mcp-ted

There are multiple pairs of tools with overlapping purposes that will cause confusion. For example, 'browse_by_deadline' and 'browse_tenders_by_deadline' appear to do the same thing but target different data sources (general tenders vs TED tenders), and similarly 'find_similar_ted_tenders' vs 'find_similar_tenders', 'get_ted_stats' vs 'get_stats', and multiple search tools with 'ted' vs non-ted variants. While descriptions clarify the distinction, the naming doesn't make this clear at first glance, leading to misselection risk.

The naming follows a mostly consistent snake_case pattern, but there are significant inconsistencies in verb usage and structure. Some tools use 'browse_' prefix while others use 'search_', 'get_', or 'find_', without clear semantic distinction. Additionally, there are redundant tools with 'ted' vs non-ted variants that break consistency, and some tools have slightly different parameter names for similar concepts (e.g., 'notice_id' vs 'tender_id').

With 23 tools, this server feels overloaded for a procurement tender domain. Many tools appear to be redundant variations (e.g., separate tools for TED vs national tenders that could be unified with parameters). The count suggests poor API design with duplication rather than well-scoped functionality, making it harder for agents to navigate and increasing the risk of tool misselection.

The tool surface provides comprehensive coverage for searching, browsing, and analyzing procurement tenders across multiple dimensions (deadline, buyer, CPV, NUTS, value range, etc.). It supports both TED and national tenders with statistics and similarity search. The main gap is the lack of update/delete operations, but this is reasonable for a read-only data exploration server focused on public procurement data.