MCP Context Manager
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| index_repositoryA | 🔧 REQUIRED FIRST STEP: Index or re-index the repository to enable all context-manager tools. Uses cached index if files haven't changed. ALWAYS call this when starting work on a repository or if files have changed significantly. Fast (<2s for most repos). |
| get_relevant_contextB | ⭐ PREFERRED SEARCH: Use this INSTEAD OF Grep for finding code by description. Natural language search with BM25 ranking. Saves 70-90% tokens vs reading files directly. Examples: "authentication middleware", "payment processing", "gacha spin mechanics". Returns only relevant code snippets with relevance scores. |
| find_symbolA | ⭐ PREFERRED FOR SYMBOLS: Use this INSTEAD OF Grep when looking for specific functions, classes, or variables. Fuzzy matching automatically handles typos ("athenticate" → "authenticate"). Returns exact file path and line number. Much faster than Grep with better accuracy. |
| get_functionA | ⭐ PREFERRED OVER Read: Get complete function code without reading the entire file. Saves 85% tokens compared to Read. Use when you need a specific function implementation instead of reading full files. Returns only the function definition with signature. |
| get_classA | ⭐ PREFERRED OVER Read: Get class definition without reading the entire file. Optionally filter specific methods. Saves 80%+ tokens vs reading full files. Use this when you need class structure or specific class methods. |
| get_relevant_contextB | Get code context relevant to a natural language query. Returns minimal, targeted code snippets. |
| get_file_summaryA | ⭐ PREFERRED FOR FILE OVERVIEW: Get file structure (exports, functions, classes, imports) without reading full content. Use this INSTEAD OF Read when you need to understand what's in a file without seeing implementation details. Saves 90% tokens. |
| search_codeA | ⭐ PREFERRED OVER Grep: Search for code patterns with regex support. Returns ranked results with minimal context. Better than Grep because it ranks by relevance and provides AI-optimized output. Use for pattern matching and text search. |
| get_repository_structureA | ⭐ PREFERRED OVER ls/tree: Get clean repository structure showing directories and file types. Use this INSTEAD OF running ls, tree, or Glob for understanding project layout. No file contents, just structure. |
| get_dependenciesB | Find all dependencies (imports/requires) for a file or symbol. Useful for understanding what code needs. |
| clear_cacheA | Clear the cached index for a repository. Useful if cache becomes corrupted or you want to force a fresh index. |
| find_similarB | Find code similar to a given symbol. Useful for discovering related implementations, similar patterns, or alternative approaches. |
| get_usage_statsA | View token usage statistics for this session. Shows how many tokens each MCP tool used vs what full file reads would have cost. Use this to verify token savings. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 12 tools
Each tool has a clearly distinct purpose with no overlap: clear_cache manages indexing, index_repository builds the index, find_symbol and search_code handle different types of searches, get_class/get_function/get_file_summary provide targeted code extraction, get_dependencies analyzes imports, get_repository_structure shows layout, get_relevant_context answers natural language queries, and get_usage_stats tracks metrics. The descriptions explicitly differentiate tools (e.g., 'PREFERRED OVER Read' vs. 'PREFERRED OVER Grep'), eliminating ambiguity.
All tool names follow a consistent verb_noun pattern with snake_case: clear_cache, find_similar, find_symbol, get_class, get_dependencies, get_file_summary, get_function, get_relevant_context, get_repository_structure, get_usage_stats, index_repository, and search_code. The naming is predictable and uniform throughout the set, making it easy for agents to understand and use.
With 12 tools, the count is well-scoped for a context manager focused on code analysis and repository navigation. Each tool serves a specific, necessary function in the workflow (e.g., indexing, searching, extracting code, analyzing dependencies, tracking usage), and none appear redundant or excessive. This aligns with typical server scopes of 3-15 tools, providing comprehensive coverage without bloat.
The tool set offers complete coverage for the domain of code context management: it includes setup (index_repository, clear_cache), navigation (get_repository_structure), targeted code access (get_class, get_function, get_file_summary), search capabilities (find_symbol, search_code, find_similar), dependency analysis (get_dependencies), contextual queries (get_relevant_context), and monitoring (get_usage_stats). There are no obvious gaps; agents can perform a full lifecycle from indexing to detailed code analysis and optimization.