Fix output truncation bug, add language support for agent responses

Bug 1 — Output truncation:
  _run_claude() was replacing raw stdout with parsed sub-field which
  could be a dict (not string). run_agent() then saved dict.__repr__
  to DB instead of full JSON. Fixed: _run_claude() always returns
  string output; run_agent() ensures string before DB write.
  Added tests: full_output_saved_to_db, dict_output_saved_as_json_string.

Bug 2 — Language support:
  Added projects.language column (TEXT DEFAULT 'ru').
  Auto-migration for existing DBs (ALTER TABLE ADD COLUMN).
  context_builder passes language in project context.
  format_prompt() appends "## Language\nALWAYS respond in {language}"
  at the end of every prompt.
  CLI: kin project add --language ru (default: ru).
  Tests: language in prompt for ru/en, project creation, context.

112 tests, all passing. ~/.kin/kin.db migrated (vdol: language=ru).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

tests/test_runner.py

@@ -95,6 +95,48 @@ class TestRunAgent:
         assert len(logs) == 1
         assert logs[0]["project_id"] == "vdol"
 
+    @patch("agents.runner.subprocess.run")
+    def test_full_output_saved_to_db(self, mock_run, conn):
+        """Bug fix: output_summary must contain the FULL output, not truncated."""
+        long_json = json.dumps({
+            "result": json.dumps({
+                "summary": "Security audit complete",
+                "findings": [{"title": f"Finding {i}", "severity": "HIGH"} for i in range(50)],
+            }),
+        })
+        mock = MagicMock()
+        mock.stdout = long_json
+        mock.stderr = ""
+        mock.returncode = 0
+        mock_run.return_value = mock
+
+        run_agent(conn, "security", "VDOL-001", "vdol")
+
+        logs = conn.execute("SELECT output_summary FROM agent_logs WHERE agent_role='security'").fetchall()
+        assert len(logs) == 1
+        output = logs[0]["output_summary"]
+        assert output is not None
+        assert len(output) > 1000  # Must not be truncated
+        # Should contain all 50 findings
+        assert "Finding 49" in output
+
+    @patch("agents.runner.subprocess.run")
+    def test_dict_output_saved_as_json_string(self, mock_run, conn):
+        """When claude returns structured JSON, it must be saved as string."""
+        mock_run.return_value = _mock_claude_success({
+            "result": {"status": "ok", "files": ["a.py", "b.py"]},
+        })
+
+        result = run_agent(conn, "debugger", "VDOL-001", "vdol")
+
+        # output should be a string (JSON serialized), not a dict
+        assert isinstance(result["raw_output"], str)
+
+        logs = conn.execute("SELECT output_summary FROM agent_logs WHERE agent_role='debugger'").fetchall()
+        saved = logs[0]["output_summary"]
+        assert isinstance(saved, str)
+        assert "a.py" in saved
+
     @patch("agents.runner.subprocess.run")
     def test_previous_output_passed(self, mock_run, conn):
         mock_run.return_value = _mock_claude_success({"result": "tests pass"})