LiuHe
Related Servers
Alternatives to LiuHe
No user-submitted related servers found.
Related Servers
- AlicenseNot gradedqualityBmaintenanceEnables coding agents to perform deterministic, model-free repository analysis and patch operations—tree-sitter repo maps, symbol/reference lookup, and SEARCH/REPLACE parsing/validation—via a CLI or MCP stdio server, without ever calling a language model.88 PyPI1MIT
- AlicenseNot gradedqualityAmaintenanceStandalone MCP server for code structure analysis using tree-sitter. Directory trees, symbol definitions, and call graphs without reading raw source files. Supports Rust, Python, Go, Java, TypeScript, Fortran, JavaScript, C/C++, and C#. Benchmarked up to 68% fewer tokens vs native tools.6Apache 2.0

mcp-reposkeinofficial
AlicenseAqualityAmaintenanceDeterministic code-graph (GraphRAG) over your repo for LLM agents — local-first, git-native, zero-infra, served via MCP. Python, TS/JS, Rust, Go, Java, C#.812Apache 2.0- AlicenseNot gradedqualityAmaintenanceEnables AI coding agents to query code structure efficiently through 16 MCP tools, including symbol lookup, full-text search, dependency analysis, and refactor planning, powered by tree-sitter parsing and index-backed code intelligence.3MIT
- AlicenseNot gradedqualityCmaintenanceToken-efficient code intelligence MCP server that indexes codebases with tree-sitter AST parsing and provides 150 tools for AI agents, using 61-95% fewer tokens than traditional grep/Read workflows.91 npm4Business Source 1.1
- AlicenseNot gradedqualityNot gradedmaintenanceCodeGraph — Open-source code intelligence MCP server. Builds a semantic graph of your codebase (functions, classes, imports, call chains) and exposes it through 31 tools. Callers, callees, impact analysis, complexity metrics, unused code detection, AI context assembly, persistent memory, cross-project search. 15 languages via tree-sitter. Single Rust binary, local-first.267 npm-
TDQS
Scored across 44 tools
Numerous tools have heavily overlapping responsibilities: multiple edit paths (write_symbol, write_symbols, edit_batch, edit_transaction, sandbox_validate, edit_collision_guard), overlapping search tools (symbol_search, code_search, references), and redundant inspection tools (inspect explicitly replaces read_outline/references/call_chain but they still exist). While descriptions attempt to differentiate them, the boundaries are often subtle, making tool selection error-prone.
The vast majority of tool names follow a clear verb_noun pattern in snake_case (e.g., find_tests, fix_imports, rename_symbol, read_symbol, write_symbol). A few outliers like 'gc', 'health', and 'inspect' break the pattern but are still intuitively named and do not introduce mixing of styles.
44 tools is far beyond the typical 3-15 well-scoped range and even exceeds the 25-tool threshold for 'too many'. The server attempts to cover an extremely broad domain, but the high count primarily reflects tool proliferation rather than genuinely distinct capabilities, overwhelming agents.
The toolset covers the full lifecycle of code development: indexing, searching, reading, editing, refactoring, testing, security review, quality analysis, and dependency management. Obvious gaps are few (e.g., no standalone git history/commit tool beyond git_worktree), and most workflows have a dedicated path. The redundancy slightly undermines usability but not coverage.