From 5e2ef821b56c78087d9c4b6d1a94f5e1cc34d8df Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Sat, 18 Jul 2026 18:35:43 -0700
Subject: [PATCH] fix(skill_engine): parse LLM JSON despite trailing brace
 prose

Greedy \{.*\} matching spanned into later prose braces, so json.loads
failed and analysis/capture dropped valid objects. Use raw_decode from
the first brace, matching agents.base.
---
 .gitignore                                       |  1 +
 openspace/skill_engine/analyzer.py               | 10 +++++-----
 .../skill_engine/evolution/capture_semantic.py   | 10 ++++++----
 .../test_extract_json_trailing_prose.py          | 16 ++++++++++++++++
 4 files changed, 28 insertions(+), 9 deletions(-)
 create mode 100644 tests/skill_engine/test_extract_json_trailing_prose.py

diff --git a/.gitignore b/.gitignore
index 27e8dfb0..73b84e09 100644
--- a/.gitignore
+++ b/.gitignore
@@ -51,6 +51,7 @@ tests/skill_engine/*
 !tests/skill_engine/test_evolver_length_recovery.py
 !tests/skill_engine/test_evolution_retry_idempotency.py
 !tests/skill_engine/test_analyzer_length_recovery.py
+!tests/skill_engine/test_extract_json_trailing_prose.py
 !tests/skill_engine/decision/
 tests/skill_engine/decision/*
 !tests/skill_engine/decision/test_analysis_adapter.py
diff --git a/openspace/skill_engine/analyzer.py b/openspace/skill_engine/analyzer.py
index 2d794b44..ada3c680 100644
--- a/openspace/skill_engine/analyzer.py
+++ b/openspace/skill_engine/analyzer.py
@@ -1508,13 +1508,13 @@ def _extract_json(text: str) -> Optional[Dict[str, Any]]:
         if code_match:
             text = code_match.group(1).strip()
         else:
-            # Try bare JSON object
-            json_match = re.search(r"\{.*\}", text, re.DOTALL)
-            if json_match:
-                text = json_match.group()
+            # Prefer the first object start; raw_decode ignores trailing prose.
+            brace = text.find("{")
+            if brace >= 0:
+                text = text[brace:]
 
         try:
-            data = json.loads(text)
+            data, _ = json.JSONDecoder().raw_decode(text)
             if isinstance(data, dict):
                 return data
             logger.warning(f"LLM returned non-dict JSON: {type(data)}")
diff --git a/openspace/skill_engine/evolution/capture_semantic.py b/openspace/skill_engine/evolution/capture_semantic.py
index a1ceb62a..aff0107f 100644
--- a/openspace/skill_engine/evolution/capture_semantic.py
+++ b/openspace/skill_engine/evolution/capture_semantic.py
@@ -399,11 +399,13 @@ def _extract_json(text: str) -> dict[str, Any] | None:
     if match:
         text = match.group(1)
     else:
-        match = re.search(r"\{.*\}", text, re.DOTALL)
-        if match:
-            text = match.group(0)
+        brace = text.find("{")
+        if brace >= 0:
+            text = text[brace:]
+        else:
+            return None
     try:
-        value = json.loads(text)
+        value, _ = json.JSONDecoder().raw_decode(text)
     except Exception:
         return None
     return dict(value) if isinstance(value, Mapping) else None
diff --git a/tests/skill_engine/test_extract_json_trailing_prose.py b/tests/skill_engine/test_extract_json_trailing_prose.py
new file mode 100644
index 00000000..f4ca5069
--- /dev/null
+++ b/tests/skill_engine/test_extract_json_trailing_prose.py
@@ -0,0 +1,16 @@
+"""Regression: trailing prose with braces must not drop valid analysis JSON."""
+
+from __future__ import annotations
+
+from openspace.skill_engine.analyzer import ExecutionAnalyzer
+from openspace.skill_engine.evolution.capture_semantic import _extract_json as capture_extract_json
+
+
+def test_analyzer_extract_json_tolerates_trailing_prose_braces():
+    text = '{"task_completed": true}\n\nReplace {foo} with {bar}.'
+    assert ExecutionAnalyzer._extract_json(text) == {"task_completed": True}
+
+
+def test_capture_semantic_extract_json_tolerates_trailing_prose_braces():
+    text = '{"ok": true}\n\nnote {x}'
+    assert capture_extract_json(text) == {"ok": True}