Redmine MCP Server

Instructions

搜尋議題 (在標題或描述中搜尋關鍵字)

Args:
    query: 搜尋關鍵字
    project_id: 限制在特定專案中搜尋 (可選)
    limit: 最大回傳數量 (預設 10，最大 50)

Returns:
    符合搜尋條件的議題列表

Implementation Reference

• src/redmine_mcp/server.py:446-515 (handler)

  The core implementation of the 'search_issues' MCP tool. This function is decorated with @mcp.tool(), handles the tool execution by querying Redmine issues via the API, performing local keyword filtering on title and description, and returning a formatted table of matching issues.

  @mcp.tool()
  def search_issues(query: str, project_id: int = None, limit: int = 10) -> str:
      """
      搜尋議題 (在標題或描述中搜尋關鍵字)
      
      Args:
          query: 搜尋關鍵字
          project_id: 限制在特定專案中搜尋 (可選)
          limit: 最大回傳數量 (預設 10，最大 50)
      
      Returns:
          符合搜尋條件的議題列表
      """
      try:
          if not query.strip():
              return "請提供搜尋關鍵字"
          
          client = get_client()
          limit = min(max(limit, 1), 50)
          
          # 設定搜尋參數
          params = {
              'limit': limit * 3,  # 取得更多結果以便篩選
              'sort': 'updated_on:desc'
          }
          
          if project_id:
              params['project_id'] = project_id
          
          # 取得議題列表
          all_issues = client.list_issues(**params)
          
          # 在本地端進行關鍵字篩選 (因為 Redmine API 沒有內建搜尋)
          query_lower = query.lower()
          matching_issues = []
          
          for issue in all_issues:
              if (query_lower in issue.subject.lower() or 
                  (issue.description and query_lower in issue.description.lower())):
                  matching_issues.append(issue)
                  if len(matching_issues) >= limit:
                      break
          
          if not matching_issues:
              search_scope = f"專案 {project_id}" if project_id else "所有可存取的專案"
              return f"在 {search_scope} 中沒有找到包含 '{query}' 的議題"
          
          # 格式化結果
          result = f"搜尋關鍵字: '{query}'\n"
          if project_id:
              result += f"搜尋範圍: 專案 {project_id}\n"
          result += f"找到 {len(matching_issues)} 個相關議題:\n\n"
          
          result += f"{'ID':<8} {'標題':<35} {'狀態':<12} {'專案':<15}\n"
          result += f"{'-'*8} {'-'*35} {'-'*12} {'-'*15}\n"
          
          for issue in matching_issues:
              title = issue.subject[:32] + "..." if len(issue.subject) > 35 else issue.subject
              status = issue.status.get('name', 'N/A')[:10]
              project_name = issue.project.get('name', 'N/A')[:13]
              
              result += f"{issue.id:<8} {title:<35} {status:<12} {project_name:<15}\n"
          
          return result
          
      except RedmineAPIError as e:
          return f"搜尋議題失敗: {str(e)}"
      except Exception as e:
          return f"系統錯誤: {str(e)}"

• src/redmine_mcp/server.py:446-446 (registration)

  The @mcp.tool() decorator registers the search_issues function as an MCP tool, automatically exposing it to the MCP server with the name derived from the function name.

  @mcp.tool()

• src/redmine_mcp/server.py:447-458 (schema)

  Input schema defined by type annotations (query: str required, project_id: int optional default None, limit: int optional default 10) and output str. Docstring provides detailed parameter descriptions.

  def search_issues(query: str, project_id: int = None, limit: int = 10) -> str:
      """
      搜尋議題 (在標題或描述中搜尋關鍵字)
      
      Args:
          query: 搜尋關鍵字
          project_id: 限制在特定專案中搜尋 (可選)
          limit: 最大回傳數量 (預設 10，最大 50)
      
      Returns:
          符合搜尋條件的議題列表
      """