parse_anthropic_response({}) == ""